Meet Golden Analytics, an AI-Native BI tool: With Francois Ajenstat

0:00Francois, how are you doing?

0:01Hey Tim, good to see you.

0:03Absolutely.

0:04It's it's been an exciting couple of months.

0:06I think we've been talking about quite a few things.

0:08But no, I'm really excited to have you on the channel today.

0:10I think you reached out to me, I think it was now

0:14three months ago.

0:16And I actually I reached out to you.

0:17Let's correct the f let's correct the let's correct the facts.

0:20I felt guilty that we hadn't had a chance to talk whilst you were at Tableau and I was on the channel and I thought to hell with it.

0:27Let's talk about what you're doing now.

0:29And you reached back to me and said, Hey, I've got a better idea.

0:32Let's talk about something different.

0:33I was like, okay, interesting.

0:36And then you re and actually you chased me because I was pretty bad at being, let's say, prompt.

0:42And you held me accountable.

0:43And then yeah, one thing led to another.

0:45And today you're here to talk about something pretty exciting.

0:48Thank you.

0:48It's exciting.

0:49I've been watching your channel since you started.

0:52And obviously I've known you for a long time, and you always have a way of getting to the core of the issue, seeing things the way others maybe don't see it.

1:03Right.

1:03And the opportunity, just talk about what we're seeing in the market, how there's new opportunities out there, I think is exciting.

1:12I'm really delighted to be on the channel for the first time with just Tim.

1:17You got it right.

1:18I think you are the first guest and video I have

1:22recorded since that announcement.

1:24So that you like obviously contextually, I don't know how the timings work with these launches, but yeah, yeah.

1:29That's pretty wild.

1:30I'm glad you reminded me of that.

1:32Because I just it never occurred to me just sitting here.

1:35So

1:36I really appreciate that.

1:37Yeah, so I guess it might help for you maybe to describe what golden analytics is, if that makes sense.

1:44What is golden analytics?

1:46That's a good question.

1:48Maybe first for the audience.

1:49I don't know if every everybody knows who I am and what I've been, but I've been in the data space for 30 years.

1:56Yeah.

1:56I I started my career at Cognos and call it the first generation of BI, and I fell in love with data.

2:03I fell in love with the power of data.

2:06And I got to continue that at Microsoft and the SQL Server team and the Office team.

2:10And then I found this little company in Seattle called Tableau.

2:14which nobody knew who Tableau was at the in the in the day.

2:18And it was just an incredible journey to be part of this little startup that

2:23Changed an industry and created a movement.

2:25And I was there for 13 years.

2:27Truly remarkable times.

2:29I then went to a company called Amplitude, which is really focused on helping people build better products using data.

2:36Again, just another application of data.

2:39And when I finished my time at amplitude, I started thinking about what was next.

2:45And I was retooling myself with AI tools using cursor, using Clog Code, and I got inspired.

2:53I saw how these tools are giving people this opportunity to be 10x, 100x better at what they were doing.

3:01It's not replacing developers.

3:03It was empowering developers.

3:05And I was like, oh, what is the equivalent for data?

3:10Is there a cursor?

3:11Is there a clawed code for data?

3:13And the answer was no.

3:15And I was frustrated with the state of our industry.

3:19I got frustrated that a lot of the tools that we know and love were trying to bolt on AI and it felt more like a marketing gimmick.

3:29Than truly transformational.

3:33So I started thinking, what would it look like if you built

3:38you know, call it a cursor for data, but a solution that uses the AI in a native way that empowers people, not replaces people, but empowers people to do their best work, to be 10x, 100x better.

3:52Right, because they have these capabilities.

3:55And so I started building, I started tinkering away, and all of a sudden the light bulb came up.

4:01And that was the genesis of Golden, uh, or Golden Analytics is the official name.

4:08But it was actually a step even further.

4:10You could use AI to automate analysis.

4:13But what if you could also add a Canva for data?

4:16And I put it in quotes just to create an ill illustration, but the idea that a lot of the time that you spend in the tools that you use today is actually spent formatting the data

4:27presenting it in a certain way.

4:29That's not the deep analysis work.

4:31It's beautiful.

4:32It's important to communicate.

4:33But what if you could make that simple as well?

4:36Yeah.

4:37And all of a sudden, you start having the brains of tools like Tableau, the beauty of a Canva, and then the empowerment of a cursor or Claude all combined together.

4:50And that's golden.

4:53It's a nice play on words as well.

4:55Yes it is.

5:00I have so many golden puns it's not even funny anymore.

5:03My team hates me.

5:06And now time for the golden hour.

5:09And you get a golden ticket.

5:11Oh, no end.

5:12No end.

5:13No end.

5:16It's amazing.

5:17I will say I love the nod to uh gesture that you've always done at past conferences and that's characteristic to the Lego.

5:25So I personally wanted to call that out because I felt like that was a very

5:28Nice touch from you because obviously, yeah, it's your little project and it's incredible.

5:32Yeah, thank you.

5:33Thank you.

5:34It's I think of this as I want it to be a gift to the community.

5:38Right.

5:38I want this to be a place where the community feels like

5:42this is a solution for them, empowers them, unleashes them, and brings us back to the height of

5:51the data community that we know and love.

5:53Yeah.

5:53And when I always do these, it's like it's rock on.

5:57It's for us.

5:58Like, yes, data is fun and we're going to have fun with it.

6:02Yeah, amazing.

6:03I know typically when I do these videos I go into a bit more context and background, but I think for this video we should look at the product and then we can talk a lot more freely about

6:13the challenges, the thinking, and where you're going.

6:15So yeah, over to you to show us the product.

6:17All right, let's see it in action.

6:19So here we are in Golden.

6:21And with every data tool, you have to start with data.

6:25So obviously you can connect to files, CSVs, parquet files, etc.

6:29You can connect to data on your Google Sheets.

6:31We can connect live to your cloud data warehouses that are out there.

6:35We have saved data sets.

6:37We even have samples because everybody needs a good sample.

6:40So I thought for today we'd use something that a lot of people in the community are probably very familiar with, and that is a data set from a fictional store called Superstore.

6:52You've heard of it before, Tim?

6:53Yeah, I think I have just a few hundred times, yeah.

6:57So we'll just use superstores so we're all familiar with it.

7:00But the first thing that happens in Golden is you don't end up in a blank canvas

7:05You end up in what I call the Discover page.

7:08This helps you understand your data.

7:11What is this data about?

7:13So you can use Superstore to analyze sales performance by customer segment.

7:17Great, it's helpful.

7:18I found that over the years, most people are intimidated by data.

7:22They don't know what questions to answer, to ask.

7:25So we put the questions right here front and center.

7:29So you can see I can ask questions about profitability and discount correlation, regional sales performance trends, product category.

7:36Like that's really helpful.

7:37And at every point I could go and create that dashboard.

7:40We also provide quick insights right out the bat.

7:43You don't have to do any work, and we pull out some insights for you to be aware of.

7:48For instance, did you know that Office Supplies and Superstore leads with 60% of the records?

7:53That's interesting.

7:54The corporate segment account is 30% of the orders?

7:57Right?

7:58You don't have to do work.

7:59We're doing it for you to help you figure out and where to focus.

8:03And of course, we give you suggested visualizations, different ways that you can visualize that data.

8:09Really simple.

8:10On the right hand side, you can see your data profile.

8:12So we have nine dimensions, five measures, two dates, and five geographic fields.

8:17So really quickly you're oriented on that data.

8:21Now, in this case, let's say we love these suggested visualizations.

8:25We can select them, we can go analyze them, but in my case, I'm going to go create the dashboard.

8:31Boom, dashboard created

8:34That's it.

8:35Fully interactive.

8:37I can engage with it.

8:38If it's not exactly what I wanted, maybe I could say add category to sales trend.

8:46And so I'll type that in.

8:48And now the agent just added that in.

8:51Great.

8:51I didn't have to figure out how to use the interface.

8:54But of course, if you want to modify it yourself, you're in control.

8:59One of the core principles that I came up with this product is what I call a slider of autonomy, which means that you can dial up or down the amount of AI that's available to you.

9:11You.

9:11It's always there in every screen at every step of the way, but it's there to help you.

9:17But if you don't like what the AI did, you can change it.

9:20If you want to do it all manually, where you want to go ahead and create your own chart.

9:26You can just do that as well.

9:28You are always in control.

9:31Now, not only that, we've also brought in AI across the experience.

9:37So let me walk you through what that means.

9:39Through the product, we started in discover, right?

9:42So here you discover the data.

9:44But there's the three core jobs to be done.

9:47You prepare your data.

9:49You analyze your data and you communicate the data.

9:52Pretty simple.

9:53We're all data people and this is what we do all the time

9:56So the way that this interface is designed is you always have a canvas in the middle, and then you have an agent on the side that's there to help you.

10:04But when we talk about the slider of autonomy concept, it means that you're always in control.

10:09You can choose to maybe change standard class to just Tim.

10:15Boom, you're done.

10:17If I wanted to make this uppercase, I can manually do that.

10:22Of course I could ask the agent to do that, but it's a lot faster to just do it yourself.

10:26Now every time you do an operation, we keep track of that operation.

10:31So you always have that confidence.

10:33Oh, we went from standard class to just Tim.

10:35And I could go back and back.

10:37It's almost like having a time machine for your data.

10:40So you're able to flow in the data and help things move forward, but you always have that visibility.

10:48Now, if we go a little bit further, maybe I have here this field that has customer name.

10:53I'd love to have first and last name.

10:56What's the calculation for that?

10:58I don't know.

10:59I can just say split into first and last.

11:04And now this is an example of where the AI is built in at every step.

11:08It knows that this is a name and it suggests how to go clean that up.

11:13Amazing.

11:14Easy.

11:16Yeah.

11:16I didn't have to learn a calculation.

11:18I didn't have to learn anything.

11:20It just did it for me.

11:22But we can go further.

11:24So let's say we have state and province over here, and I want to enrich it.

11:29I'd love to get additional data that wasn't in the initial data set.

11:34For instance, can I get the state capital of that state or the population?

11:39Let's do that because we want to normalize by population.

11:42Yeah.

11:42We'll now use our APIs and go to the internet, pull down the data, and

11:48Sometimes it doesn't get it quite right.

11:51But you get the idea of being able to bring these things together.

11:55So that's the example.

11:57Of course, you can also go in and you see your data dictionary.

12:00This is where you can go make all of your changes, clean this up and say this is an order.

12:07Right?

12:07It's everything is available at click away.

12:10Yeah.

12:11So that's the core analysis, the data prep side.

12:15Obviously, the agent can do a lot of different things.

12:18One of my favorite little parts of the agent, for instance, is suggesting new fields to use.

12:23When you click on that, it looks at your data set and says, oh, you might want to create a profit margin one or days to ship or discount impact or net sales.

12:34Yeah.

12:35Done.

12:36And now it's just added to the data model.

12:39Really simple.

12:41Now let's move over to analyze.

12:43This is an area where I was playing a lot with places where every time I saw somebody doing hacks.

12:50In Tableau or Power BI or Looker, that told me the software was failing and we could do it.

12:56Right.

12:56Yes.

12:57So the experience here starts with obviously you can type in what you want, you can double-click what you want.

13:06And it comes up.

13:08I can go in and change and look at sales.

13:12It's just

13:12gonna just really simple at every step of the way you'll see that we have next best actions for the analyst

13:20Yeah.

13:20So it says based on what we know, you might want to break this down by category.

13:26Well, that's a good idea.

13:27Great.

13:28Done.

13:29Now remember in the prep flow, we had our agent here on the side.

13:34Well, it's also keeping track of all the steps you did.

13:37So you can always go back in time and you can also see visually how things progressed in your analysis.

13:46Yeah.

13:46So you can really understand how you evolved.

13:49That's amazing.

13:50Um and if this is something I like, I can actually create a pin, think of this as a bookmark, an interesting view.

13:56And I can reuse that later on.

13:59All right.

14:01Now in this scenario, we've got a lot of different things that are available.

14:05Obviously, you have the different chart types that are available.

14:08Now this is an example where if you want to create a lollipop chart, you've seen all the blog posts on how to do that.

14:16That seems like a lot of work.

14:17Why not just make it available out of the box?

14:20Really simple.

14:22And it's just right there at your disposal.

14:26Or here's another kind of interesting one.

14:28Maybe I'm looking at a scatter plot like this, and we'll put

14:34Profits and sales.

14:36Boom.

14:37And here let's add something with a little bit more detail.

14:41Subcategory.

14:42Great.

14:42I'd love to create one of these amazing magic quadrants.

14:46Where you have each quadrant with different colors.

14:48How do you do that?

14:50That's a lot of work normally, but in our case, what we can do is we have this handy Command K function that shows you all the features on the product.

14:59And I could go and say, can you turn on quadrant mode?

15:04Boom.

15:04We have our quadrant mode.

15:06Okay.

15:07And it does that for you, right?

15:09And it does it all for you.

15:11And so it's exposing the things that are in the product, but we make it really simple.

15:16There's other examples like that, creating a SANKE, a KPI.

15:20This is actually a really nice one as well.

15:22Well, I'm gonna look at my KPI of sales, my order date.

15:27Boom, nice, we're done.

15:29Simple, super simple.

15:31Maybe I want to do a waterfall analysis

15:35All right, let's look at it by ship mode.

15:38Boom.

15:39Super simple.

15:41Nice.

15:41All right.

15:42These are just examples of different ways that you can visualize.

15:45The data, but you'll notice that you're in the flow at all times instead of trying to figure out what calculations to put together.

15:54And then on the left hand side, obviously, if you don't know where to go, the agent is there to help you and say, here are different analyses you could do to help you get started.

16:04You're bootstrapping.

16:05Yeah.

16:06Right.

16:07All right, let's go to now communicating the data.

16:11Now in communicating the data, I use the word communicate expressively because you could communicate as a dashboard, maybe in the future as documents or stories or lots of different ways.

16:22But what you see here is obviously you can create a dashboard manually, but you have now these suggested dashboards based on your data.

16:31Here's e-commerce sales overview.

16:33And it gives you essentially the prompt that the system is going to use to lay out and put the things together.

16:40And maybe I'll say, make me sound like.

16:45A pirate.

16:46That's the classic.

16:48Why not?

16:48Why not?

16:50So now that AI is gonna go in and take what you've put together and start putting in the various details.

16:58Creates the map, lays the out in different ways, right?

17:02Gold by territory.

17:05Of course.

17:06And it also suggests different ways to put together stories

17:11That are more than just dashboards.

17:13They're ways of communicating what's possible.

17:16Boom, plays it all out, addresses it.

17:20off you go.

17:20It's more of a report in this one.

17:22It's just a sort of top level summary as you read down more detail.

17:26Correct.

17:27And if something you don't want, just get rid of it.

17:29Just delete it.

17:30Yeah.

17:30And you can go add maybe that pin that we saw earlier that we thought was interesting.

17:35Nice.

17:36It's added right there.

17:39So what really what you've seen in all this is the way that

17:43We're putting the power back into the hands of the analysts, putting the power back into the data people's hands, but making it really simple.

17:51It's point and click, but it's also AI enabled.

17:54And I know Tim, there's always the question of what about chat?

17:58Can't I just talk with my data?

18:00Of course, you can chat with your data, but it's integrated into the flow.

18:06So if I say show me sales by month.

18:09Right.

18:09Of course it'll give me the answer.

18:13But we give you also the context.

18:14Like, did we get it right?

18:15What do you mean by month?

18:17Oh, did we mean ship date?

18:18And it starts learning over time.

18:21Break it down by category.

18:23It knows the previous question and highlights that in here.

18:29Boom, there it is.

18:31And this is pretty tight, so let's go side by side view.

18:33It's a little bit easier.

18:34So you can see the flow of discussion.

18:36Yeah.

18:36And if it's something you like, just add it to your dashboard.

18:39Yeah.

18:40And the text response is really good as well for accessibility as well.

18:44Yep.

18:45So really the joy of Golden is to make data fun again, to make it fast, to make it intuitive.

18:55to make it something that we would be proud to share with others.

19:00And that's really what it's all about.

19:02It's not about the number of charts you have.

19:06It's not about right the various whiz-bang features.

19:10It's how do we help people answer the questions they have as fast as humanly possible?

19:18Because they do it's fun.

19:19Yeah.

19:21It's pretty impressive, I have to say.

19:23It's very impressive.

19:24I uh it is wild to me.

19:26I think I've I've said this in a couple of contexts.

19:28Like

19:29this concept of the intuition of what people need to be doing with their data and then creating a tool that

19:38understands that basically, right?

19:40So if you ask people what they want, you'll get something very different to understanding the intuition behind what they need to do.

19:47And I think

19:48the like your demo of that has really clearly shown like how you've understood that problem in the modern context, in the in an AI context.

19:57Because a simple thing you would have thought is we've all seen these AI textbooks.

20:01boxes, right?

20:02And you go into them and that's all you get and what comes out is what comes out.

20:05And I think that's been the mode that most people have have understood.

20:09But I think, shoot me down if I'm wrong here.

20:11What I see in your product is

20:13Actually, the user and the AI can interact with the product if if I can use that sort of paradigm

20:20And there's a common surface that the user can work with to interact with both.

20:24So they can interact directly with the asset or they can interact with the AI, which also knows how to interact with the asset.

20:31This sort of Trinity, if I call it.

20:33No, I think you got it right.

20:34It's about humans and AI working together to augment human potential.

20:41Yeah.

20:41And this is where I think it was interesting when I was looking at the different solutions that exist on the market, there were a lot of chatbots for data that were being created that are aimed at replacing the analysts.

20:54Yeah.

20:56And that's not the spirit of what we want to build, what we want to do.

21:00We want to empower people.

21:02We want people to do their best work, not replace them, but give them superpowers they didn't have before.

21:08for.

21:09Yeah.

21:09And sometimes, you know, when you look at AI solutions, they do things incredibly well.

21:16It's unbelievable what it does.

21:18But sometimes you just want to be able to do the work yourself and change what the AI did.

21:23Maybe you got it 80% right or 90% right.

21:26That's already impressive.

21:28But doing the last mile, the last 10, 20%, you need to have that ability to just get it the way you need it.

21:38So it's putting you back in control.

21:40It's putting you at the center of it with AI as a superpower for you.

21:45Yeah.

21:46Absolutely.

21:46That's that different mindset.

21:48And it's pretty profound.

21:49Like at no point in the demo did you show me a function, right?

21:53It was just a prompt.

21:54We just did it.

21:54And but the functions are there.

21:56They're always there.

21:58Of course.

21:59So people gotta watch the slider of autonomy.

22:01It's there.

22:02The point is that doesn't have to be your only way of interacting with those things.

22:06And I guess an interesting question around the

22:11experience.

22:11I think you really nicely laid out the four parts.

22:14It was discover, prep, analyze, and then communicate.

22:17Communicate.

22:18Yeah.

22:18Correct.

22:18Exactly.

22:19That's a really interesting path because I think we see in today's, let's say, landscape this very, let's say, modular approach.

22:28Where actually you do your semantic modeling over there, you have it exposed over here, and you take your story and you share over there.

22:37And I think you've tried to think of the whole experience as one

22:41let's say timeline.

22:42Do you have any sort of thinking on how that works, how that looks when, for example, analysts maybe want to bring, let's say, databrew.

22:51into this platform or they want to bring DBT or import what they've done in DBT.

22:55How does that interface with the paradigm you're putting forward?

22:59Very much inspired by my beliefs at Tableau, at Amplitude, at Cognos.

23:07It's about choice.

23:08If a customer has a semantic model built already, we should be able to leverage it and extend it further to the users.

23:18But if they don't, they shouldn't be blocked from starting.

23:21So it's always this and in everything that we do.

23:25So if you have Databricks, guess what?

23:27You can connect and go.

23:29If you have data in the Unity catalog, you'll be able to leverage it and bring that in so it's available.

23:36Yeah.

23:36But at the same time, it means the users are going to be able to be empowered.

23:41And then the data teams or IT can see what the users are doing with that data and continue to enrich their models.

23:49I think sometimes we have these polarizing debates about A versus B.

23:54And I think no, actually, the two need to work together because we're all working on the same thing, which is data.

24:01But we come at it from a different point of view.

24:03And my view of it is how do you empower people to want to use that data, make it so simple that they want to use it, which makes the entire work that the data teams have done even more valuable.

24:16Yeah.

24:17We don't want to lock it down.

24:18We want to bring it to the people so it can have the most value.

24:21Yeah.

24:22And another question that I've had

24:25A lot of debates over, let's say, is what is the future surface of work?

24:30And so if I contextualize this a little bit, so many companies like

24:35pick a let's say pick a side sometimes they work in Slack sometimes they work in Teams sometimes they're in Google workspace sometimes they're in Office 365 work happens in a lot of different places

24:46One of the biggest, let's say, challenges I see with AI is actually it's challenging that paradigm because you've got the, let's say, the big companies like Claude and ChatGPT saying, actually,

24:55We'll bring those experiences into our platform.

24:59How do you see that sort of challenge with Golden, if that makes sense?

25:02I think the suggestion is here's a browser window.

25:05This is where you experience the product.

25:07But do you see a world where Golden travels into those sort of sorts of platforms?

25:12Or yeah, absolutely answer.

25:15My two favorite words are yes.

25:18And then the second word is and.

25:22Okay.

25:23It shouldn't be an either or.

25:25You should let people consume the data in the tools and apps that they want.

25:30It's not a fight.

25:31It's if they need it there, you should make it available there.

25:34So obviously we have to start somewhere.

25:36So we're starting with the full experience because if you can't use our experience

25:41It won't even work anywhere else.

25:42So you gotta make it great in the first place.

25:45But of course, and we already have prototypes of this, it works natively in Claude already.

25:51Just with an MCP, you can start asking questions.

25:54But all of the if you the magic behind Golden, the suggested visualizations, the auto insights become available.

26:04So we're creating essentially this enrichment layer that makes that data travel even further.

26:10But now when something happens and you have to go discover further, you have a place to go.

26:16Yeah.

26:17Amazing.

26:17And some users prefer one interface versus the other.

26:20Yeah.

26:21I remember the whole debates when the iPhone came out.

26:24Said, all right.

26:25PCs are gonna go away.

26:27We don't need laptops.

26:28You're gonna do all your work on your phone.

26:30We all have a phone and a laptop or a desktop, and we live in a world of ant.

26:36It's not an or.

26:37There's some people that are purely mobile, but the majority of us are using different experiences different different everywhere.

26:44Yeah.

26:44And I think it's the same thing.

26:46And so I think now we're very much in the world of, oh my God, this is incredible.

26:52It's a new thing.

26:54But it'll work in many different places.

26:56Many places.

26:57Perfect.

26:59Okay.

26:59This is yeah, I'm again I've seen this demo and actually I saw a lot of new things today, so I will just call out to the audience like

27:06The distance you've traveled since our last saw the demo is wild.

27:11It's unreal and you've barely seen.

27:13Like the I just wrote down what we shipped last week and it's unreal.

27:18Yeah.

27:18But I think actually this is the other really amazing thing of this new era is as a builder, I can build so much faster than ever before.

27:30The one time a year, three times a year, four times a year, that is glacial pace these days.

27:36Yes.

27:36Like the pace of development has accelerated.

27:40accelerated and I think that's great.

27:43It means that you're able to do more, you're able to solve more problems.

27:47But as a user, you don't have to wait 10 years for a feature to get finally delivered.

27:54Yeah.

27:55It can get delivered in a matter of hours, days, or weeks.

27:59Yeah.

28:00And that that's em again, it's empowering.

28:02And it's not to say that the previous generation was bad.

28:05It was just built in a different way in a different.

28:08And now there's a new way of building that is just remarkably different and better.

28:14But it means you can move so much faster.

28:16Yeah.

28:17And I remember in the tableau days, we had this incredible ideas forum.

28:22It was an amazing way of generating feedback from the community.

28:27Yeah.

28:27But our ability to take that feedback and do something with it was limited by the complexity and the number of develop developers that we have.

28:36Yeah.

28:36Well, if I can move 10x faster.

28:40How quickly could I go through that list?

28:43That's just again, I'm just getting giddy, excited about what we can do for the community.

28:48And you can take on bigger challenges.

28:50Like I think one of them sort of, let's say

28:54Underlying stories behind a lot of features in analytics have always been the investment required to bring about the innovation to support that feature.

29:02So it's not often the feature isn't the challenge, it's the

29:05underlying infrastructure to support that feature in a reliable way that can go on.

29:09I think just talking to some colleagues who work with AI a lot more frequently and who are building products elsewhere.

29:16That's been a common challenge.

29:18People are taking on what are what they would have considered side quests a lot more frequently and they're freeing up the time to take on much, much bigger sort of conceptual challenges because they've got this little G

29:29Which is which I think it's amazing.

29:30I think it's exciting.

29:32I think if I change tack a little bit and if we step back and maybe just look at the industry you're enter entering, you mentioned earlier on that like you were seeing things that weren't being solved in the industry.

29:43What else would you say is probably like a big challenge for the BI industry in general, probably yourselves included, because it's not maybe let me rephrase it this way

29:53The pace you can now move at means there's so much more available to you in terms of direction and focus.

29:59And I think all BI companies have this shared problem, which is now we can achieve so much.

30:04Where do we even start?

30:05So how do you

30:07Focus your team, how do you focus your product?

30:11How do you say, no, let's not do that, let's do this, when anything's possible, if I could put it up

30:17Anything is possible, which is great.

30:20What is different today is what becomes more important is are things like what we call product taste or product sense.

30:26Do you deeply understand the customer problem?

30:29Because you can build anything you want, but if you don't deeply understand it, if you haven't walked in their shoes or felt their pain, it's hard to build the right thing.

30:38And I think that's a difference today.

30:40But if we separate out a few things that you said, first, it doesn't matter what tool exists, we still have the same challenges that we've always faced.

30:51Data literacy challenges, change management challenges, data governance challenges, they're still there.

30:58So while tools may make it more approachable,

31:02We still have to get more people to get comfortable with data.

31:06And I think my view with something like Golden is to try to reduce that barrier even more so we can have even more people in this great community that we're part of.

31:16But it's changed, right?

31:17You have to do things a different way.

31:19I know that when I started putting this together, I brought in a couple of people that I know that have used Tableau and Power BI for years.

31:26And the hardest part for them was to untrain themselves for.

31:30Or how they've done things for 10 years.

31:32Yes.

31:33And saying, is that the right way of doing it?

31:36Not saying it's the bad way, but there's a different way.

31:38And it was, by the way, it's exactly the same in the early days of Tableau when people were coming from business objects and cognos.

31:45expecting things a certain way.

31:47No, it's a slightly different way.

31:48So you have to get comfortable with it's a new way and then be willing to lean in and move forward.

31:55Yeah.

31:56But the other aspect that's happening these days is when you're starting fresh in a post-Chat GPT era versus pre, right?

32:05Where you're not burdened by

32:09Decades of technical debt and issues.

32:12It's actually quite liberating because now, you know, we're building with the LLM as our silicon.

32:19Yeah.

32:19And so as new capabilities come alive, we can exploit that really quickly and move faster.

32:25Whereas in the previous generation of tools and

32:28This is true for a Tableau or Power BI, a Looker, a Metabase, and all these different tools that exist, which they're plenty, it's a big industry.

32:38They have a lot of legacy coat that is going to be hard to move over.

32:45And I think you have to be willing.

32:50To make big bold bets to transform because it won't just move forward.

32:55And it's really hard

32:57There's companies like Intercom who were growing really fast for a long time and then stagnated.

33:03And the only way that they could move forward was to burn the boats and re-

33:08And they re-accelerated because they restarted in right an AI native way.

33:14Yeah.

33:15And that's hard.

33:16Truly, because you have customers, you have to bend a lot of culture, history, land experience, and happy customers as well

33:24That's right.

33:25Absolutely.

33:26So we get to start from scratch, which is a benefit.

33:30It also means we're gonna have to build a lot more things that maybe didn't exist before.

33:35But as long as the things we build are delightful, empowering, liberating, and help people move forward, I think that we can build something truly remote.

33:47Amazing.

33:48And so as I've said, like again in month in weeks you've gone what I would have considered like a quarter of the stuff shipped, and that's just the stuff I've seen.

33:58There's more I haven't seen, I'm sure.

34:00What does a roadmap look like?

34:01Because that's also probably a different challenge, right?

34:07Which would suggest that you can hit that shit point sooner.

34:11So yeah, what does the roadmap look like?

34:13How does that work?

34:15You know, Tim, you've never changed.

34:17Every time I share what we're building, you always want to know what's next.

34:21You're never satisfied.

34:25Check out what we're doing in the next twelve months.

34:27Okay.

34:28What's b after that?

34:32The the there's two there are many parts of that.

34:34The first thing is we have to delight customers.

34:36We're gonna ship something, there will be bugs, there'll be issues, there'll be feature requests, and we have to deliver on that.

34:42And so a lot of that roadmap will come from what the customers tell us.

34:47us and I think we have to be deeply connected to customers and that's actually one of the values in the company is there is no separation between customers and developers.

34:57Yeah.

34:58Everybody's accessible.

34:59They're all connected.

35:00So you have to move forward.

35:01Yeah.

35:02But as if I look at what we're going to be building next, there's some subtle hints in what I showed today.

35:08I use the word communicate, dashboards.

35:12Yes.

35:13So you'll see other ways of communicating.

35:16All right, maybe it's through slides in stories and documents.

35:20You'll see a lot of leverage in the platform.

35:24Obviously, we have a concept we call shared connections.

35:27So you can build a connection that's leveraged everywhere, but it also includes things like metrics.

35:33and being able to have curated metrics that land everywhere that you can consume anywhere.

35:38There'll be things like monitoring agents and workflows and

35:43I hate to use the word catalog because that's it's an older word.

35:47There's a lot of metadata that's being generated in the platform and how do you leverage that to build knowledge?

35:54that can be shared.

35:56Because the more knowledge there is, the better the environment becomes.

36:01So we part of the company values is not only no separation between customer and

36:07and developers, but speed is an attribute.

36:12Yes.

36:13We are going to move fast.

36:14We're going to innovate fast for customers.

36:15And I think now we have that core possibility of doing that.

36:19Yeah, that's amazing.

36:20And that's exciting.

36:21I think that's but if you've used any experience, a really good experience, you like your phone or whatever, I think one of the things you've always appreciated is yeah, when you're

36:30gets better and f software update, improved features, and or saves you battery and or gets faster.

36:36Like very simple things, but actually means a lot to experience.

36:41If I touch on one thing you talked about there, which was the catalogue and metrics, I think I often see clients, especially in the consulting world, deliberate over these two concepts.

36:50Maybe not deliberately, but deliberate on them a lot.

36:54And one of the opinions I've had for a while is that actually we build too many dashboards.

36:59We we focus too much on the end state before really grounding ourselves in like the

37:05The thing that matters, and I put metrics in that category.

37:08If you can define a metric, where it goes from there is totally game.

37:12That's exciting.

37:13That's easy.

37:13But if you don't define it, you will burn a lot of time in the end solution.

37:18trying to make it right.

37:20And in Golden, are you thinking of ways of really trying to get the user to do the work in the right place?

37:26Because your interface is very flexible.

37:28You can go to any part really however you want through the prompt window at any time.

37:32And so how do you instill

37:36let's say this catalog mindset to close the loop, how do you make a user say, actually, before I create this, let's go find what's already there.

37:43How do you get them

37:45to do it in the prep stage, but not in the analyze stage, right?

37:50Because that's these are the really bad patterns that B Idols of today have instilled in us, right?

37:54And how do you break that model to really make what you're doing sing?

37:58Because I think

37:59Your platform does need that structure to do things consistently well.

38:03Right.

38:03That's a great question.

38:04And what's my favorite word?

38:07And yes, I'm sorry.

38:09I did it in the wrong way around.

38:16So it it's an end.

38:18Look, I think part of what happens with a lot of these solutions is without governance, you have anarchy.

38:26Too much governance and you have zero adoption.

38:29because it's too limiting for folks.

38:31So you have to find that Goldilocks balance between.

38:36But what I think is actually interesting is being able to harvest the organic development that

38:42users do combine with the centralized approaches.

38:46And so by that I mean that if you think of you have a table called accounts.

38:51It could be perfectly curated, great, amazing we have accounts, but you don't know yet how it's being used.

38:58Is it these are the accounts that are at risk of trading?

39:03Are these upsell opportunities?

39:05Are these accounts that are up for renewals?

39:08Like

39:09Yeah.

39:09It's the same table, but used in different ways.

39:12All of that human enrichment, which call it a dashboard, call it whatever asset they create, that becomes knowledge.

39:20So, what that really means is you're able to start leveraging that and from a centralized point of view, see why are people creating these derivatives?

39:29Yes.

39:30Is it that there was a need that wasn't met?

39:34And if there is, then great, let the users define it because they understand their business, and then harvest that out, put it centralized so that others can benefit from it too.

39:45Yeah.

39:46But limiting somebody for to do the thing that they need to do won't work because if you do that, guess what they're gonna do?

39:53They're gonna do an export to Excel and gonna they're gonna find a way to do it anyways.

39:57Yeah.

39:58So if you can do it in a way that helps you understand what is actually going on and continuously refine, I think you end up in a better state.

40:07Yeah.

40:08So again, the view is do an end, enable people to centrally do it, whether they do it in golden or they do it in DBT or Snowflake semantic views, it doesn't matter, right?

40:19But now they have visibility into how people actually

40:21actually use it.

40:23Yeah.

40:23Yeah, that's it's a good balance.

40:25And I think also y you you'll probably come across, say, customers I call them, clients who

40:31have a very let's say opinionated view of where it has to happen.

40:36Not necessarily where it should happen.

40:37That's sometimes a different I often have different views to my plans.

40:41I'm not sure that's what it should happen, but fine.

40:43We'll just go with what you're suggesting.

40:44And I think, yeah, supporting that approach will will really help and maybe make it easy for them to A-B test.

40:50That's something I wish clients did, right?

40:51Like instead of choosing one, actually let's go with both approaches in two different use cases and see.

40:56Which one works rather than having this sort of top-down approach and yeah, I'm I feel I can't know this, I haven't used a product.

41:03I feel like Golden kind of enables that if that makes sense.

41:06So the collaboration element of any of these solutions is really important.

41:10Because we might be seeing the world differently, and that's okay.

41:14But can we communicate it together and debate it together and see where you're coming at it?

41:20Oh.

41:20You calculate this way because of these conditions.

41:23Okay, that makes sense.

41:25But have you considered this?

41:27And that collective knowledge makes us all better.

41:31Yeah.

41:32Yeah.

41:32A hundred percent agree.

41:35Yeah.

41:35I don't again, I'm so impressed.

41:37I've I'm amazed how much it's come along.

41:39It's really incredible.

41:40I'm excited.

41:42to to to also use a product that I will be at some point in the future, I'm sure.

41:45And I guess how would uh how would people find golden?

41:48It sounds like a really

41:50Strange question to ask because for all I've known you've been obvious way to go.

41:53But yeah, like how do people find you?

41:54How do people get in touch?

41:57Well here's my phone number.

41:58Call me anytime

42:01We're in the early phases right now.

42:03You're all getting the first peek at what this product will be, and there's a lot more work to get done.

42:10We're starting with a wait list on the website.

42:13So that people can nominate themselves to want to join in and we'll be starting to pull people from that wait list.

42:20Yeah.

42:20And then product availability will be relatively soon right after.

42:24Right.

42:25But we have a website, goldenanalytics.

42:27com.

42:28It's really incredible.

42:30You can sign up for the wait list, see some videos, but we'll be revealing more over time.

42:35And there's definitely a few surprises

42:38To come and one thing that we did is we actually hid some golden tickets inside the product.

42:45Ooh, very nice.

42:46So if you do the right combination of keystrokes and

42:51movements, you can actually get some golden tickets and open up the wait list for other people.

42:55So hopefully there'll be a fun like game that people

43:00You really have pushed this golden pun to like it's right another level.

43:04Golden tickets.

43:05It's creating golden moments at a golden time.

43:09Exactly.

43:09It's the golden era of analytics.

43:12Maybe it's the golden hour, yeah.

43:15You wake up at five AM just to work in the ideal time.

43:18That's right.

43:20But like Tim, like this is all about joy.

43:23It's fun.

43:24Data shit.

43:24It is, exactly.

43:25We're laughing about it.

43:26Exactly.

43:27It's a really good way to think about it.

43:29Yeah.

43:30I should have thought of that as I was laughing away, but yeah.

43:33It's perfect.

43:34So listen.

43:34And you've got a golden hoodie already, so you know you're set

43:40I'll wear that on my way in, although I'm not sure it's welcome, but hey.

43:45You earned it for life.

43:46Yes, I I I I do love it.

43:48I love that moment.

43:49Uh but I'm excited to use it.

43:50I can't wait to try it out.

43:52Obviously, once it's more generally available, I'll be making content about it.

43:56I think

43:57it's really important to explore tools and how they're approaching this.

44:01And I think you're the you're definitely the first, let's say

44:05I'm gonna call it, I don't know if this is even the right word, you tell me multimodal tool, because typically a lot of AI tools have said work in this text box, and then BI tools have said we'll add a text box to our thing.

44:17But no one's really supportive.

44:18What if I just want to bring my typical workflow to a modern way of working, like an AI approach?

44:23And that's what I mean by multimodal.

44:25You can use it whichever mode you want to work with and it supports both to drive the experience.

44:30I'm excited about that.

44:32I'm also sure that

44:33you will let's say trigger some sort of response from the market, the world, the community, just to see what challenges turn up to

44:41to take you on because I I think that's healthy, right?

44:43I think you saw that absolutely coming from Microsoft to Tableau and Tableau versus Power BI was always a big thing.

44:49I think now the game is filled with many mortals.

44:53I think

44:54even just in the traditional BI space, there's now at least ten competitors that can turn up to any RFP to take it on.

45:01So I'm also excited to see what

45:02turns up in in the space you're you're targeting.

45:04Because I think that is where we're going.

45:06And it'd be interesting to see how quickly companies pivot to this sort of direction, which I think is a great challenge.

45:13But

45:15But I'll add that my point of view is I love learning.

45:18I love innovation.

45:19And it's actually great that there's competition

45:22It's great that there's different companies that are approaching the problem from different points of view.

45:28And I think for everybody in this community, just being exposed to what's out

45:34Just like you're just him now.

45:36Yeah.

45:36But because there's a lot out there for people to look at, learn from, be inspired by, start using.

45:45And I think that makes us all better.

45:47So having an open mindset is really key.

45:51And really understanding what's that company's or that product's point of view of the data space.

45:58And it's not that one is right or wrong.

46:00They're just different.

46:02And where do you lean into its strengths and its differentiators?

46:05And that I think is part of the joy of building.

46:09And I think the more that people keep an open mind,

46:12That there might be something different, the better it is.

46:15Yeah.

46:15Uh myself coming at it from years at Tableau, years at Microsoft, seeing things in a different way, I think it's important for people to not.

46:26be the things that they tried to be in the past, right?

46:30Yes.

46:30We replace the report factories to only turn into another report factory.

46:34Yeah.

46:35We gotta break that mold again and not get stuck to our old

46:38ways and be open to other ways of doing things.

46:41That's amazing.

46:43Yeah.

46:43Amazing.

46:44Good.

46:45I'm excited to see how you take on that challenge.

46:47And yeah, obviously I'm hoping you come back.

46:49At this rate you'll need to come back like quite often and to show us what's next.

46:53What I'm looking forward to is your analyses of our next product announcements so that you can start reading the tea leaves of where we're going because you're very good at that.

47:05And say connect A and B and C and then turn it this way.

47:09Yeah.

47:10Exactly.

47:11Exactly.

47:11Let me know.

47:12I'll be there in your first keynote.

47:16I can't wait.

47:17I can't wait.

47:18And one of the reasons I did think about my brand deliberately was, hey, actually, we are moving into this sort of multi-tool ecosystem.

47:24And it

47:25it would be dangerous to assume that you can sit behind one direction and just keep going with that.

47:31I think you people will have to adopt many different ways of working in different parts of the data ecosystem.

47:36And Golden is just one of them, but I'm sure that's really more.

47:39Yeah.

47:39Yeah.

47:40Amazing.

47:41Listen, thank you finally.

47:42Thank you, Tim.

47:43I know we we made it out.

47:44It's our first time.

47:45It's crazy after knowing each other for how many years and

47:49Yeah, I should have taken that opportunity sooner, but I'm glad you you've showcased your product here on the channel for the first time.

47:54That's super special to me.

47:56Thank you again for sharing that opportunity.

47:58And yeah.

47:59And keep doing all the great work that you're doing and sharing the love with the community and helping people discover different ways of using data that I think is really important.

48:10Yeah, amazing.

48:11Thank you so much.

48:12See you soon.

48:13Thanks everybody