Tableau's Dreamforce 2026 agentic analytics keynote reviewed

0:00Okay, Dreamforce was last week.

0:02I'm checking out the keynote.

0:04I'm just gonna go through this.

0:05I'll make another video of this with a much shorter take on this.

0:08My takeaways as it were.

0:10But uh yeah, I find this

0:11interesting.

0:11I have

0:11I've literally been out of the loop of Tableau for the last

0:14four months.

0:15You've probably seen the cadence of videos drop down a little bit.

0:18That's very much deliberate.

0:19I've been trying other tools.

0:20I've been exploring other technologies.

0:22More on that

0:23maybe in a separate video.

0:24This video is about Tableau.

0:26Let's check out the keynote.

0:27As ever, they always start with a forward-looking statement.

0:30I find this a little funny, but yeah, here we go.

0:32Let's see what's

0:32what's in store.

0:33I might skip this.

0:35Executive Vice President and General Manager

0:37Tableau

0:38Salesforce

0:39Mark Record.

0:42So Mark Wrecker is the new CEO

0:44equivalent.

0:45To the Tableau

0:47keynote

0:47in the next 50 minutes

0:50We're going to share with you how you can take your organization from the data-driven enterprise to the

0:57agentic

0:58analytics

0:58enterprise with trust and speed.

1:01But we're going to start with a thank you.

1:04I want to say a huge thank you to our Salesforce customers, to our Tableau

1:08customers, to our partners, and to our communities, our Trailblazers, and

1:14It's always interesting, just as a context, just to look at the room.

1:17Because the thing is, this is not the Tableau keynote

1:20at the Tableau Conference.

1:21This is the Tableau

1:22Keynote at the Dreamforce conference.

1:23It's a lot smaller audience just out of the gate

1:26But it's also sad to see that there aren't as many failed seats.

1:29I think sometimes the camera always does it injustice.

1:32I have seen worse in terms of attendance, so

1:35it speaks to sometimes a difficult challenge that I think Salesforce have, which is how do they bring Tableau

1:40to their existing customers who are primarily just interested in

1:44It's a bit like I don't know sending a Tableau user to I don't know

1:47a Databricks conference.

1:48It's going to be a tough crowd because Databricks is a very different product to what Tableau is.

1:52Yeah, it's just an observation, nothing really to take away from me.

1:54You also can't take it as empirical evidence of anything.

1:56It's just an observation.

1:58customers, to our partners, into

2:00our communities, our trailblazers, and the data

2:03fan.

2:04If you are in this room right now,

2:06You are a part of the Data Fam.

2:09So let's start with a huge warm welcome and thank you to everyone in the Data Fam.

2:15Make some

2:15noise.

2:19I will say,

2:21maybe this is definitely a separate video, I think this term

2:25has become

2:28a little, what's the word?

2:31I think there is

2:33toxic is the wrong word.

2:34That's what that's the word I wanted to use, but that's not true.

2:38I think it's become a very

2:41There's something about it that doesn't sit right, if that makes sense.

2:44I know

2:45I'll talk about it in a separate video.

2:46This has got nothing to do with the members of the community.

2:49Let me just be absolutely clear.

2:50It's got everything to do with the way that Tableau

2:53have used this term from a marketing standpoint.

2:57And I think what it's come to represent, if that makes sense.

2:59Yeah, more on that later.

3:00But everyone on this front row, legends of the Tableau

3:04ecosystem and world, I know some of these faces, again, it's not the Tableau Conference, so it's not your typical faces, but needless to say

3:10Yeah, this is a pretty well established crew.

3:12There's also some Salesforce employees here in the crowd.

3:15They sit very close to the front because they are monitoring the keynote.

3:18They're part of making sure it runs smoothly.

3:20So it's not all datafam here at the front

3:22It has also got a few staff and

3:24colleagues of the Salesforce ecosystem.

3:26So yeah, let's carry on.

3:29Because

3:31we have

3:32something

3:32to celebrate.

3:34Because together, we have built something incredibly special over 20 years.

3:40You just heard Matthew McConaughey describing it.

3:43We built the data-driven enterprise.

3:46Why

3:47So

3:47we could make better decisions

3:50that led to better outcomes, more

3:53leads, more pipeline, better win rates, improved customer experience.

3:59And there were three foundational elements

4:03to the data-driven enterprise.

4:05The first is our trusted data foundation.

4:09Data

4:09deterministic semantics

4:11in single sources of truth

4:14determinants.

4:14So that we could power

4:16world-class analytics.

4:18The term deterministic has become like this word where

4:22it's like a safety word, it's a warm blanket.

4:24What it really means is hey, can I use semantic models in a predictable way with AI on top?

4:29So

4:30Even if a hundred different people ask the question, do they come out with the right answer, the same answer and the right answer?

4:36That's really all it

4:38The thing is, in analytics, everything was always deterministic to start with.

4:42So

4:42I feel like we've just created this buzzword.

4:45to describe what we've been doing this whole time.

4:48That was the whole point of things like Tableau and Business Intelligence.

4:51How do you enable everyone to answer the question and arrive at the same answer

4:55and get away from the spreadsheet ecosystem?

4:57And for some reason

4:59It's back on the table again when it was already a solved problem.

5:02Yes, thanks to AI, but yeah,

5:04it's

5:05sometimes

5:06history rhymes as they say.

5:07It doesn't repeat.

5:13so that we

5:14humans

5:15could look at insights and go take action.

5:18Test.

5:20And most

5:21importantly

5:22Number three, a data-driven culture

5:25through every function of our organizations, sales

5:29,

5:30service

5:31marketing, product, finance, every decision rooted in data.

5:37And we should celebrate that.

5:39Nothing

5:41Ever held us back, not our imagination, not our possibilities, except for a few simple things.

5:48Resources, time, and technical skill set

5:53But

5:53today, with

5:55agentic

5:55intelligence, those constraints have eliminated to basically zero

6:01We have

6:02incredible new possibilities

6:04with a genetic intelligence.

6:06Basically

6:06zero

6:06is that's a misunderstanding of the fact that

6:09last two

6:10percent is the hardest two

6:11percent to get

6:13If you give someone Tableau, they can go from 0 to 70 very quickly with very basic understanding of Tableau

6:20However, I'll tell you, I know Tableau

6:22visionaries, Tableau

6:23experts, Tableau

6:25gurus

6:26who still do not fully understand

6:29the power of the data model.

6:30I've made videos with Kirk

6:32and even I still don't fully understand the power of the data model.

6:35I keep discovering new things every single time.

6:37And to say that it's gone to nearly zero, I think is a misunderstanding of two things, enablement

6:44and

6:44I think the true challenges businesses have from a cultural perspective.

6:48It's not that the tools can't do it

6:50It's that actually it's really hard to bring people up to that level and that's just always going to be an inherent problem.

6:56You can't go to basically zero

6:58So I know it's

6:59a genetic intelligence.

7:00I know the point here is that hey, you can now prompt and get the answer you want.

7:03But I hear that and I'm like, oh

7:05no, you need to come and check out some of the businesses I've helped.

7:08Basically

7:09zero

7:10will never exist.

7:11What you can do is bring everyone one step closer to the kind of things they want to do, the kind of workflows they want to finish, and support them in making sure that more people can get across the line.

7:22But basically zero

7:23is a bit of a bold claim.

7:24So yeah.

7:25Anyway, that's just my take.

7:27A business user doesn't need to know SQL.

7:30You don't need to be a data engineer.

7:34But you do need to be a data engineer if you build a data model that person uses.

7:37So again, it's

7:38yeah,

7:39business user doesn't need to know SQL, but when that thing is wrong, yeah, suddenly the SQL gets handy.

7:45That's also never been a premise of Tableau.

7:47I don't think business users

7:49end users of Tableau

7:50have ever needed to know SQL.

7:51That's literally never been a requirement.

7:53Even well-established Tableau developers, I would argue

7:57sometimes have never touched SQL.

7:59So

7:59again, this is not a new thing.

8:01This is not a premise that we've heard before.

8:03So

8:04sorry, a premise that

8:05that's really new to the industry.

8:06And I'm reacting to these things as I hear them because I think what I'm trying to do is construct

8:11a view, a perspective on look, what is the standpoint of the offering they're about to give us?

8:15And so if the underlying premise

8:17about those things are not really truly hitting the mark, then I think there is a

8:24what

8:24we what we're presented can miss the mark and I can I'm happy with that

8:28I'm happy with not agreeing with what's in the keynote and I can evaluate it for what it is.

8:33But if I don't even agree with the original premise, then it's really hard for me to have a perspective.

8:37So I'm not just

8:39reacting to it to be a skeptic.

8:40I'm trying to pick apart the core premise of the key arguments.

8:43And this is the keynote, this is the first two minutes.

8:45This is where you make your stance.

8:47I've watched so many Tableau keynotes, I've been in Tableau Keynotes.

8:50I know how some of this works.

8:51In the past, I've helped make these keynotes better.

8:54I know how the stories are structured.

8:56This first five minutes is about setting your store.

8:59Even if you go way back to when Christian Chabot was doing keynotes, the first ten

9:04minutes was about setting the store for what we were about to see.

9:06I think it's fair to

9:07unpick each and every one of these points

9:10You simply express in plain English

9:14what you'd like

9:15in an agent

9:16executes that instantly

9:18It gives us

9:20incredible

9:22new speed in intelligence.

9:24Speed in intelligence that gives every single person in the organization

9:30The power to create, build, and consume analytics

9:34in whole new ways.

9:35This

9:36fundamentally changes where

9:38and

9:39how we

9:40work.

9:41Right?

9:49It's pretty interesting.

9:50Giving everyone the power to create, build, consume

9:53analytics in whole new ways

9:58Yeah.

9:59Yeah.

10:02What I feel is disingenuous

10:04about this is like where is this possible?

10:07Tableau is a brand name now.

10:09The ecosystem spans so many different parts of Tableau and Salesforce, but it's not clean things.

10:15So for example, when you look at AI-powered modeling, that absolutely does exist in the Tableau

10:20ecosystem.

10:21But it's in action

10:22it's in Tableau

10:23Semantics, which is in Salesforce.

10:25I will argue

10:26companies like TorSpot

10:27actually have it as part of their core product.

10:29We don't quite yet have that as part of Tableau

10:32Desktop.

10:32So

10:34yeah, again, I

10:35'm being pedantic here.

10:36Let's just move on.

10:38Agentic

10:39analytics is embedded in your chats

10:42It's in your flow of work.

10:44It's proactive, delivering you an insight and recommending an action.

10:50What it really empowers us to do

10:53is

10:53to see, understand, and act on data at the speed of thought.

11:00And that's Tableau

11:01's mission statement.

11:03Helping

11:03it.

11:04It didn't always used to be

11:05the mission statement.

11:06See and understand data was it.

11:08You didn't need to need

11:09you didn't need to put in act

11:12on

11:12at the speed of thought.

11:13Understanding happens at the speed of thought.

11:15You always needed to act on your data.

11:17That's

11:17it's a bit like saying, I'm going to be the world's best, fastest sprinter.

11:24It's like

11:25you add that additional qualifier that is inherently implied in the

11:29outcome.

11:30So

11:30when you see and understand data, you act on it

11:34and you do it as quick as you possibly can.

11:37People see, understand, and act on data.

11:41And that is why the Tableau team is so excited

11:45for today, because it takes our mission to entirely new levels.

11:50All

11:51those things that we imagine, those possibilities of the future, we can deliver

11:56today, now.

11:58So

12:00This is

12:01the question.

12:02How do you become the Agentich Analytics Enterprise?

12:07Like we're operating here at Salesforce today.

12:10So I have some very, very good news.

12:13All right.

12:13You've already built the hard part, which is the data-driven enterprise, the foundation of that.

12:20The culture.

12:22And what was that culture built on?

12:24It was built on trust.

12:26Trust that the data was accurate.

12:29Trust that the dashboard was right.

12:31Trust

12:32that

12:32the right users

12:34could see the right

12:36sensitive enterprise and customer data.

12:39Trust.

12:40that it

12:41understood

12:42the specific

12:43nuances of your organization, how you define revenue, how you defined a lead.

12:50your

12:50business's context, the trust of deterministic semantics.

12:56So

12:56how do you become the Agentic

12:58Analytics

12:58Enterprise?

12:59You take probabilistic intelligence and the speed and power of that

13:04and marry it with your trusted

13:07data-driven foundation.

13:10And

13:10that enables you to

13:12take

13:13your enterprise to entirely new levels.

13:17Your trusted data foundation with the power of agentic

13:21intelligence.

13:23So

13:23you can in real time

13:25see, understand,

13:27and act on your data at

13:29the speed of your own thoughts.

13:32Okay?

13:34This is precisely what we have built at Tableau

13:38and why we are so excited today.

13:41Because for the next 40 minutes, you're going to see a lot of new innovation

13:46That enables you to become the Agentic

13:49Analytics

13:50Enterprise with speed

13:52and trust.

13:54Okay.

13:55We have four layers of our platform.

13:58All of them are powered by agenc intelligence.

14:01Man, I

14:02have to

14:02say something about platform diagrams.

14:04It's

14:05I feel like

14:05someone

14:06Salesforce has forgotten what a platform diagram is.

14:10Because

14:11I miss the old ones.

14:13I I genuinely do.

14:14And they're boring.

14:15Platform diagrams are boring

14:17They're not designed to be exciting.

14:19And I feel like what they try and do here is they try and show you the platform in a in an oversimplic way.

14:26Which makes the boundaries between how the different components come together opaque.

14:31And so you struggle to see your own enterprise in this diagram

14:35because you can't easily see

14:37Where certain workflows you have in your analytical team sit.

14:41So I can see my data here at the bottom

14:44and I can see headless analytics at the top.

14:46To me that

14:46just

14:47wait, what?

14:47No, I don't have any headless analytics

14:49solutions

14:50Okay, what's an action engine?

14:52Okay, a decision engine, a knowledge engine.

14:53Oh, I know what a data engine is.

14:55That's what's Powers

14:56Tableau

14:57Server.

14:57Or

14:57is it?

14:58It says it's Powers

14:59Agentic

14:59Analytics with Hyper Data.

15:01Okay, so that is just a data engine.

15:03It's always been there

15:04But then the knowledge engine, decision engine, action engine, these are all radically foreign to me and I wouldn't know how to equate them to a modern analytics

15:11stack.

15:12I know what they're trying to do.

15:13They're trying to

15:14ref

15:15reframe

15:16that platform diagram in a new light, but it loses people immediately, I think, because I think a lot of people who look at this are people who run analytics teams and they're used to seeing

15:26platform diagrams that really easily show what's possible with a platform rather than sort of the aspiration if that makes sense.

15:34That said, I will also say

15:35headless analytics

15:36sounds awfully familiar

15:39Anyway, open and extensible stuff, MCP's API skills, bring your own LLM apps.

15:44To be honest, Tableau

15:45has been on an absolute tear on extensibility.

15:49So much has happened.

15:50The community has developed so many different things.

15:52By community I'm talking about developers, partners, and even people now doing sort of vibe coding solutions, building their own thing.

15:59It's now so easy.

16:00Genuinely anyone could be a Tableau developer in today's world

16:03A lot of the code

16:04bases are very easy to access.

16:05But yeah, this this sort of middle section of this diagram always loses me because I'm like, oh

16:10come on.

16:11What

16:12like what is this actually?

16:13But anyway, let's carry on.

16:15our

16:16hyper

16:16data engine, which allows you to bring in data from any source, structured or unstructured, to power agentic

16:24analytics.

16:25our

16:25knowledge engine that takes our human

16:28semantic understanding of our organization, our unique business context.

16:34and puts that into a graph where agents can thrive on and drive accuracy

16:41so you can

16:42use our decisions engine

16:44and empower everyone in your organization

16:47to build, create, and explore freely

16:51on any single

16:52surface.

16:53And then

16:54why do we look at analytics?

16:57We look at it to take actions, our

17:00actions engine

17:01that automates workflows with agents.

17:05So when you have the right insight delivered to the right person at the right time, the action is automated.

17:11And if you joined Mark's keynote yesterday, you saw this everywhere.

17:15Headless analytics, because we want you to take your trusted knowledge so you can drive decisions on any surface your organization is working on

17:24Whether that is Slack or Coworker

17:27or CloudForce

17:29or Microsoft Teams or

17:32did

17:32he say co-worker?

17:33I think he did.

17:35Teams or ChatGBT

17:46Or Microsoft Teams

17:48TVs

17:49co-worker

17:49co-work

17:51So

17:51in Mark's keynote, you saw analytics

17:54is omnipresent.

17:55It's everywhere.

17:57So this is the Salesforce.

17:59platform diagram, in a way

18:01this makes more sense to me from a Salesforce centric perspective.

18:04I can actually see where each part of Salesforce ecosystem sits

18:08in the whole thing.

18:09What I don't like

18:10is

18:11the

18:12The forcification of everything.

18:14Everything is a force.

18:15Agent force, clawed force, AI force, dream force, tableau

18:20force

18:20Agent force.

18:21Just so many forces.

18:22He's forcing stuff on people all the time.

18:25It just gets a little bit much.

18:26Yeah.

18:27Okay, it's fine.

18:27It's a Salesforce diagram, but yeah.

18:30Tableau's knowledge, context, and semantics are powering our customer 360, they're powering Agent Force, and they're accessible in every AI

18:40force

18:41surface.

18:42Okay, so

18:45we know this.

18:46We built the data-driven enterprise.

18:48That wasn't easy

18:50The technology

18:51alone is insufficient.

18:54This is also about cultural change management.

18:57And so for the rest of the keynote, we're going to share our playbook that you can use to become the agentic analytics.

19:04Enterprise.

19:05And there are three steps.

19:07And in each of those steps, you are going to see some

19:11really exciting new innovation.

19:13You're going to see four products that you can use now.

19:16to transform your organization from the data-driven enterprise to the agentic analytics

19:22enterprise.

19:23And we're also going to be joined by three

19:26of our most inspiring customers who are going to share how they're using Tableau

19:31today

19:32to transform their

19:33organization.

19:35So quickly, what are these three steps?

19:37Number one, empower every single person

19:41to be a builder in your organization.

19:44We are in the greatest era

19:46of universal self-service

19:49analytics.

19:50Number two, this era of

19:53analytics is about having intelligence come to you.

19:56It's about eliminating the time from insight to action, whether that's by a human or an agent.

20:04Delivering the right insight to the right person with the right recommended action, where

20:09they are working.

20:11And number three, and most

20:14important, is build on what we created

20:17In the data-driven enterprise.

20:20Give agentic

20:21intelligence a semantic understanding of your organization.

20:26So it can thrive not just in analytics use cases, but in any use cases across your enterprise, which will enable you to allow everyone to be a builder and drive intelligence to them where they're actually working.

20:41And so

20:41I have a saying, which is

20:43show is way better than tell.

20:45So we're going to show you here in a minute.

20:48But first

20:49I would like everyone to give an incredibly warm welcome to Tableau's new Chief Marketing Officer, Brock Jones.

21:03All right.

21:04Thanks, Mark.

21:05Now

21:06we heard from Mark's opening that AI is changing everything in analytics.

21:10And that's especially true when it comes to how we create and build.

21:15Because now

21:16everyone is vibe coding with analytics.

21:19Think about that last time you were in a meeting.

21:21You had a great idea for a dashboard or an interactive data app.

21:25You no longer have to wait for that to become a reality.

21:27The reason that phrase doesn't sit quite right with me is because I don't think everyone is vibricating with analytics.

21:33I think analysts are vibe coding with analytics, but business users are

21:45Vibecoding is such a loaded term because yes of course business users are taking Excel dumps putting into Claude

21:51and say

21:51build me a dashboard

21:53That is not vibe coding.

21:54That's like data exploration inside of ChatGPT or Claude, whatever

21:59you use.

22:00Vibecoding to me

22:02is

22:06It's the next level up, which is persisting that and having that be something that you can come back to, have it update in the background and actually work.

22:14Essentially

22:15replace the pipeline you used to have with a solution you

22:18built.

22:19Analysts are vibe coding solutions that update in the background.

22:23Absolutely.

22:23I've seen it, I've done it myself.

22:25But I

22:26I wouldn't say

22:27it's so weird when you talk about the audience because sometimes you've got to remember that the majority of people who use the product

22:33are not analysts.

22:34They are end

22:35users.

22:35And so

22:36I struggle sometimes to tell the difference between whether Salesforce are talking about the analyst, whether they're talking about the end

22:43user.

22:43If it's the end user I'd argue

22:45they're not really vibe

22:46coding.

22:47But if they're analysts absolutely they're vibe coding, but then they shouldn't be.

22:51In some cases, definitely not.

22:54So yeah

22:55You don't have to go to your data team.

22:56Instead, you can sit down in your favorite AI tool, describe what you want, and see it magically appear on screen.

23:03And that is honestly incredible.

23:05At Tableau, we're really excited about this new reality because it's helping more people see, understand, and act on their data.

23:13But we think we're just scratching the surface here.

23:16We've been asking ourselves a question, what if we can make it better?

23:21What if the world's best visualization engine met vibecoding?

23:25What if the design best practices that actually make a chart make sense, the ones we've developed over the past 20 years, were automatically leveraged inside every prompt?

23:34What if you had control

23:36right when you needed it to switch between a simple chat bar and a powerful drag and drop canvas for more precise editing?

23:44And what if every answer and output we got from AI was trusted?

23:48No second guessing, just

23:50pure confidence that what you're building is accurate and ready to be used across

23:56your teams.

23:58That's not a hypothetical reality anymore.

24:01That's here today.

24:03Tableau

24:03Studio is a completely reimagined version of Tableau.

24:06We've built it from the ground up.

24:08We're reimagining how desktop works.

24:09And it is the ultimate workspace for vibe coding analytics.

24:13With Studio, you can author

24:16refine and publish metrics, dashboards, calculations, vibe-coded extensions, all in the same workspace.

24:22And the best part is

24:23always in control.

24:25If you ever want to jump out of the AI chat to tweak a mark or manually adjust a feature, the drag and drop canvas is right there, seamlessly available for you to switch over to.

24:34So

24:35this feels like a reimagined interface for Tableau

24:39with the LLM

24:41chat window almost at the core, but also a mix of some of the other stuff that you have in the Tableau ecosystem.

24:48brought into the center.

24:50Now

24:51I will say, if I may say so, I have talked about something similar to this.

24:55Nowhere near as good as this.

24:57I did it before AI was a big thing.

24:59But I've always talked about the ideal version of Tableau, a one unified workspace.

25:04And I've got a video on it.

25:05It's up on the channel.

25:06I'll link to it.

25:07I'll put a thumbnail up on the screen.

25:09I really hope this is actually that.

25:11I really hope this is a fully realized version of that.

25:14Let's keep watching.

25:17And Studio also comes through analyst

25:19best instincts built in, and we call these skills.

25:22And you can turn them on when you're building with AI.

25:25Say you've got a slow running dashboard and you need to increase the performance.

25:28Why not turn on the performance skill and have the agent get to work for you under the hood?

25:33Nowhere else do speed and trust converge so seamlessly to create the ultimate experience for vibe coding analytics.

25:41You all want to see it in action?

25:43Yeah?

25:44Okay.

25:45Give a warm round of applause for our demo drivers today.

25:48We have John

25:49and Caitlin on the stage.

25:51Okay.

25:52There it is.

25:53There's a tableau card.

25:54And in today's

25:55demo.

25:56I'm

25:56picking it

25:57apart straight away.

25:58So on the left hand side it looks like we have

26:01Almost the equivalent of

26:02the server

26:04side pane, but it's been compressed.

26:06But what is really stark and different is this search box in the middle.

26:09What do you want to start doing?

26:12It immediately makes me wonder, oh God, what are the AI credits going to be like for this?

26:16Because before in my analytics experience, I didn't need any AI credits to build a solution.

26:20It was just free in my time, right?

26:22But

26:24yeah, there's so much to unpick here.

26:25Let's keep watching.

26:26We're at Fluxwell, a fast-growing AI startup.

26:29And I'm a new sales operation analyst

26:31there.

26:32So I'm still getting my bearings.

26:33I don't know everything about the data yet, but my CRO has come to me and he needs help immediately

26:38understanding how the quarter is performing.

26:40Okay.

26:40And so

26:41he's going to ask me some questions about the quarter, but before we even get into those, I want to take a moment and look at this screen here.

26:46Does this look different to you all?

26:47Does it

26:48seem new?

26:49Yeah.

26:49That's right, this is a completely reimagined UI for Tableau.

26:53Okay, it's purpose-built for the modern AI era, and it all starts with a simple chat bar.

26:58And so my CRO comes to me and that first question he asks is hey Brock, can you help me understand

27:03how w

27:04things are looking with our top small business deals for the quarter?

27:06He wants to see which deals are rising to the top.

27:08So we're gonna go ahead and ask

27:10Studio for some help there

27:12It's gonna do some thinking and perfect.

27:14Very cool.

27:14Tableau

27:15brings us the right visualization.

27:16Okay.

27:17I don't know if this is like

27:19it looks like it is the real thing, if that makes sense.

27:22The reason this

27:22let me show you why I have this weird

27:24um

27:24tool at the top.

27:25I have this tool that hides

27:28icons, that's why it has this blue bar at the top, but needless to say

27:32Yeah, this looks like it's running proper tableau.

27:35This is not uh

27:36what's the word

27:37Figma demo or something like that.

27:39But

27:40really interesting.

27:40This also feels like it's in a web browser

27:42It feels like a browser

27:44first

27:44experience in there in that little window, which means you can have it in server or cloud, whatever.

27:48But yeah.

27:49Now what I also like about this is, okay, this chart here.

27:53What I hope he's about to say is that when I click open, I get the tableau interface and I can customize this.

27:58But I also see things that I've never seen before.

28:00For example, these icons related to this attribute.

28:03So

28:04That suggests that there's also some new capability coming into desktop to enable a much richer experience.

28:09But instead of guessing, let's keep watching

28:11Just got a stacked bar chart here.

28:13We can see Northwind aerospace at the top.

28:15That's the big deal.

28:16We want to keep some eyes on that one this quarter.

28:19And then my CRO is asking me, hey

28:21I actually want the bigger picture, Brock.

28:23I want to look at the full view of my pipeline and how it's progressing globally.

28:26Can you help me with that?

28:28So let's ask Studio.

28:32Okay, great.

28:33Another visualization.

28:34This is perfect for our purpose, right?

28:36We want to see how pipeline's been progressing over time and with this Sankey chart we can see how things move from stage to stage.

28:42This is Tableau

28:42's power picking the right visualization in the moment.

28:45But if I want to peek under the hood, maybe I want to take control.

28:47Maybe I want to change the color or apply a filter.

28:49I can do that.

28:50I can seamlessly switch from the conversational UI

28:53straight to this drag and drop canvas that you all know and love, right?

28:57This is what gives me the power to build in my way when I want.

29:00And let's go ahead and change the chart colors.

29:04And I'll pause it here and I'll say this is actually a welcome refresh of the Chrome.

29:08Tableau Chrome has been a little bit dated, so

29:11just giving it a new lick of paint.

29:12I think I remember it was Tableau

29:1310 where the interface took this sort of.

29:16modern leap and then I believe in 2020, 2021 there was another little bump to the icons.

29:22Things kind of

29:23vector shapes started coming back in.

29:25They changed some of the icons for links between data sources.

29:30Things went to the web and it got a bit of enhancement, but it's never really had like a hey, well if you just blow away the canvas and we put everything back on, but we did it with a little bit more thought and consideration.

29:41This feels to me like what Tableau

29:43Next should have been.

29:45Can you just imagine if they'd just taken this move and they'd done it four years ago?

29:50I think that would have been much, much bolder.

29:53Man

29:54Yeah.

29:55Yeah,

29:55yeah, story

29:56for another day.

29:57Anyway, let's keep going.

29:58Alright, cool.

30:00Yeah, so this is looking good.

30:01But here's one other idea I have.

30:04I know that this data is for a global view.

30:07I think it'd be really cool to plot it on a map or a globe.

30:10Now, I'm really good in Tableau, but I'm not a developer, so I'm dropping out of that drag and drop view.

30:15I'm just going to ask the agent for some help.

30:16Let's see what it can do for us.

30:19Can you visualize our revenue on globe?

30:21Alright.

30:21Okay, cool.

30:22Whoa.

30:23This is awesome, right?

30:25Look at this.

30:26We can see this view of the globe.

30:27We can

30:28hover over charts, fully interactive

30:31Studio is able to help me build this.

30:33This is gonna be amazing to show.

30:36Can you visualize our revenue?

30:38Can I just

30:39Qualify what he said to justify this.

30:49Sorry, big

30:50spots.

30:52Okay.

30:54I'm just gonna ask the agent for some help.

30:55Let's see what it can do for us.

30:58How did you

30:59talk about

31:01the

31:02global curve?

31:02Uh

31:04data

31:05is for a global view.

31:06I think it'd be really cool to plot it on a map or a globe.

31:09It would be really cool to plot it on a map or a globe.

31:12Say

31:13no analyst

31:14ever

31:14when analyzing worldwide sales.

31:18Like literally.

31:20I kind of just wanted to double check the qualifier there.

31:23Yes, it would be really cool, but it would be also really

31:26not Viz

31:27's best practice.

31:28Stephen Phew would be turning in his grave right now.

31:31Now I'm really

31:32having

31:32Tableau, but I'm not a developer, so I'm dropping out of that drag and drop view.

31:35I'm just gonna ask the agent for

31:36a

31:36And by the way, no Tableau

31:38developer

31:38can build a Globe map.

31:39They have to go get an extension and install it.

31:41They're not vibe coding the the the the Globe map.

31:43It's a little it's a bit

31:45Oh yeah.

31:46Let's see what it can do for us.

31:51All right.

31:52Whoa.

31:53This is awesome, right?

31:55Look at this.

31:56We can see this view of the globe.

31:58We can hover over charts

31:59fully interactive.

32:01Studio is able to help me build this.

32:03This is gonna be amazing.

32:043D

32:06Balcharts

32:07mapped on top of a 3D globe.

32:09It's the worst of both worlds.

32:12On one hand

32:14Very nice execution.

32:16If I just take away the 3D

32:19-ness

32:19of it, like that is a really cool app.

32:22Let's just be clear.

32:23Okay, it's built something, it's built

32:24in there.

32:25I don't know if it's

32:25vibe coded this app in real time

32:27in the hands.

32:28It probably has.

32:30But you

32:30go and show a 3D globe

32:32and a bar chart.

32:33You don't take you basically won't get taken seriously, at least in the analytics

32:36field

32:37As it relates to business best practice, just helping people see and understand stuff.

32:40Literally the motte, right?

32:42This is not helping with that.

32:43It's a disconnect between the mission and the actual purpose.

32:47Cool demo, bro.

32:49Completely useless in a business context.

32:52Fine on Tableau

32:53Public.

32:53Completely fine on Tableau

32:54Public.

32:55But that's not what we're seeing here.

32:56We're seeing the full paid business

32:58facing

32:59product.

32:59So

33:00you could have shown something equally impressive that a developer could not have done as quickly

33:05that could have been taken more seriously.

33:09So that's the power of studio.

33:12Pretty cool, right?

33:13Yeah?

33:16Alright.

33:18Now look we know not everyone's gonna want to sign

33:20up.

33:24Big space

33:24out

33:24screen to a tiny little screen.

33:26So I don't know

33:26this is there may be a different laptop with the zoom

33:29scaling set right up.

33:30But yeah.

33:38For Chat GPT.

33:39And we want to give them the same freedom to build with Tableau

33:42directly inside these interfaces as well.

33:45And so that's why we're bringing that same trusted foundation you saw in Studio directly into your favorite AI tools.

33:51with

33:52Tableau

33:52data apps.

33:54Tableau

33:55data apps will allow anyone in your organization to easily build custom applications inside third-party AI tools.

34:02That you can actually trust and operationalize at scale.

34:05That's all thanks to the power of Tableau.

34:08So let's go back into our demo to take a look.

34:10And this time

34:12I'm a regional sales leader at Fluxwell.

34:14But I don't start my day in studio.

34:16I start my day in Claude.

34:18And I've just had my morning coffee and I'm gonna ask Claude

34:21the same thing I do at the beginning of my day

34:22every day, how are my accounts looking by segment?

34:25Let's start there.

34:28Okay, great.

34:30We have a really easy to understand visual.

34:31I see enterprises

34:32leading the way from my accounts.

34:34No surprise there.

34:35But this has got me thinking, because I've got an upcoming sales on site with my team, and I actually want to build something more interactive with

34:41this data.

34:41Something I can use to actually select accounts, drill down to them, do some planning with my teams in the room.

34:47So I need more than a simple chart.

34:49I need a full-blown application, and this is where Tableau

34:51Data Apps comes into play.

34:53Now I'm not sure exactly what I want, but I've got a general direction, so I'll give Claude some thoughts here

34:59And what we'll notice is that because Cloud's connected

35:02Tableau, it can act like my best analysts and help guide the experience.

35:06It understands my semantics, models, how my business runs, and so it can make suggestions like

35:11which outcome am I hoping to drive?

35:13I've been focused a lot on expansions, select that.

35:16It can also recommend actions we can build into the application because Tableau.

35:20So what this must be doing is it's using the Tableau MCP.

35:23to hook into the data sources inside of the platform and then be able to have this experience.

35:28What I've not yet seen is

35:31Can it build assets in the Tableau

35:33ecosystem for you from here, but then present them back here?

35:36So for example, if I vibe

35:38code a chart in here

35:40and I say yes, save this.

35:42When I go back to Tableau

35:43Cloud, is it sitting there waiting for me in my favorites or in my most recent

35:47basically?

35:47So let's wait and see.

35:49Natively integrated with the Salesforce environment.

35:51So we can actually select an action like run an expansion play directly from the application.

35:56So we're going to click build.

35:58Cloud's going to take all this input from the chat, all the knowledge it has from Tableau, wrap it up into

36:04UI, and here's the cool part.

36:06It's going to say ready to publish to Tableau.

36:10Not locally

36:12to my computer, not via some HTML file.

36:15Publish it directly to

36:16Tableau

36:16so it's production ready.

36:17This is the power of Tableau data apps.

36:20Okay, let's check it out

36:24Whoa, all right, this is awesome, right?

36:26Look at this interactive application.

36:28We can hover

36:28over each account.

36:30All right, we can see the accounts.

36:33If it published it to Tableau, why am I

36:36in

36:36Tableau next?

36:41This is sometimes when I don't get

36:43come

36:43on.

36:44For for proper uh tableau users

36:47is so confusing.

36:48What's more, if you look along the top, check this out.

36:50If we switch to data

36:52view.

36:52I can actually see the semantic models, the formulas.

36:56So publish the tableau means publish the tableau next.

36:59That's what this actually means.

37:01And the data

37:02app lives in the Salesforce ecosystem.

37:04Not the tableau

37:05desktop that we need.

37:06This is this gets me every time.

37:08It's so annoying.

37:09I'm sat here getting baited left and center by Salesforce every single time.

37:14We're like, oh, this is a great idea.

37:16Oh, let's go

37:18Ah

37:181000

37:20force.

37:20Come on.

37:23Every time it gets me.

37:25It's so annoying.

37:28This is an application I can actually trust, right?

37:30It's not hallucinated metrics.

37:32We know we can use this.

37:33This is the power of Tableau

37:35Data Apps.

37:36And if we click into

37:36any given account, we see something else.

37:38We see this button launch expansion plate.

37:40That's that suggested action that Cloud Cloud recommended we built.

37:43And they

37:44have me good because they switched

37:45between

37:46Tableau studio

37:48to data apps

37:50and so in your head you you blend this experience into one

37:53amorphous thing

37:54But again, if you work in this stuff, like I know this is not Tableau

37:58Desktop.

37:58We can't lay things out like this.

37:59I wish we could.

38:00I wish it was this good.

38:02Ah

38:02,

38:03it's so annoying.

38:05I was like, great mate

38:08of building with Tableau

38:09as your foundation.

38:10And what's more, this is not just reserved for me.

38:12Okay, I can share this

38:16and my favorite channels like Slack.

38:17Yeah, Slack.

38:18And because it's all built on Tableau, it's gonna honor the governance and permission rules I've set.

38:23So frustrated.

38:29Pretty cool, right?

38:29Yeah?

38:30Okay.

38:31So look, we've seen a lot

38:33We've seen building in studio.

38:35We've seen building in your favorite AI tools.

38:38But

38:39we need to make sure that we're building across every interface, right?

38:42Because in order to power everyone to build

38:44We need that same trusted foundation to work across all of our interfaces.

38:48So

38:48what I want to do next is welcome to the stage our first customer, Jake Schlon from CrowdStrike, who's going to talk about how he's empowering

38:55everyone in his organization to be a builder.

38:57Jake, come on up to the stage.

39:01So

39:02whenever I guess for a customer, I I respect

39:05that people who come onto the stage are doing a a pretty big thing.

39:10They're talking about their own account.

39:13I cannot critique them and what they're doing and how they're running their teams.

39:17I don't walk in their shoes.

39:19So I tend to skip these customer sections because

39:22Customers don't ask for people like me to dunk on them.

39:24They definitely don't need it.

39:26But

39:26Salesforce and the people who present the products that we all pay for and use, yeah, that's that's that I think that's very fair.

39:33So

39:33I will listen to a bit of this.

39:34If there's a little bit of product, I will watch it, but I won't really comment on

39:39on

39:39the customer if that makes sense.

39:45Thanks, Jake.

39:46Absolutely.

39:47Okay, Jake.

39:48So

39:48why don't you tell the room a bit about how you're using Tableau at CrowdStrike and what are some of the initiatives you're most excited about right now?

39:55Absolutely.

39:55Tableau is the source of truth, our trusted partner

39:59for everything that's powering our business at CrowdStrike on the analytics

40:03front.

40:03Something that I'm very excited about is we launched what we call 360 dashboards.

40:08They're very similar to

40:13But it's a whole list of

40:14the customer conversation

40:15of everything

40:16from

40:16if it's on a count

40:17Okay, or my reports.

40:28Cool metric here, reduction insight.

40:30I always find it interesting how these are validated, but yeah, hey, the customer's saying that this is the case.

40:34So

40:35you can take a screenshot of this and say CrowdStrike

40:37said

40:3750%

40:38reduction in insight.

40:39And this is

40:40I think they're talking about Tableau

40:42Next.

40:43in this

40:44context, especially with the data apps piece, it's hard to know

40:49because it says data 360, agent for

40:52SalesCoard, Slack, Tableau.

40:53So that I think they're talking about all of the SARS-Fource

40:56ecosystems.

40:57Not just Tableau.

40:59And that includes

41:00so instead of the agent

41:02using

41:03the rocky

41:04board is

41:06a little bit more.

41:06Let's just see

41:06keep skipping.

41:08Okay, there you go.

41:09Customer bit

41:10over.

41:11Libby

41:11now.

41:12Libby

41:12is one of my favorites.

41:13Libby is an awesome product manager at

41:16I've worked with her on a couple of things.

41:18I do believe I've seen what she's about to show here.

41:20Very early stages.

41:22She used to also work on Tableau

41:23Prep.

41:24I love my Tableau Prep.

41:25It's still the best analytics experience on the planet.

41:27I'll defend that

41:28Till the day I die.

41:30Libby was on the team that helped make that happen.

41:32Yep.

41:32Big fan of Libby.

41:33Let's see what she's got for us in this.

41:36As a product leader here at Tableau, what I love most about my job is the opportunity to listen and learn from all of you.

41:43Our customers.

41:44And what I'm incredibly excited to share with you today comes directly from those conversations.

41:49Capabilities that differentiate agentic analytics

41:53from traditional BI.

41:54Rock just showed you how today everyone can be a builder, from a business user to a data team.

42:00We're moving faster than ever with LLMs.

42:03And that's great

42:05for when you know what to look for, what question to ask, and when to ask it.

42:11But that burden, it's still on you.

42:15So let's go ahead and flip the script.

42:17What if, instead of looking for an insight, intelligence found you

42:23at the right moment

42:25in your flow of work.

42:27As a business user myself, I know that by the time I remember to ask a question, it probably means that I'm already falling behind.

42:35And that's exactly why Tableau

42:37is giving everyone

42:38an always-on

42:39analyst

42:40available 24-7

42:42for prompted or unprompted analytics.

42:46We're extending from just self-service

42:49to at your

42:50service.

42:53Today, I'm really excited to share

42:55with you.

42:56Oh, sorry, next slide, please.

42:58I'm excited to share with you

43:00Proactive Intelligence.

43:02We're bringing together the foundations with data alerts

43:07To bring truly autonomous insights.

43:09Okay.

43:10So while data alerts, they monitor the latest changes that you've told the system are important, things like the metrics that matter to you, the goals, the thresholds.

43:19Autonomous Insights goes one step further.

43:22Okay.

43:23An agent is continuously looking through my data, analyzing the most urgent and important changes.

43:30and delivering that to me

43:32in my flow of work.

43:33That's actually quite useful.

43:35If it's done well, this can actually be very nice, especially in industries where

43:40You blink and you miss the insight because the data is moving so fast that you don't even have time to react to it.

43:45I think of consumer goods as an example.

43:47You're doing promotion, it lasts for

43:49seven days, you're getting sales

43:51data sometimes daily

43:53But sometimes you're getting it in retrospect, so you run the promotion and you see what's happened a long time after.

43:58You don't sometimes get that fidelity.

44:00Having an agent that can bring the insight to you as soon as it's available, but then also tell you what is changing, so

44:07You don't need to look at the th

44:09you don't need to look at everything to know what's important.

44:11You can get

44:11like a first pass at it by an agent who

44:14's generally well

44:15reasonably trained to say

44:16Okay, this might be important to you.

44:18That's actually quite cool.

44:19Another thing is

44:20Tableau has a long heritage of let's say

44:22alerting.

44:22So you can build basic alerts, you set your thresholds, off you go.

44:26metrics and Tableau

44:27Pulse and then metrics prior to Tableau

44:29Pulse as well.

44:30Metrics also in the Tableau Next

44:32ecosystem, it's super confusing.

44:34So I don't know if we're still talking about Tableau

44:36next year.

44:36I think

44:36I feel like we actually are, but I think it's also true that this proactive intelligence probably spans both ecosystems.

44:43because

44:43I can see it working well.

44:45So we'll see more of the demo, but yeah it does actually say Tableau

44:48Next here on the thing.

44:49You say FluxWell Business.

44:50Actually

44:51something for you

44:52FluxWorld Salesforce

44:54My

44:54gut

44:54is this is in the Tableau

44:56Next ecosystem, but hey, let's see.

44:58Let's see.

45:00We're shifting

45:01from reacting to what happened

45:04to acting on what's important.

45:07So with that, how do you get started?

45:09Let's go back to the beginning.

45:10The foundation is familiar for most of us.

45:13It's just CRM.

45:14There's no add-ons, no bolt-ons, so let's see it in action with the demo.

45:21So here I am

45:22Tabet next

45:23looking at sales insights.

45:24Now

45:26we're gonna

45:26find it

45:27out of the box.

45:28I gotta

45:28keep getting

45:29my team's most important go to my team.

45:31Hold me out

45:31here.

45:31Come on

45:32This would have taken us so much time to bring things together

45:36out of the box.

45:37I don't know why I'm getting so excited.

45:39And it's not just sales.

45:40We also have it for service.

45:42Things like case resolution time.

45:44This is a data

45:44app.

45:44Team metrics.

45:45I think they're

45:46performance.

45:47I've not seen Taboy next in a while.

45:49But

45:51I'm going to

45:51ask a question with coworker.

45:53Now

45:53I just got back from some time ago.

45:55I'm overwhelmed.

45:56My pipeline has moved without me.

45:58So I'm just gonna go ahead and ask coworker

46:00What should I focus on?

46:02Instead of looking through dashboards or trying to look through specific accounts, Tableau

46:07is connected to coworker.

46:09It knows who I am, the accounts that I own, and it's analyzing for the most important changes and risks

46:17To ultimately surface opportunities for me to take over the finish line.

46:22Cool.

46:22So I just went from

46:24what did I miss?

46:26To something that I can really act on.

46:28And that's super powerful.

46:30So

46:31I can

46:31I will say this is actually quite useful.

46:34I wish this existed in the Tableau

46:36Core ecosystem as well.

46:38It makes sense in the Salesforce ecosystem that this exists as well.

46:41It's like a nice like a nice

46:43it's a nice way of sifting through Salesforce because Salesforce in itself is a pretty difficult

46:48platform to navigate your way through and once you've found your way through specific things you likely bookmark everything.

46:54Having something like this that can just go across the different data sets and just

46:58bring the proactive insights to you.

46:59It's a little bit cool.

47:00And it it likely becomes the sort of landing page for most people, especially if they've just got questions they really want to quickly answer.

47:06Without having to, I don't know, pull out the stops and weave your way through data apps and all of that stuff.

47:11Just go ahead and

47:13track everything I care about and bring it into

47:15one universe.

47:18I'm actually gonna stop here, 20

47:20through this, and let's see what happens

47:22Conversation within the opportunities page, or I can switch over to Slack to have a conversation with my team.

47:28So

47:29I can see that data app that

47:31Brock shared earlier.

47:32I'm gonna check that out, but I just received a notification

47:35Tableau is telling me that something needs my attention.

47:39There are three expansion opportunities

47:42that are highlighted for me

47:44These are accounts that have similar agent projection usage as Everpath

47:50systems, an account that we recently closed.

47:54So I want to take a pause here.

47:56Here is the important part.

47:59I did not ask for this analysis.

48:02Cool.

48:02I did not go through every account.

48:05I did not compare usage history.

48:08I did not build a model, I didn't have to ask my account team.

48:12Tableau

48:13was looking out for me

48:15An agent in the background

48:17brought together data from different sources, analyzed the different trends, ran

48:22deterministic prediction models and statistics,

48:26thinking about memory and personalization.

48:28This is not just a notification.

48:31It's a prioritized reason to act.

48:36And this is quite this is handy.

48:37In an ideal world, this is what all analytics looks like.

48:40Do nothing, have the right insights

48:42land on your plate.

48:43The key question is

48:45the thing I always struggle is, unless you have a live use case, this use case can sometimes feel far-fetched.

48:52And it only feels magic when it actually works

48:55when you need it.

48:55So for example, it tells you something you didn't know about the data that you

48:59didn't even conceive.

49:00And so

49:02The reason these kinds of things tend to fail when it comes to adoption

49:07is that learning time to figure out what's useful to someone.

49:12sometimes too long

49:13and by then the users

49:15sort of moved away and they've just gone back to doing things the way they do it.

49:18So really the success of something like this isn't the idea.

49:21The success around this is

49:23How good is this tool at figuring out what's important for you?

49:27And then

49:28it needs to do it

49:30however many thousands of times there are people in your organization.

49:33And if you think

49:35about what

49:36that takes behind the scenes.

49:38That is

49:39quite an intensive application, at least in the sort of constructs of how I understand AI to work today, especially assuming that you've set up all the semantic models correctly.

49:50So many things have to go right.

49:52for this sweet spot to exist.

49:54Which is why this can

49:55feel for

49:56fetched.

49:56I'm not dissing the idea.

49:57The idea is absolutely perfect.

49:58And in a logical way, Tableau

50:00Pulse does this already.

50:02It's just taking it to the next step.

50:04You can see how stringing together several innovations in the Tableau

50:07ecosystem has led to this.

50:10But everything I just said there is still valid.

50:12Every single piece has to work like clockwork.

50:15The analysis it chooses, the things it spots, the way it tells the story, the way it drills into it, then the way it presents it to the user, then where it sends the message, then where the person takes action.

50:24And then when they come back, do they see that message

50:26was the data moved on?

50:27And so they can't actually see the thing you alerted them.

50:29And if they see that, okay, you've gained one trust point.

50:33And then you go again.

50:34And it's only when you get 10 or

50:36yeah

50:37A number of those consecutive hit rates does the user then go, oh, this is actually really useful.

50:42Okay.

50:43Yeah, that's with some context, I think.

50:45For

50:45why features like this

50:47Can struggle.

50:48So from here I can go in and dig deeper into the why, or I can go ahead and just take a follow-up

50:54action.

50:55The agent has a playbook that I can kick off in just one click.

51:00This would have taken me so much time to do manual effort and connect the dots

51:04completed for me, just like that.

51:07And it doesn't stop

51:09with just these three accounts.

51:11Tableau

51:12is suggesting I take this one insight and turn it into an always

51:16on advantage.

51:17Okay.

51:17Saying, hey, do you want me to keep checking for these types of opportunities in the future and send it to your team?

51:23Yes, please.

51:25The value isn't only that I moved faster today, it's that the system is finding opportunities for my whole team for tomorrow.

51:34Yeah, that's good.

51:35So this same loop, it's available wherever your users work.

51:39So

51:40if you work in Microsoft Teams, Tableau

51:43will not wait for you to ask the question.

51:46We'll deliver the insight at the right time

51:49with the recommended action.

51:51The same thing goes for ChatGPT.

51:55And the same thing goes for Claude.

51:58That's good.

51:59That interface.

52:00I do

52:01one

52:02theme of this whole showcase so far.

52:05And it goes back to a topic that I have not quite got my

52:09head and hands around, which is

52:11this concept of where

52:13is the surface of work?

52:15For a long time in analytics, we've done our work where the tools are.

52:19So you go to the Tableau

52:20platform, you do your stuff there, you get the insight there, then you go to Slack and you send the message, then you go to Word and prepare the presentation.

52:27we've been bouncing between tools to do our job.

52:30I think the future looks very different, especially AI, because AI seemed to for some reason

52:35command

52:36all other tools to support to it.

52:38So you have MCPs, you have the ability to call, bring your own LLM as well in some cases.

52:44All of these things, if I take Claudas and examples

52:46typically

52:47The official term for this kind of capability is called a harness.

52:50So the harness is the thing that sits on top of the LLM that then allows it to do stuff.

52:54Clawed code is a harness that specifically allows you to focus on coding.

52:58You might have had of other coding tools, codecs

53:01as well.

53:02What I'm essentially talking about here is

53:04what is the surface of work and it seems to me like there is this evolution to move towards being able to support doing work inside of an LLM tool

53:14exclusively

53:15because it's almost like

53:16the interface is disappearing from lots of applications.

53:19Which I think is bad because then you lose fidelity with that simplicity.

53:22But then on the same token

53:24I do not think teams complete their workflows in these tools.

53:28They still have to go and do things in the real world.

53:30They still have to go and talk to people and stuff like that.

53:32But I think it's a really interesting question

53:34where you see here the multimodality of the features they're shipping.

53:37They're nearly all available in

53:39Claude, Team, Slack, whatever.

53:41And that's a really important architectural decision because it's an acknowledgement that, yeah, it

53:46people are not coming to the Tableau

53:47platform anymore to get their insight.

53:49And so if we want to be successful, we actually have to go to the places they work.

53:53I think that's a topic that's not often talked about.

53:56But look at many other tools.

53:57They're actually not as good at this.

53:59And

53:59when you

54:00really drill into this, even

54:01tools that I'm quite fond of recently, they're still not quite there.

54:04So this is where Tableau's extensibility can stand out really well.

54:07And I like it.

54:08I think it's a nice, it's a nice

54:10it's a nice philosophy to have across all features.

54:12Whenever we ship a feature, let's make sure that we can ship it, not just in Tableau.

54:17But in every surface of work that a user's like to be in, whether it's team, our own infrastructure, Slack, an API, whatever it is, let's make sure we can get the feature there.

54:27It might change, but that trusted intelligence, that proactive experience

54:33will meet your team where they work.

54:44So

54:44that was the proactive intelligence loop.

54:47Where an agent is continuously observing the changes in my data.

54:51It's analyzing for the latest trends and the most important things that it ultimately needs to recommend an action to me

55:00This is the difference between agentic analytics

55:03and traditional BI.

55:06Analytics

55:07stops just being a place that you go to

55:10And it starts to become something that's truly working for you.

55:15So with that, I want to welcome to the stage a customer that's already changing the way that they run their business with agentic analytics.

55:21Let's see.

55:22Welcome to the stage, Tableau

55:23Chief Product Officer

55:25Suther Jones, and from MasterCard, Sabel Bresnik.

55:28Okay, so

55:28it's MasterCard.

55:30Sabelle and Southard.

55:31So

55:32the Chief Product Officer.

55:33I say that because the CMA has changed, but Southard is fairly renewed to the Chief Product Officer

55:36role.

55:37Thank you so much for coming.

55:39And thank you and welcome the Dream Force.

55:42It's been a pleasure to work with you and get to know you over the last couple of years, actually.

55:46So let's start.

55:47Tell us a little about your journey.

55:48You started with CRM Analytics

55:49and you're moving towards Tableau Next.

55:51Maybe you can share with the audience how that journey has been for you.

55:54CRM Analytics is our single source of truth for sales and we adopted CRMA back when it was introduced as Wave Analytics in 2014.

56:04And I'm sure some of the Dreamforce veterans in this room are now

56:08remembering and smiling.

56:10Who remembers those surfboards?

56:12Yes.

56:15Over the years we

56:16built a lot of CRMA dashboards to support sales, but what ended up happening was our users were

56:22The reason I'm listening to this is I'm

56:24I'm trying to figure out how they're gonna pivot from C RMA to what we've just seen.

56:27So

56:27I'm quite curious what the customer story is here.

56:30I'm not gonna critique it.

56:30I'm just gonna listen to it.

56:39Just an answer to your question.

56:40So that pushed us really hard to look at the conversational and

56:44that's where Tablon

56:45X changed the game for us.

56:47Last November we conducted a POC and we landed on Agent

56:51Force

56:51and TabloNix

56:52as our agent

56:53analytics platform.

56:55What made it click was having an agentic

56:58tool

56:58Right inside Salesforce, in

57:00sellers

57:01flow of work, built in a semantic data model, and tailored to MasterCard business preferences through Data 360.

57:08By mid-February, we had 100 users live.

57:11Now we are scaling to 1,000 users

57:15with the new use cases on the way.

57:17Wow.

57:18A thousand users.

57:19Not bad.

57:19That's amazing.

57:20I'm gonna

57:22see if we can

57:23journey.

57:25No, I'm here for the products.

57:26I just want to see the next one.

57:37See if we can skip ahead

57:39No

57:40decks.

57:40Okay, here

57:41we go.

57:42So I think

57:43it's

57:43carrying here.

57:48Here we go, right back.

57:49But many of you are probably wondering what a lot of folks and that I speak to, the CIOs I speak to are wondering, do I really have that trusted data foundation?

57:58And Mark talked about that earlier.

58:00So

58:00I have really good news for all of you.

58:03All of you have already built that.

58:05You are three-quarters of the way there to having that foundation.

58:08And we're going to get you all the way there in the next five minutes.

58:11So let me explain.

58:14So

58:14when we look at

58:15a Tableau

58:16dashboard or

58:17CRMA dashboard

58:20and we make an insight, we see the metric, that means something very specific to us and our business.

58:28And then we map that to the know-how we have about our business, how our business is structured, how our organization runs, what markets we're in.

58:35And that means that we as humans have implicit knowledge.

58:40That is something that a regular agent doesn't have.

58:43A regular agent can look at your trusted semantics, but doesn't have all the knowledge that you have

58:48So let's make it real.

58:49ROI, return on investment.

58:51It's a metric we all have in our business, but each of in the audience have a very unique way of defining it.

58:56And you know, when you look at that ROI and you compare it for a particular product in a certain stage of its life cycle, you think about it differently than if you think about a product late in its life cycle.

59:06Or when you look at a campaign in one business unit, you might evaluate that ROI value differently than a campaign in a different business unit.

59:13That's your business know-how.

59:15That is what you need to couple with that trusted semantics that you've already built.

59:21And that is why we build Tableau Knowledge.

59:23Nothing to add there.

59:24Tableau

59:25knowledge

59:26is

59:26your

59:26one

59:28correct, accurate,

59:30governed, trusted.

59:35I can't tell whether this is in the Salesforce ecosystem or not.

59:38But before I see the feature, let me have a riff.

59:41Let me have a go at this.

59:42I think this exists in both Tableau Desktop and

59:45Tableau

59:46Next.

59:46Tableau

59:47Core, sorry, in Tableau

59:48Next.

59:49In Tableau

59:49Core, it's basically the metadata API, the REST API, and a bunch of other capabilities,

59:55the VisQL

59:56data engine.

59:57all coming together to build you

59:58give you this perspective.

59:59The thing that builds them together is probably the innovation here.

60:02How do you stitch all the constituent parts together to be able to tell a story?

60:06This sort of graph thing, we saw it at a Tableau conference, it's come back here in a new interface inside a Tableau

60:11Studio.

60:11kind of

60:12useful because it's basically just a semantic map, semantic sort of journey, and it's here inside a studio, which is a a good hint of where it's going to be focused.

60:20In terms of the Tableau

60:22Next

60:22ecosystem, this is just naturally the semantic capabilities in the Salesforce ecosystem.

60:26Except for

60:27Again, you are adding that layer of, let's say, intelligence on top of it.

60:31So whether it's in Salesforce or whether it's in the Tableau

60:34Core ecosystem, it's essentially the same capability.

60:36It can be seen the same way almost exactly

60:39because

60:39Everything starts with a data source, some sort of model, and then some sort of output.

60:43So the same map can apply in pretty much every instance.

60:45So

60:46now that's my sort of take on it.

60:47Let's see what it's about to be.

60:48It does say Tableau

60:49Cloud and Tableau

60:50Server general availability basically

60:53now, which is pretty cool.

60:54So let's have a look.

60:56Foundation for knowledge and

60:57organization.

60:58How does it work?

60:59So as

61:00Mark talked about this a little bit earlier, when you try out and talk to all your semantic models, your tableau, published data sources, data inside of serum analytics, even metrics that might be inside of DBT or other semantic areas.

61:11We automatically create a graph, a knowledge graph on that.

61:14But we don't leave it at there.

61:16We also reach out to all the knowledge

61:18that's in your know-how.

61:19It's

61:19since you have your Google Docs, your PDFs.

61:22your

61:22data catalogs, your glossaries, and we bring that in also, and we automatically create that graph.

61:27Once again, automatically.

61:29You are not manually building this graph.

61:31We're building it for you.

61:33Across all of your assets.

61:35Now you may be wondering

61:36what

61:36the other CIOs I talked to

61:37wonder

61:38and that is

61:39wait a minute.

61:39I'm not so sure I believe you.

61:41That sounds pretty cool, but eh, not so sure.

61:44Let me show it to you next.

61:46I see that.

61:46Let's do the demo.

61:48Alright, so where am I gonna start?

61:49We're gonna start where your semantics models are.

61:51All of you have already built that.

61:52By the way, you should give yourself a pat on the back

61:54because

61:54You've built the hardest part.

61:56That is the trust.

61:57So let's just start here with a a

61:59a couple of contextual things.

62:00So collections are something that existed in Tableau Cloud for some time.

62:03I'm not sure how well they've been used.

62:05But they are an equivalent of a project.

62:07A collection can span multiple projects essentially.

62:10You can collect different things from different places, workbooks, prep flows, data sources.

62:14And they come in a collection.

62:15It looks like they've repurposed it specifically for this purpose because it's a much nicer way of cutting the crap of the tableau.

62:22cloud server like hierarchy, which can be a little bit restrictive, especially if you're new to this sort of CMS layout for data.

62:29So collections just mean, hey, this is my universe, marketing and analytics.

62:33If they're actually done and curated well, they absolutely do exactly that.

62:37You go to these collections, you get the right access.

62:39You still need individual permissions and access to see the relevant data.

62:42But if you've got row level security, that shouldn't really be a big problem if you've done that right.

62:46So again, it's one of those things that done right, it can work really well.

62:50But again, most people probably don't spend the time setting them up.

62:54Semantic models, your metrics.

62:55I if I'm in a Salesforce customer, I have my sales metrics, my service metrics, my marketing metrics, they

63:01come pre-built for you.

63:02Okay.

63:03That's what Libby

63:03mentioned earlier.

63:04I also have other metrics like HR and finance and other things in my semantic models.

63:09But where's the graph?

63:11Just click on the button.

63:12That's your graph.

63:15Automatically built for you across all of your semantics.

63:19Is

63:20this useful?

63:22Okay.

63:22It's a bit it reminds me of Obsidian.

63:24I don't know if Obsidian is a no-taker.

63:26It allows you to take notes, but it allows you to link your notes.

63:29And the reason it's typically uh

63:31used in certain fields, especially research, is the ability to cross-reference notes is actually quite handy, especially if you're building

63:37some

63:37sort of idea, some sort of research, and then it creates this sort of visual diagram to show you what your notes look like in this sort of thing.

63:44And it looks cool.

63:45Again, like I said earlier, it looks cool, bro.

63:47But what does this actually enable?

63:49So this is what I want to see next.

63:50Okay, you've given me this visual interface.

63:52It's a cool thing to see.

63:53Good bit of marketing.

63:54Okay.

63:55What does it actually help me solve?

63:58We're connecting the relationships.

64:00We're showing how each one relates to its different spots, what

64:03dashboards or workbooks it depends on, what metrics are available, the

64:07specific business preferences that are in there.

64:10The specific way I

64:11calculate different metrics, even odd ones like historical velocity, there's the description, that's the sequel, that's the data source that comes from.

64:20One place where I can get all of that.

64:22We're not building this.

64:23You're not building it.

64:24We are extracting it from all the work you've already done, leveraging the work you've already done.

64:30But I said something else.

64:31I said

64:31that's not enough.

64:32We need other data sources.

64:34We need know-how.

64:35So let's connect to them.

64:36We can connect to a number of other sources.

64:38Maybe it's my Google Drive, maybe it's Informatica's data catalog, maybe it's Data 360, maybe it's my Teams

64:45or Slack conversations, my emails, my glossaries, my training manuals.

64:49In this case, I'm going to grab a training manual on how we do renewals and expansions.

64:53It's a PDF.

64:55Unstructured documents.

64:56That'd be really valuable for an agent to use when answering questions.

65:00So

65:00let's connect to it, load it up,

65:02and that is now going to enhance the knowledge in that graph.

65:06I have just improved the accuracy of my graph.

65:09That's pretty awesome.

65:11Wait a minute, do I have to do that all the time?

65:12Does a human have to manage that graph?

65:15No, you don't have to.

65:16We actually built it so that the graph can help manage for you.

65:20This is our homepage for our graph.

65:22I'm talking about the health of our graph.

65:24It's 88%.

65:25That's pretty good.

65:26My kids think 88 isn't good by the way.

65:29So this is interesting, funny

65:31anecdote

65:31about your kids.

65:32If they think 88% isn't good, then

65:34you need to help them lower the standards.

65:36But anyway, Tableau

65:38knowledge.

65:38This is an interesting way at approaching the semantics problem.

65:42This is what fundamentally is.

65:44The graph is like the analyst

65:46sort of Fisher

65:47Price

65:48version

65:48of

65:49hey

65:50what

65:50is your semantic model

65:52in

65:53your ecosystem?

65:54And it probably tries to draw relationships across things that maybe shouldn't be related.

65:58And so I still

65:59I still think it does need an element of curation and a good steward of a platform would do this curation.

66:04Now you can leave it to it, but there are going to be unique instances in your business where you definitely want to make sure you curate these examples

66:12It's an interesting premise, but I was

66:14put caution on that element of letting it do all of the work.

66:16Now, like all AI assisted tools, it should be doing the majority of the work, taking you to that point where actually it's now worth spending the time doing the work because it's done the hardest bits for you.

66:26essentially discovering those relationships that are incorrect, porting those to the correct setup, and then giving you a visual interface to essentially curate and manage that.

66:34I think that's a really good philosophy again

66:36for how to go about the problem of solving this.

66:39Other platforms, especially the data warehouses, still have a really, let's say, archaic approach to this perspective.

66:45where

66:45yes they have really powerful features but they're heavily landlocked behind interfaces that are predominantly driven by people who are very comfortable with highly technical tools.

66:54Which means the everyday analyst

66:56would never go in there.

66:57And at the end of the day, in small organizations and large ones, it's the everyday analyst

67:01that's being asked to do this stuff, not the admins, not the data engineers, not the server admins.

67:05And so

67:06you have to strike a balance

67:09I

67:09don't know what's going on in this world, but that's not good enough anymore.

67:11But 88

67:12's pretty good, I think.

67:12We can make it better.

67:13How do we make it better?

67:14The agent's gonna tell you

67:16How to make it better?

67:17Do I have conflicts on my metrics?

67:19Do I have places where descriptions are lacking?

67:22Do I have places where there's broken metrics?

67:24What

67:24a broken metric.

67:25The metric is different today than it was yesterday.

67:28So I can click in there and the agent's gonna tell me how to fix it.

67:31Wait, the agent's gonna tell me how to fix it.

67:32It's gonna suggest something.

67:33It's gonna say, hey, listen, net

67:34ARR

67:36It looks like the underlying source system changed.

67:38Someone made an update to a column.

67:40Maybe they changed the field tape, maybe they did something different.

67:42Tell me how I should fix it.

67:45I can accept that fix.

67:47And with that, I have now fixed my knowledge.

67:52Again, this is pretty cool.

67:53I like this.

67:54I like this.

67:55I really like this.

67:56I still worry whether we're inside a Celsource desktop.

67:59Am I is this the biggest bait?

68:01Is

68:02Tableau

68:02Studio

68:03The desktop that I know, if that makes sense, just connected to Salesforce.

68:09I don't know.

68:09Yeah, but I didn't just

68:10fix my knowledge graph.

68:11I also fixed my semantic layer

68:14So

68:14the agent is actually helping you manage your knowledge graph as well as your underlying trusted semantics.

68:19And that is how I get trusted knowledge

68:21on which we can build all those wonderful experiences that you saw earlier from Brock

68:26and Libby.

68:28That's the secret to having that trusted data foundational, which

68:31you can build a trusted agentic enterprise.

68:33Now

68:34with that

68:35We're going to move back to another customer.

68:36We're going to talk about how IBM has levered their trusted data foundation.

68:41Please welcome Michael Peeler, head of GoToMarket Analytics and Engineering at IBM.

68:49Michael, welcome.

68:51Thanks for being a Dreamforce.

68:52Thanks for representing the great state of New Jersey.

68:54Give me

68:55some.

68:55Yeah

68:56Anybody from New Jersey?

68:57Woo!

68:58All right.

68:59Mike and I live like 20 minutes apart.

69:01So

69:01at IBM, you guys also, similar to Sabelle, started with CRM

69:04Analytics.

69:05And are on this journey

69:06to Tableau

69:06Next.

69:07I know your journey is a little bit different than MasterCard, but maybe you can share a little bit

69:09about your journey as well.

69:13Sales cloud.

69:14Am I being baited?

69:14That is absolutely the thumb.

69:17Am I being baited

69:18ports that were used for sales clouds with tablets CDA?

69:23Skip ahead.

69:26Another cool stat to look at, pretty cool, interesting

69:29sort of adoption here.

69:30Again, for customers like this, this makes total sense.

69:33Of course you'd use Tableau Next.

69:35It's just innate

69:35uh

69:36instinctively

69:37where you're already doing most of your work.

69:39So if it works for you, why would you change?

69:41It's a very easy win.

69:43You can update dashboard

69:44fifteen minutes.

69:45Even with that, you know

69:45someone

69:47goes for the key market.

69:48Um so

69:48we look like a checker, you have to run it up.

70:01Okay, we're back to the end.

70:02So now we're closing the sessions, just

70:04another two minutes, and that's pretty much the end of the keynote.

70:06So

70:06let's

70:06see back.

70:13And

70:14our incredible presenters, Brock

70:16Southern

70:17in Libby.

70:19Okay, so

70:21you just saw it.

70:22That's how you take your trusted data foundation and the data-driven enterprise to become an agentic analytics

70:30enterprise in three steps.

70:32You empower everyone to be a builder with Tableau Data Apps and Tableau

70:36Studio and Skills.

70:38You have intelligence come

70:40to you where you're working.

70:42So

70:42you can drive actions and decisions with proactive intelligence and headless analytics, and you build it all on your trusted

70:51foundation with Tableau

70:53knowledge.

70:54And our customers are doing this today at scale.

70:57Organizations of all industries and sizes, 97%

71:02of the Fortune 100

71:04are trusting Tableau.

71:06To transform them from a data-driven enterprise to an agentic

71:10analytics

71:10enterprise.

71:11I will say that

71:13for a lot of these companies, Tableau is not the only technology that's there.

71:19It's just the case.

71:19It's just true.

71:21So

71:22strictly speaking, I would love to know how many of these people exclusively rely on tablets.

71:30Probably a much smaller percentage.

71:31And I'd also love to know how many of these companies use another business intelligence tool.

71:36Of course

71:37Salesforce would never really show that here, but it's just a really important detail to show.

71:41For example, I know four companies on here.

71:44that I know

71:45use

71:46in

71:46balance

71:47Tableau

71:47and something else.

71:48So

71:48some orgs

71:49prefer something else, some orgs prefer something else.

71:51I can't really say that because I'm

71:53supposed to share this information publicly

71:55But they're sharing here that they will use Tableau, which is good.

71:58But the other question is

71:59what else do they use and why do they use it and where does it

72:02where is it successful?

72:03Also, do they mean Tableau

72:05Salesforce, Tableau

72:06Next?

72:06Do they you know there's so much more context here and so

72:09You can cast this

72:10net

72:10super wide because they could be using Tableau

72:12Core, but not Tableau

72:13Next, which is what we just saw.

72:14They could only be using Tableau

72:16Next, which is part of Salesforce, but not Tableau Core.

72:18So it's like, ah

72:19I think this 97

72:20%

72:20is very easy to arrive at because generally speaking, it's just one big sort of melting pot, as my wife likes to say, of

72:28some

72:28parts of the Tableau

72:29ecosystem but not what you and I would think is like

72:32the complete setup, if that makes sense.

72:34So

72:36Always read these numbers cautiously.

72:38There's no validation style or something like that.

72:40There needs like a definition.

72:42Use tableau.

72:43Tableau

72:43products.

72:44Tableau.

72:44Use a tableau

72:45brand.

72:46Who knows?

72:46Who knows?

72:47Anyway, let's carry on.

72:48So what can you go do now?

72:50Guess what?

72:52You can go build right now

72:54here at Dreamforce.

72:56Go to the campground, go to the forest, get

72:58started with our pre-built apps.

73:01Start building on Tableau

73:03with Tableau Data Apps with MCP and Headless Analytics

73:07and join our Slack community.

73:09As we drive every organization from a data-driven enterprise to an agentic

73:14analytics

73:14enterprise.

73:15And I will finish this the same way I started.

73:18Thank you.

73:18Thank you.

73:19Thank you to our customers.

73:21Thank you to our partners.

73:22Thank you to our community.

73:23Thank you to the Trailblazers.

73:25And thank you to the Data

73:26Fan.

73:27Let's have a great dream

73:28for us.

73:28We'll see you at the concert.

73:31Very cool, very cool.

73:32And I'll take an extension and a note from that and I'll say thank you for watching this video.

73:37Thanks for joining me.

73:38gonna get to dip in again.

73:39Interesting keynote, interesting keynote.

73:41I think what I expected from Dreamforce from Tableau.

73:43I think they're saving their powder for next year.

73:45I think we'll really see the Tableau

73:47Studio thing come to fall properly in Tableau

73:51Core

73:52I didn't quite see

73:55sort of the full vision if that makes sense.

73:57And it's again, it's getting harder and harder to separate between

74:00Tableau Core on Tableau Next.

74:02There really shouldn't be any difference.

74:03Like

74:04obviously

74:04the future of this

74:05of the all of this stuff is that you just don't know what you're using.

74:08You're just in Tableau

74:09full

74:09stop.

74:10And I actually think that's a good product direction.

74:11You shouldn't have to know

74:12whether you're in desktop, prep, server or cloud, as far as you're concerned, you're just in Tableau.

74:18One product, many experiences.

74:20In fact, I've done a video about that.

74:22I'll put it

74:22up again.

74:23Go ahead and check that out.

74:24So I do like the step towards that direction.

74:28I want to make a video that to follow up this one.

74:30I probably won't follow this one.

74:31It'll probably come one or two videos after this one.

74:33And this is about

74:36product

74:36value.

74:36There's a lot of what we saw today that's about product marketing in a very different way than what we typically expect.

74:42Marketing to me is

74:43hey we've built this feature, let's go tell people how cool it is.

74:46What I actually think sometimes goes on at Salesforce

74:48is actually product marketing and leads the agenda of what to build next.

74:52And sometimes that's not intuitive

74:53because you'd think that would come from the

74:55product managers and

74:57developers understanding users' needs and going off there.

75:00But actually it's the product marketing team that seem to drive these

75:04understandings, feed that back to product managers who then build features to meet that demand that is being identified by product marketing.

75:11What can sometimes then

75:13happen

75:15is

75:15what I've previously called a disconnect, a lack of intuition,

75:18many words I've used in the past.

75:20Between like

75:21what users want

75:22and what's actually gets delivered.

75:24So you get the genuine sense of what users want, but you deliver it in the place they didn't expect.

75:28Happens a lot with Salesforce and Tableau

75:30recently.

75:31They put a lot of the stuff you want inside of Salesforce, but not where all the analysts are, in my eyes, in Tableau

75:36core.

75:37But

75:38it creates this gap between what I call perceived value and realized value.

75:42And I think this is the core reason why people

75:45who

75:46are evaluating other tools and moving away from Tableau

75:49are moving away

75:50and often

75:52quoting the price

75:53when in fact it's actually what's causing this is this perceived value gap.

75:58Okay.

75:58So

75:59yeah

76:00Video scene on that.

76:01Thanks for watching.

76:02Uh

76:03take care and I'll see you in the next one.