Trulip: governed AI dashboards built on Tableau by Infotopics

0:00Twenty years ago, we built BI platforms to help solve one problem.

0:05We wanted to make sure that everyone sharing data and information in our organizations was looking at the same thing.

0:11And we built SQL to make sure that when we asked those questions, we arrived at the same answer.

0:17Fast forward 20 years, and it feels like we're back in the same place, but this time AI is the new challenge.

0:23You see, we've seen the expansions of platforms adopting MCPs and capabilities that allow us to build solutions.

0:31But what's happened is that we've sort of, you know, loosened the belt a little bit on security and governance.

0:38And so a lot of these vibe-coded solutions are moving around through the organization.

0:43And chances are that IT are just probably here, heads in sand, or they're struggling to keep up.

0:48And so whilst

0:50People want to keep innovating.

0:52Whilst people want to keep up with the moment and take advantage of what they can build and what they can do and what AI enables them to do, there's this equal friction of wanting to do it the right way.

1:03Now, I wanted to talk to two people, Merlin and Klaus from Infotopics.

1:09You see, they've spent the last 10 years

1:11building on top of the Tableau ecosystem, which means they've they've suffered essentially learning all the ins and out of how to build

1:20extensions on top of a BI platform.

1:22And that has given them an intuitive understanding of how to do it right, how to do it well, and importantly they've worked with customers to solve all these problems.

1:31by deploying these inside of the Tableau ecosystem.

1:35And in speaking to them a couple of weeks ago, they showed me something really interesting.

1:40They call it Trillip.

1:42To me, it's a good response to this exact problem.

1:46How do you allow people in your organization to take advantage

1:49of the AR moment whilst giving them the same governed access to the same data sources that you've already got inside of your Tableau ecosystem.

1:58In this conversation, we go into that into a lot of detail.

2:01I sort of didn't know what to call this video because to me it's a little bit of a platform showcase, but at the same time, it's something that sits on top of Tableau

2:10I asked some pretty tough questions throughout this conversation as well.

2:14I even asked them, hey, why why can't you just use this thing on its own?

2:17Why do you need Tableau?

2:18Anyway, that's a question for you to ask and find out more in this video.

2:23As ever, let's get stuck in.

2:25Merlin, Klaus, how you doing?

2:27Good?

2:28Very good.

2:29Excellent.

2:29Welcome back.

2:30I think we've had Merlin on the channel before.

2:32We've had infotopics on the channel before, generally speaking.

2:35And it's really good to have you back.

2:37It's really exciting because I think you have a new product to share

2:41Let's start with introductions first, just in case anyone's watching and they haven't met infotopics or either of you separately.

2:48So maybe Merlin, if I hand over to you and yeah, introduce yourself.

2:52Sure.

2:52So my name is Merlin.

2:53I am the CTO and founder of Infotopics Apps for Tableau.

2:57And our company specializes in supercharging Tableau with game-changing riper capabilities, powerful visualizations, and all kinds of AI accelerations.

3:06And we are known for products like Super Tables, which is a powerful table for Tableau that let the enter end user interact with it like it's Excel.

3:14But also WritePack Extreme where you can build any WritePack use case for Tableau and many more extensions.

3:19And we have been doing this since 2017 already.

3:22That's where we started.

3:24And we have customers all over the world using those products.

3:28Amazing.

3:28And Klaus, over to you.

3:30Yes, I'm Klaus.

3:32I help Merlin and Richard to managing the team at Infratopics and I take care of everything.

3:37Growth, bringing our products to our customers and help them to make the most out of it.

3:42People might also know me from the Tableau community, so I've been around for a while already.

3:47I think like all of us.

3:49And it's been a fun ride so far and I'm really excited about TrueLith, the thing we are going to talk about today, the new product, and what it brings to our customers.

3:58Amazing.

3:58And I must add that I think when I was learning Tableau, God, it's a long time ago now, twelve, thirteen years, I came across your names and I always looked at your work and I was always inspired by the things you were doing.

4:09Merlin was

4:10hacking something.

4:11I think you came up with you're the first person I saw who made a game inside of Have Lay, which is which which I was I found pretty incredible.

4:18And Klaus

4:19You're one of the visionaries I looked up to for a very long time and I think you helped set the standard for what I think visionaries are today.

4:25So I'm gonna add that because I know you're both humble, you won't say that, but so yeah, no

4:29It's been an absolute privilege to have you in the community and I've really enjoyed sort of your work and I'm looking forward to see what Trillip has to offer.

4:36So maybe let's get into that.

4:38What is Trillip?

4:39if you can explain it briefly before we have a look.

4:41Yes.

4:42So w just like I said, we are building products for Tableau since twenty seventeen.

4:46And we also so we have quite a lot of separate solutions for Tableau.

4:51And last year we started to think like what would happen if we start to combine these these applications together.

4:58And we also had customers that started to ask this.

5:01For example, can we do right back in super tables?

5:04Which means if we combine Ryback and Supertables, you get a whole new product.

5:08And we kept on doing that for the last year, so more integration between the products

5:12And then last year we said what if we combine everything in one platform?

5:16What do we get?

5:17And that's when TrueLib actually came, because if you combine all our visuals

5:22we have the tooling we have like the meal schedule with pixel per with reporting right pick as a data engine we've got all the analytics arts you actually start to get an analytical product

5:32Like a full platform.

5:33And that's where the excitement came from.

5:35Like, hey, I think we got something here.

5:37And then of course we got the whole AI revolution where it's which we were also adding into our products.

5:42Which we thought like, hey, but what if we also add more AI into this?

5:46and answer those questions we get from our customers.

5:49And that's actually the the whole story and also the product that we've got to show you today, which is not just a Tableau extension anymore, but it's an analytical platform.

5:56You can use.

5:57Yeah, and and and with this platform we are also able to help our customers with use cases outside of Tableau, which which has been

6:06uh highly demanded feature requests, so a super table that runs in the A BI portal or in in in some other application.

6:14We were not able to to deliver this so far, but with Trulib we are now eventually able to do that.

6:21That's one maybe the third reason.

6:24after yeah integrating everything into one platform, adding AI capabilities to the product and to the platform, yeah, the integration in for other use cases.

6:34Amazing.

6:35Amazing.

6:35And I have to say, I think this comes at a moment where it seems like Tableau's Salesforce is opening up the Tableau platform a little bit more, right?

6:43In the last release

6:44four years APIs have really accelerated, but like in the current moment with AI and all these capabilities, there is suddenly an interest in building things on top of

6:54tableau and I think maybe that's a really good time for your product because a lot of what I see at least is very all over the place.

7:00People are just vibe coding this or that and that they don't have the structure that I think

7:05I understand your product has to offer.

7:07So as we get into this, I'm really excited to see how that works and set it all up.

7:11Maybe let's just show people the product.

7:12Let's just see what it is and how it works.

7:15So I'll hand over to you to Yeah.

7:17Sho show us what it does?

7:18Good.

7:19Yeah.

7:19Yeah.

7:22Sure.

7:22That's one of the reasons or why we have built this

7:25Because we were talking to customers that were actually saying we want to do more, we're using Tableau, but we want to like have dashboards faster, we want to play with chat GPT to generate dashboards.

7:35Yeah.

7:36What we saw was that those customers were actually generating dashboards and I got the feeling we were going back to the Excel hell again.

7:45Because they were sending around HTML files and putting them everywhere in the coun in the company on on file shares, on hosting platforms, internally and maybe yeah, externally.

7:55And if I like oh no, we're going back into the Excel hill again.

7:58That's the whole idea of those platforms to not do that and have authoring and authentication stuff in place.

8:03And that's exactly what I'm gonna show you.

8:06Like you got the five coding

8:08And dorphine, like the quick building, the good stuff, and then in a safe and secure platform.

8:14Right.

8:14Running SaaS and on premises.

8:16So that's also a big thing.

8:18So let's start with a little demo and start.

8:22We can start where we come from and then I can we can build up to where we're going, okay?

8:27I see.

8:28So this is the platform.

8:29It's called Schrule.

8:30And as you can see, it's already starting with a big how can I help you, Box?

8:34So this is our Visi assistant, where you can ask anything about the platform or let it do things to you.

8:39I can show we will show to you in a second what it all does

8:42But first thing what I want to start with is with extensions because that's where we're coming from.

8:47Yeah.

8:48And many of the customers will recognize these lists.

8:51So we go to extensions.

8:53Yeah.

8:53And I will show you one of our products as an example how this runs in Tableau.

8:57So this is the list of products we have.

8:59And one of those products many people know us for is super tables.

9:02So I'm gonna download this with this extension.

9:04And I will show you how this runs in Tableau.

9:07So if you get in here, go to the extensions and then download it right here.

9:14Open it up and it shows super tables, right?

9:17So we can drag columns on it, and this is what people love to do.

9:21Measure on it, maybe in one more dimension.

9:24And then I can actually product in there as well.

9:27Here you can do things like this.

9:28So you can group columns, you can pivot, you can do all the cool things you can add.

9:32There's this is one of our products.

9:34And if you've been

9:36Looking at the list, you also see there's a new button here, which is the embed now button.

9:41Okay.

9:41That's the thing that is not present in our current solution, but in TrueLoop it is.

9:46And if I just hit button.

9:48Yeah

9:49It creates an analytics view.

9:51And this is like a brand new environment you didn't see before.

9:55So this is analytics, TrueLib Analytics.

9:57And here we have a full-fledged analytics platform where you can use R extensions, add filters, different visualizations, extensions from other vendors if you want, which are also supported.

10:11And basically anything you like.

10:13So maybe I will quickly show you what you can do.

10:15We can connect to a data source here.

10:17We didn't do that yet.

10:18Let's connect to superstore and we can drag in.

10:23As simple as this

10:25Just like you saw in the tableau as well, a dimension and a measure.

10:29And it's exactly the same product.

10:31You can also still configure it with all the settings you have and super tables.

10:35And you can then if you want ship it out by sharing it either with an embed code or if you like you can actually

10:43Put it back into Tableau as a as an extension.

10:46And the reason why we do this is because many of our customers are using Tableau and they want one platform to go to, right?

10:53So they want to go to your analytics.

10:55And with this feature, they can still use our products inside the Tableau environment.

10:59So people can still go to one environment and then look at Tableau dashboards or embedded analytics true lip analytics dashboards.

11:06I see.

11:06Yes.

11:07And yeah, since we also support vibe coding and generating new extensions, we obviously didn't want to reinvent the V the wheel, like super tabled.

11:17Has been on the market for seven, eight years already, and we have so many customers and so many feedback went into this product.

11:25So we want to open that up.

11:26Of course also for all our customers.

11:28And yeah, this is also nothing that you can vibe code.

11:31If you Merlin just showed the configuration, there are so many there are so many configuration options that all come from actual use cases that are custom.

11:41But this is nothing that you that you just generate with a prompt.

11:44So this is real craft.

11:47It's tried and tested and it's it's actually a really good point, because that's the thing that vibe coding.

11:52It doesn't give you, it doesn't give you everything that comes after you vibe something.

11:56So shipping it, testing it, feedback, the many iterations of feedback.

12:02surviving releases and version releases, surviving security changes by the vendor platform, surviving changes in the underlying platform like Snowflake change authentication.

12:12I think twice in the last five years, right?

12:14And so all of that comes in the product.

12:17You don't get that if you vibe kit, which is important.

12:20Yeah.

12:20Yeah.

12:20Yeah.

12:21Yeah.

12:21So to to go a little bit more onto this this this product, this true analytics, let's just create a new page.

12:27And see this is my dashboard.

12:29And so as you can see here on the side, we got a

12:31Long list of visualizations we support from Cartesian charts, radial charts, there's slow charts, Senki's, there's even geo charts, like maps

12:41And there's this whole list of all our extensions as well you can just drag in here and use like we just saw super tables but do we control entry of course and ripe extreme you can use in here super tables

12:52And all the show me more visualizations were known for.

12:55So there's like a whole list of products you can use.

12:58And besides that, quite exciting to show

13:00show you also these elements we got filters images and we got the HTML component which can do anything you want so it's a basically a template the AI can use to build templates and use as visualizations.

13:10Yeah.

13:11And yeah, shall we just show that with uh yeah, okay, good.

13:14Well yeah.

13:14What is the great and the the cool thing is I can drag in the HTML component and try to build it myself with HTML.

13:21But I can also just open VIS AI

13:23And just ask him like create me a a let's see we can actually do something.

13:30Yeah, create me a uh uh uh KPI grid

13:34for category with sales, profits, and quality using the HTML components.

13:41Let's see what it does.

13:45So now we're talking to Vici AI that's using our own AI engine.

13:49It understands what I'm trying to ask and understands our analytics platform, understands the visuals we have.

13:54And as you can see, it already built a

13:57A list of KPI cards here.

13:59Right.

14:00Just in one second.

14:01And now I can actually go in there and I can change it.

14:04So we have the AI that generates components, and now I have the power to actually change them myself.

14:10So for example, it's grouped by category now.

14:13We can also group it by subcategory or any other dimension.

14:16So then it will generate a card KPI card or row of data, I have to say, for each of them.

14:21Yeah

14:22It's using these measures.

14:23I can actually format them here as well.

14:25So we get good formatting.

14:26And this is the template it generated.

14:28So you see it's using those little brackets to generate like smart tags.

14:33And it generates the full card.

14:35And I can for example also go in here and say, okay, I don't like a two-grid column.

14:39I want, for example, this.

14:41Or actually I want KPI cards below each other.

14:43And then I can create all kinds of different KPI cards.

14:45But you can also imagine that you can use

14:47Anything here that is HNO.

14:49You can maybe use like player ranking for football teams or with images, everything you can imagine you can put in here.

14:55Just it's just one prompt and it works.

14:57Okay.

14:57And I think this is already a very exciting feature using the Visi Assistant.

15:01But the Visi Assistant can do way more because if I just grab a little piece here and I say let's make something fun.

15:09Using my favorite restaurant, of course.

15:12So I'm gonna ask it to create a one-page McDonald's team superstore sales dashboard, folks, sales with McDonald's style titles and using dollar as a formatting.

15:21And I want some specific charts and it should focus on category a second.

15:24I can actually ask that as well.

15:27And it will understand my question and it will create a plan on how to build this dashboard, which we can review.

15:34And after that, we can actually let it so we can look at the plan here.

15:38Okay, write some things that it wants to do.

15:41And if I like it, I can say approve plan.

15:43And it'll start build start building this full dashboard for me.

15:46Amazing.

15:47Let's see, here we go.

15:53And so what's happening behind the scenes here is obviously you've got the data source and it's working with the prompt and then also

15:59I guess an AI agent behind the scenes to give all bring all that context together and then create what we've just seen here on the left, right?

16:05Exactly.

16:05Yeah.

16:05So we'd be working really hard to to build an engine.

16:09for and we also have MCP for example that have all these tools available for the AI to understand like how can I build a dashboard, what charts are there, how can I configure them?

16:17How what is actually a dashboard?

16:19Because it's a JSON object in the end.

16:20How can I build it?

16:21How can I do title?

16:22Everything is in here.

16:23Just to show you that that will show you more.

16:26And that results into just one prompt and it builds a dashboard, which is quite fun already to show it.

16:34So I need this is of course I'm just asked for specific charts to show up a little bit.

16:38And the fun thing is from here I can actually just I can just play with the interface and it will just update.

16:43So I can again take over control of my charts.

16:46And do anything I want.

16:47So I can make this one wider or um and I can actually if I don't want it just undo it again with the command C.

16:53I can remove charts if I want, if I don't like them.

16:56Right?

16:56So I can just do the things that I

16:59I'm used to in analytics tools.

17:00So it's not only an AI that's vibing, it's an AI that's vibing, and I can actually play with it myself.

17:06And one of the fun things is I can actually dive into a chart like this one.

17:10I said, you know what?

17:10Okay, I like this chart, but I would like to have at a border radius.

17:18It's really contextual prompting for the chart, so you can just customize it.

17:22Yeah.

17:23Yeah.

17:23So now it will now dive, it will now understand that it only has to look at this specific chart.

17:28And add a rounded corner to this specific chart here.

17:32But of course you also have access to the settings directly.

17:36So it's not needed to do everything through prompting.

17:40If if you know where the setting is, it might be even faster to to set up this rounded

17:45corner directly and then asking for it.

17:48But exactly.

17:48Yeah.

17:49I think with this we also help to lower the barrier to people who are not

17:55Who don't know the settings.

18:00Yeah, this will also di democ democratize the building of analytics products.

18:05Yeah.

18:06So you see it's just actually

18:08Let me put it in the air sets with it.

18:10It's actually using our see you can using all our settings we have here.

18:15You got a lot of settings.

18:16So for example, data labels, there's a prefix here, we don't need it.

18:20See we can add that small places here

18:22Yeah you can maybe make it a it's a currency but you can also make it a thousandth number if you want.

18:27But there's formatting here, so we can use different formatting.

18:30We can color by a dimension.

18:33Yeah.

18:33We can say, you know what, I want to color by and let's colour it here.

18:36So you should color by dimension, you can change that as well.

18:39Let's reset.

18:40So ever everything you would expect from a analytics dashboarding platform.

18:44Yeah.

18:44Yeah.

18:45Exactly.

18:46Giving people the ability to dial it in.

18:47And you're using your infrastructure that I think you've perfected over multiple sort of versions of building extensions and y you know it works, which is another thing.

18:57Yeah, we got a lot of experience.

18:59And I think one of the great things is it's interactive as well by default.

19:02So I can just click on these things, right?

19:05And that you see that the charts were updating automatically.

19:07Oh it's basically using the same dimension, but you understand.

19:10It's it that it's out of the box interactive as well.

19:12And we can add filters to it and all the things that I lost.

19:15And then this is basically and and just small feature will also love this.

19:18But for example, if I just grab a chart, I can just command C, Command V it, and I got a copy of it.

19:22And I can just edit it and it just is I can just backspace it and so on.

19:26Small things to really it quickly build charts and build dashboards.

19:30Together with the AI.

19:32I think one more feature I want to show you here.

19:34This is also what we really.

19:35So if I'm just making say example, for example, we ask it something like, show me the top 10 customers.

19:44for sales.

19:45So this is a a another question where we're not really using analytics, but we can just talk to our AI and it will generate cards in line.

19:54Yeah.

19:54But now comes the power because now I can

19:57If I want to have this chart in a dashboard, I would have to recreate it, right?

20:01Correct.

20:01But that's not the case.

20:02I can just grab it and I can just put it on my dashboard

20:05Nice.

20:06It will also format in my dashboard format automatically.

20:12And I can actually also do this.

20:13I can just grab the text.

20:16And that is also in here.

20:17And this is now an HTML component.

20:19The one we showed you.

20:20And you see, it's using those dynamic tags out of the box

20:24It's got the context and the metadata to go with it.

20:27Yeah.

20:27It's just it's a drag and drop experience.

20:29We don't have to do any coding and if we then add a simple filter here.

20:33You go to category, for example, just a simple filter.

20:36We change the filter to let's see technology.

20:39Oh, Sean Miller is still the king.

20:41That's the new office supplies.

20:43So you see can long steel now.

20:45So it's dynamic tight dynamic piece of text.

20:48Just from AI, you just drag it in.

20:50Done.

20:50Yeah, it works

20:51And I can then play again with it.

20:53Yeah, and this is of course addressing that that need for so when you do conversational analytics, you have your findings, your insights.

21:00How do you share them with your colleagues?

21:03And there is no need to s to ask the same question over and over again.

21:06And if you have that insight, I think this is a great way of sharing it.

21:11So you can create a dashboard, add more to it if more is needed, and then just share embed where you want to embed it to you have to share the findings with with somebody who needs to see it.

21:23Amazing

21:24And it's really fluid, I have to say, I wish Tableau could uh learn a thing or two from this.

21:31They've had twenty years and like we people have been crying out for this stuff.

21:35You your Canvas alone, if

21:37Your canvas is just it just it to me it's proof that it's i if you just spend a bit of time thinking about it, these are obvious things, right?

21:46And you make it easy, you make even dynamic text in Tableau today

21:51This is a much better implementation because it just lets you get to the analysis first, then you can play around with anything after, change your analysis, the text changes.

21:59And that is super powerful.

22:00But yeah, I love it.

22:02It's really good.

22:03We are also quite excited about it and we will add so much more to it

22:06So at the moment this is still in preview.

22:09So the focus for now was to support these embedding use cases that Merlin started with

22:14But all of what Merlin just showed is just the start and this will uh this will grow weekly.

22:20And yeah, we have sh a lot of exciting stuff in the pipeline.

22:23And and by the way, this is also real, so there's no faking here.

22:26This is all real

22:27Working.

22:28I've got access.

22:28I can validate that.

22:30I've been in there.

22:30I'm doing it.

22:33I'll make my own video as soon as I'm competent.

22:35Can we do one thing, Melan?

22:37I don't know if you could hit

22:38Command plus just make this the icons a little bit larger.

22:42Oh.

22:43And then if we go back to the chart selection, I just wanna

22:46go through that list exactly this list.

22:48And so these are all the different chart types and the groups of charts.

22:52And so is it my understanding that a lot of these you've built into these are coming from the infotopic

22:58that say universe, right?

23:00And you've brought them into Trillip.

23:03But then as well, if I scroll down, I see extensions, right?

23:06So maybe if you talk a bit more about the groupings here and how that works.

23:10Yeah

23:11Yes, so we started with extensions in the analytics platform.

23:14So we said okay, we need to have a platform where we can also share our products outside of Tableau.

23:19For example, if you have a super tables embedding use case somewhere in a portal, we should be able to do that.

23:23So that's where we started.

23:24Right.

23:25And of course we've got a whole lot and this are by the way need to c not the correct names, but we got all kinds of charts.

23:30You show me more, like waterfall charts, sunbursts, funnels, all these things.

23:34But what we always did is we only build extensions we built with the extensions for Tableau that were not in Tableau.

23:42So we never build a line chart extension because you can do a line.

23:45So these charts were still missing from the platform.

23:48So that's why we added all the deep charts as well.

23:52So we also got line charts.

23:56So you can just add an order date here.

23:58Yeah.

23:59And a sales.

24:00And you will get a line chart as well.

24:03And then that is how build in the product.

24:06Yeah, yeah, it makes sense.

24:07That explains it.

24:09That makes a lot more sense.

24:10And that's why you have your own extensions at the bottom and then other components further down.

24:14Yeah.

24:15Here we go.

24:15Yeah, now it works.

24:16Yeah, good.

24:16There you go.

24:17Yeah, perfect.

24:19You if I show you this list, this is analytics, right?

24:22So if I want to get dates, I want to be able to look at continuous months, for example.

24:26Yeah.

24:27And then if it I only have data in twenty twenty one and nothing in these years and then here again, I need to see gaps that's continuous.

24:33If I go to month discrete

24:36Then I wanna see the months, right?

24:37So it's yeah that's all built in the platform by default.

24:40So we all got these kind of features in here.

24:42If I wanna look at quarter quarter day, I just see the quarters and I got the gap in here as well.

24:46And I wanna be able to also fill that up with null value, stuff like that.

24:50It's all by default in the platform.

24:52And there's more.

24:53So if we, for example, look at the data here, we actually are able to calculate things as well.

24:58So we can create a profit margin or profit ratio.

25:02And I can just say profit divided by sales.

25:07Right?

25:08Calculation works.

25:10But we can do more.

25:10We can also say percentage of total.

25:14So we got also a how do we say it?

25:17It's my window sum of sales.

25:22Divided by sales, for example, or the other way around, what you want.

25:26We got percentage of total.

25:28There's a whole list of calculations you can use.

25:30We can actually use the min max average counts, health statements.

25:35There's string calculations you can do.

25:37And there they've got window accurate calculations like the window sums, averages.

25:41All that data.

25:41Yeah.

25:42Yeah.

25:42Running sums.

25:45Ranks, everything is in here.

25:47And I think people think that those calculations are exclusive to Tableau because they're only ever used Tableau.

25:52But actually a lot of these are just functions you'd run through SQL.

25:54So they're not proprietary.

25:56They're just

25:57SQL functions that are done in a specific way.

25:59If you open up Tableau, run custom SQL, the performance recorder, you'd see the SQL the Tableau generates to do these functions.

26:06So these are all exactly these are all just based on the because we're just connect to the databases, these are just based on connection, the SQL uh statements.

26:13Yeah.

26:13And yeah, exactly.

26:14Yeah.

26:15So that that means we somehow still have this differentiation between like basic chart, like bar chart line charts and extension.

26:23And I totally get your question because why would you have an extension in a new platform?

26:29Yeah, this is trying to bring our history together with our future and that's why the these extensions are uh also available here.

26:37Yeah.

26:38I think we will also look into what we call TrueLib Studio.

26:41Oh yes.

26:42That is a a a studio to generate new extensions.

26:46Then of course

26:47Again, looking at Tableau, which is our primary focus at the moment, but also something that we can use in analytics right away.

26:56Yeah, and here we try to address the need for for creativity and speed and creating new extensions that we see everywhere in the community.

27:05With yeah, providing governed governance and security.

27:09So everything that gets generated here is behind our authentication, beyond our security.

27:15And I think that makes it really powerful and let's see if we can build something.

27:20Yeah, sure.

27:22Just gonna read all the page ones.

27:25Yes, very good.

27:26We're there.

27:28Alright, so I got a prompt ready because I'm not gonna type this in a recording.

27:32Yeah

27:33But this is again a a real example.

27:36And what we want to build is we want to have a tableau extension, which we can also use in analytics, of course.

27:41But we want to have a tableau extension where we show the US grid as a the US as a grid

27:46W with different shapes, which I want to be able to also choose what kind of shape.

27:50And yeah, that that's actually my prompt, but then a bit more text

27:54And with this prompt, we will again use our tooling to build a fully interactive Tableau dashboard extension or Tableau Vish Viz extension.

28:04That can interact with other charts, that has tooltips, that has settings by default.

28:09And so the AI understands, okay, there's a settings window I can use, and I can add settings to it, which I can reuse in the product.

28:15So a full fledged, full viz extension based on our imagination, what we want it to do.

28:22Yeah, and this is of course leveraging our experience.

28:24So we've built quite a few extensions and

28:27We know the building blocks, we know the APIs, we know how to set up such a thing.

28:32And this is now happening and being leveraged in the back.

28:35And yeah, you can follow along.

28:37what what our AI assistant does and it should result in a in a good uh result, which it doesn't.

28:45It kind of does, but it's playing with the positions of the states.

28:53It will show it how you say it with dummy data.

28:55So you have a preview here because this is of course not in Tableau, so that's the dummy data.

28:59Yeah and if we then go to let's say go here to Tableau I can actually grab that file

29:09C US script map, that one.

29:12There you go.

29:15There we go.

29:16Again, here it renders better.

29:18Yeah.

29:18I can go to location.

29:20Let me add the state to it.

29:23And a measure like SILS.

29:25Very cool.

29:26Beautiful.

29:27Interactive one.

29:29And so and what I find super cool about this is be and that's why I think

29:34Using a platform like ours is beneficial is because we got the knowledge of how to build these kind of things.

29:39It's correct.

29:40Not a one-shot to to open the eye to chat, build me an extension, you hope it works.

29:44This one just it works.

29:45And the the thing is that

29:46It has a settings wheel here as well.

29:49I can just and all these settings are the I think okay, I have to add some settings, these things will be handy, these things will be handy, and I can actually ask it to add more settings.

29:56So I can change, for example, the shapes here, which is an out-of-the-box feature of this one-prompted extension.

30:03I find that quite mind-blowing.

30:04Yeah.

30:05I can change the colors to let's say orange, for example.

30:08There's things about scales.

30:10I can change the spacing, there's labels, there's even formatting.

30:15So I can do state plus value and then make the value for example abbreviated with an or like uh with 1.

30:222 case.

30:22So you see it's it's cancel case.

30:25And I can add a densimal to it if I want.

30:27It's all out box.

30:29And I think Florida should sit a little bit more to the right.

30:32Is that also possible?

30:33Yeah, I understand this is I think

30:36I can also just drag it.

30:38Oh wow.

30:40That's so cool.

30:41So you can move Alaska and Hawaii if you wanted to.

30:44I was gonna say they're tastefully put there at the bottom because obviously they don't quite sit on the map, but

30:49Yeah, you could move them if you wanted to.

30:51Yeah, yeah.

30:52And let's just make this our uh our demo, put it on a dashboard here.

30:58So this is a tableau dashboard again.

31:00And show you also the normal map of Stableau of course and make this a sheltered map

31:10Oh, you're using show me.

31:12Yeah, of course.

31:13There you go.

31:14There you go.

31:15Let's also put it on the dashboard.

31:17Make it a bit bigger.

31:20So yeah, we'll see if I add the filter features and click on New York, that it also filters, right?

31:26It's an interactive extension that builds.

31:30All by default part of the platform.

31:32And then and this is what I really if we then go back to to analytics for example, so we go to analytics and we create a new app.

31:41It's refreshing my session.

31:43Yeah, here we go.

31:44Create a new app.

31:45Connect to data.

31:46So this is a new session of analytics.

31:49Going scrolling down.

31:51It should also be in this list, is it?

31:54You first have to publish it.

31:55I know I'm squat, yeah.

31:58Very good.

31:59Zipping through.

31:59There you go.

32:01We have to publish it, so I have to go back.

32:03You say I have to go back?

32:05It was I think in the um your script map this one.

32:09I have to publish it, you say?

32:10Yes, publish.

32:11Ah okay, I didn't do that.

32:12Hit the publish button.

32:13Here we go.

32:15Now let's go one more time back to the dashboard, create a new dashboard, connect to the data.

32:21Done

32:22There it is.

32:24Perfect.

32:24Cool.

32:25Look at that.

32:25Can just drag in that one again also here.

32:29And now we should be able to use a province with sales.

32:35Uh oh, sales.

32:37We got a dimension.

32:39Okay.

32:39We have to do one more prompt on it because the T-Rex encoding says this has to be a dimension and it needs to be a measure.

32:45So fine.

32:46We should go back to the and like change the encoding, maybe.

32:49Adjust it.

32:49Yeah, but that that's expected.

32:51That's part of the iteration.

32:53This is really incredible.

32:54Like I love how much the extension gets out of the box.

32:58immediately because these are typically afterthoughts.

33:01This is literally how you waste credits in Claude or ChatGPT because you build something, you great, and then you go back and you do it again and you just

33:10Every time you're doing that in one session, it's sending the whole entire history of the previous thing again.

33:14And that's how you get token maxing.

33:17But to be able to do it efficiently here also shows that it's a more efficient way.

33:23of using those tools to get an outcome because you've got all the I'm assuming you've got the context of how the the extensions you've built have worked

33:30So you get to that first solution faster and you stop messing around in these tall burning credits, which then justifies their use more generally, right?

33:38Which is nice.

33:39And so I think the big thing is that we have this whole framework ready and it only has to inject like the middle part.

33:44We got the whole starting ready with the extension association, all the functions are there, everything, it all knows how to use it.

33:50You the AI builds this part and the out part, the last part also what we have.

33:54So it can just there's something in between that the AI does and we got the whole framework around it.

33:58So it doesn't have to come up with

34:00Embedding your extensors API, making a file with a typed file with all the methods, for example, that it's all done.

34:05It's all there already.

34:06You can just use it.

34:07It makes super and that makes it also reliable.

34:10And and of course it's also security is also a thing that we implemented here.

34:14So it it will also there's a whole list of a blacklist of things that cannot do, for example, like scripts.

34:21Or post messages or retrieving data externally, posting data, that kind of thing is not possible.

34:27There's a whole code scanning after the creation.

34:31Okay, is this a save extension?

34:33Yes, then you can use it

34:35And and that's also part of that whole platform building.

34:37It is safe, secure and it is also for the big companies, it's fully on premises available as well.

34:44So you are fully in control of this.

34:46What we do support for Visi Studio is whitelisting uh charting libraries.

34:52Okay.

34:52Like you can whitelist D3, chart.

34:54js, yeah, so all the

34:57All or the known libraries.

34:59So again, we do not want to reinvent the wheel.

35:01If there is something you can use, then just use it.

35:07You are able to generate also the three charts for example.

35:10Yeah.

35:10Yeah.

35:10So if you have a high charge license, for example, you can use it here.

35:13And then only if these scripts are here, then it is actually able to use these things.

35:19So yeah, what we are built what we were building with the examples were just native SVG.

35:24But imagine if you can start extending it with these kinds of scripts and tell the AI, you know you have to build a a high chart.

35:31It will also do it.

35:32Yeah.

35:33Yeah.

35:34Or whenever you want, actually.

35:36Yeah.

35:36And of course, that's maybe you didn't even say it.

35:38It also has write back out of the box.

35:40Oh you can also ask it to build a write back a write back table, for example.

35:46Or any write back use case and then it will connect it will also have the understanding of how our writeback engine works because there's also an API that it can use.

35:54So it can

35:55Write back data, it can retrieve data again and show it.

35:58So you can have we have examples of bar charts that you can drag on the bars and to say that I want to adjust the bar, make it a bit higher, add a little reason to it, save it.

36:06Well you could save back to the database with my name attached to it, and that's also reflected in the bar chart again.

36:12And it's just one or two prompts.

36:13One prompt and a follow-up one and done.

36:16It works.

36:16It's funny because uh Matt Miller was on stage.

36:19I think of the last conference at the end they showed this capability that was basically like parameterized charts or something like that.

36:26And it's it was exactly that.

36:27But the way it was solved was like, ah, this has never been done before in tablet.

36:31And in my head I was like, but I swear right back already exists.

36:34Come on.

36:35So yeah, this is very much that.

36:37And I think the other thing is you as you said, it's here, it's live, it's working, it's built.

36:41It's not something that we hear about six months ago and we've still not seen it.

36:45two and a half years later, as is the case with a lot of features sometimes in the tablet systems.

36:50So and we actually write to a database.

36:51So we're at them right into local storage.

36:53It's a real database write back action.

36:56So yeah.

36:57That's that.

36:57And it's fully out of the yeah.

36:59So that but that's what we've been doing for years, of course.

37:01And all that knowledge is being put into this platform.

37:03So it's out of the box working in this platform as well.

37:06Super cool.

37:07Amazing.

37:08Maybe do you want to show the MCP uh integration?

37:11Yeah, sure.

37:12Yeah.

37:13Yeah, so

37:14Because this is built from the ground up, it has been built AI first, or how you can say it.

37:19API first, AI first.

37:21So everything

37:22All the API calls that are there are also tool calls for an What does that mean?

37:27But it actually means that I can actually go to ChatGPT and talk to Trulib.

37:32So we already are thinking about you can use this hapless.

37:35So the old interface, you can you will be able to use this without the interface as well.

37:41So maybe we can again ask something like what can we do?

37:45Let's make it a bit bigger to improve profitability in region central for my Superstore MySQL.

37:50And I have to say to you my true lib demo environment, of course.

37:55So what this will do is it will use all the MCP tools we have available.

37:59It's around 300 already.

38:01It will figure out like, okay, how can I connect to a data source from Trulib.

38:05And what are what is the the data engine?

38:07So what are the query engine?

38:09It will use our query engine, which is also an MCP tool, and start running multiple queries on our data engine, figure out an answer, and then of course the beauty of an LLM.

38:19Make a beautiful answer outfit.

38:21Yep.

38:21And there we already got an answer.

38:23There you go.

38:24Perfect.

38:24Yeah.

38:25And this is all on our

38:27MCP tooling.

38:28So there's no other application here.

38:30This is all true.

38:31Yeah.

38:32So we already got information here.

38:34So it's something about tree tables as a special case.

38:37So we all know that this is superstore data.

38:40But we can say table looks suspicious, provide a visual analysis what it's writing this, for example.

38:45And it will again run multiple queries, analyze the data it gets.

38:49And come up with a with a visual analysis of this.

38:53So it's creating charts.

38:54That's the beauty of LLNs, of course.

38:56Ah

38:57It's very fast as well, like for the And exactly that's what I want super chain.

39:02Yeah, yeah.

39:03Super fast.

39:03And we get already answers here, which I think is

39:06Yeah, this is very cool.

39:08Yeah.

39:08Yeah.

39:09In in a next iteration we will be able from here also to save these insights directly into an analytics dashboard.

39:17Well if we have it once, you wanna keep it, you wanna share it, and this is what uh CP will allow us to do.

39:22Yeah.

39:23And it's save it in your studio or something like that, right?

39:25Yeah.

39:26Yeah.

39:26Yeah.

39:27But so did this whole platform is

39:29AI first or how you would call it.

39:31You can use it in your favorite an AI tool.

39:35Amazing.

39:36So I have to ask the question.

39:37You are doing a lot of

39:40really innovative foundational unique things with the platforms.

39:46I guess you you don't

39:48have to use Tableau.

39:49Like a customer could come to your platform without Tableau and still use it, right?

39:53Because you've got all the underpinnings in order to be able to host them essentially.

39:57They can connect to their data source.

39:59It it doesn't have to be in Tableau.

40:00It can be in something like Snowflake or whatever.

40:03You've got your charts, you've got your MCP, you've got your Visi tool.

40:07Like it's its own platform.

40:09Am I hearing that right?

40:10Tell me if that's a bad idea, but to me, that's actually the appealing thing here.

40:14It's a nice way

40:15to start to think about moving some of my abilities away from Tableau if I have Tableau, but I'm doing it again in a safe governed way and I have a bridge.

40:23because you've got that kind of capability.

40:26Yeah?

40:26Yeah.

40:26Yeah.

40:27So the way we position it.

40:28So the answer is of course, yes.

40:30Yeah.

40:30The way we position it is is that we are basically a like a

40:37Um an addition to Tableau?

40:39Because Tableau is like an incredible analytics platform that has a lot of amazing features that has been working for what you say 20 years.

40:48We bring these new capabilities like these these fast analytics dashboards, the iterative, that Visi Studio, the generative extensions, generative AI, generative apps.

40:57We bring that to Tableau as well.

41:00And so customers basically have the best of both worlds.

41:04Yeah.

41:04And combine it because this is all running on Tableau virtual data source.

41:08sources as well.

41:08Yeah.

41:09And that's also where we want to launch with.

41:10So we will launch with double virtual data sources.

41:13Yeah.

41:13And after that, of course, based on customer demand, we will extend.

41:18Yeah.

41:18Yeah.

41:18But the whole data is already built for right back.

41:21We already do years.

41:23writing to Snowflake, Databricks, all the popular ones, SQL Servers, all the ones.

41:27So we our data engine is already using that, all those data sources.

41:31But we start with Tableau Visual Data Sources here to be like the ultimate add-on and have those generative applications right there in Tableau or besides Tableau and embeddable in Tableau.

41:43Yes.

41:44And of course we are still very much committed to Tableau.

41:48We love the tool.

41:49It's our history.

41:50And supercharging Tableau has been our mission for many years.

41:54And also with TrueLive we still do that.

41:56You mentioned I think it's called Table Souls.

42:02Yeah, that that is also doing right back, but again, we I think we come with an experience of

42:09I think six, seven years for right back extreme.

42:12And yeah, we've seen a lot and and we bring this capability to the Tableau ecosystem and we're still here to do that.

42:19But we also see how AI can really accelerate also the use cases because setting up a write back application is something already and it

42:29You need some knowledge about Tableau, about data modeling, to bring that data back in.

42:35And with the platform and with Studio and Visi AI, we can help our customers and any Tableau customer

42:43to build out these write back use cases with just a prompt and have a production ready product in just no time.

42:51Yeah.

42:53Yeah, it's good.

42:54I'm really impressed actually.

42:55Like I it I say this because like I sat with Clams just last week.

43:00We spent an hour going through the platform.

43:02I felt like I understood it.

43:03And then we've gone through it again and I'm like, wait a minute.

43:06It's just so much more.

43:07It's incredible.

43:08It's really cool.

43:09Still showing like a little bit of it.

43:11Maybe there's a lot of course there's much more to come.

43:14Yeah.

43:15But I think one of the cool things as well is that we got a full AI trace as well.

43:19So if you talk about transparency is you can actually see like this is my prompt and what is the I actually doing to get this

43:26Prompts very full trace here of all our tool calls and how it gets to its answer using all our analytical capabilities.

43:35So

43:35Transparency is one of those things that we really worked on.

43:38You see all the tool skills and how it's building visuals, how it's analyzing and giving it to the city.

43:43Everybody leans into this deterministic

43:46Exactly.

43:46It's a viewpoint, which is if I was going to try and do this, could I repeat it?

43:52And what are the steps I need to repeat it?

43:54Yeah, this this trace essentially gives you that.

43:56But I also look at this as hey, I like this pattern.

43:59I like this way of doing this, actually.

44:02So when in in this specific business domain, this to me is like a very good way to always think about the problem and let's say insurance or finance what you have.

44:11Certain tables set in queries that are used by everyone.

44:14It's almost like creating an a different version of a data model.

44:17Instead of creating a data model

44:19You're actually just creating, let's call it a query path, right?

44:22That is the same thing, but it leads in a deterministic outcome in terms of how the analysis is always done, which is super powerful.

44:28Yeah.

44:29Yes, and what we are doing here is rolling over what we've learned and and done for years and and write back since we are very strong and regulated industry.

44:37So auditability.

44:39is very important to our customers.

44:41So they want to trace every write back action in the past and how they are also able to trace

44:47any AI action in the future with this platform.

44:51And this is all c this is customer data and remains custom remains customer data.

44:57Also this is stored in the repo and yeah our customers can use this data also to learn about how their users use the platform and yeah and improve the way of working with it.

45:08So that's I think well very powerful.

45:10Yeah.

45:11And there's two features coming which we didn't have do not have in the product yet, but one is inside.

45:16It's right here.

45:17I'm not gonna click it yet, but it's yeah

45:20So w coming soon.

45:21Have this this AI this engine now that AI is using and stuff.

45:26And what we found out is if we like extend it, we can actually let it mine our data sources, so data mining, and come up with insights.

45:34data insights and explain them to our end users in a user-friendly way.

45:39So little cards that are clickable.

45:42And by combining all our tools we have, we already got prototypes of this.

45:47It generates insights.

45:48Let's say I'm the CEO of this company.

45:50Yeah.

45:51I'm the CTO, but the CEO.

45:53And by clicking on different things, it understands okay, Merlin is interested in these KPIs because he's clicking on them multiple times.

46:00It's building a profile of me and with the mining it will generate interesting insights for me.

46:05And those insights also have a analytics dashboard.

46:08And that's I think what the cool thing is.

46:10You can click the insights.

46:11It shows for example like tables just as a downward trend of profit is with a little line chart.

46:16Click it, you get an analytics dashboard with more details about it.

46:19So it also generates an analytics dashboard because it knows how to do that.

46:23And then from there you can materialize that.

46:25So you think this is a really awesome dashboard.

46:27Materialize it, save it, make it your favorite, and then so you've got this whole tool.

46:31That's one thing.

46:32The second thing is communication layer.

46:34So that's the part we're adding next.

46:36We got Mail Scheduler, which is our application where you can build pixel perfect reports

46:41with tableau dashboards that it sends to your email or to teams or just plaque with PDFs personalized.

46:47So if you have 300 stores, for example, as a company, it will send personalized

46:51Yeah.

46:52A report to every manager every Monday morning, for example, how the store is doing with precise message, inline images, or sent it to Teams or Slack.

47:00That whole application we're also gonna integrate into TrueLib.

47:03So you can imagine what kind of power that brings to this platform again.

47:06Like subscriptions, sending out messages proactively.

47:09Those insights can be got going to your email or to your teams or something like hey

47:14Yeah, th th you have to look at this kind of thing.

47:16So it becomes more of an action platform as well, or like even more action platform.

47:20So that's coming.

47:20These two things are coming as well.

47:23Those are logical next steps as well.

47:24That makes a lot of sense.

47:26It's exactly what people will come to you after you should have let them into the platform.

47:29So it makes sense to be building those already.

47:31Yeah.

47:31Yeah.

47:32Yeah.

47:32And it already works, but we just have to now make it

47:35Reliable.

47:36Production ready, yes.

47:38Yeah, exactly.

47:38Exactly.

47:39The prototypes are there.

47:40Yeah.

47:40Amazing.

47:41Before I forget, why Trillip?

47:43Why the name?

47:44I didn't ask earlier on.

47:46Yeah, TrueLib is is a little bit of a play with our heritage.

47:50So we are a Dutch company, but it's of course also a play on what we provide with this platform.

47:57So trusted analytics, trusted AI that's

48:01one very important part.

48:03And if you look really closely at the name, so there is a very popular flower in the Netherlands that that that is connected to this name.

48:13Tulip.

48:13And yeah, we think it

48:15It reflects in a nice way what we are up to, where we come from.

48:19And yeah, again, that that trust and governed analytics is is a very important part of our platform.

48:26Yeah, and that explains the logo of the icon.

48:28I was trying to figure out what is this icon?

48:30Okay, it's a tulip.

48:31But there you go.

48:33The little Dutch joke.

48:35Yeah, it's good, it's good.

48:36Yeah.

48:37Amazing.

48:38I'm really excited to use it myself.

48:40Obviously, I've got access.

48:41I've started playing around with it last week.

48:43I spent some time with Klaus setting it up.

48:44It's funny that the most challenging thing of setup was actually all the other platforms, right?

48:49It was snowflake, it was tableau being difficult.

48:52But actually once we got in everything was really smooth.

48:55And I like also the

48:57It goes without saying, and I know it's not something we've talked about, but the the UX that has gone into this, the user experience and the thought is really good.

49:04And I think it shows a little bit of your heritage as a team and just building stuff for people.

49:09You've thought about these things.

49:11And

49:12They look easy and they almost you almost don't notice them.

49:15They're not something that we'd spent time, twenty minutes talking about.

49:18But when you use a tool that doesn't have those things,

49:21it really does frustrate you and it really grates upon you.

49:24So we talked about simple things like copy and paste, the canvas, being able to move things.

49:29That comes from an understanding of UX and it's all the way through the product.

49:32So

49:32I'll say that on your behalf because again I know the two of you are humble, but you know, throughout the whole entire platform, even the configurations setting things up, you really thought about the the user flows, the things they need to do, and also

49:46It's not too intimidating, right?

49:48Like the there's a sort of a stepped process throughout the whole product, which I really say.

49:51Yeah.

49:52That's of course a bit also the beauty of a fresh start, right?

49:56Yeah, true.

50:01then all this experience has gone into this n this new platform and yeah just being able to start on a green field and make the product

50:11Yeah, according to these experiences and how we would how we would want it to be.

50:16That was really amazing and fun thing to do.

50:20And yeah, yet now we want to bring it in front of our customers.

50:24So

50:24Everybody who sees this is is able to sign up for a trial.

50:29Very easy, very easy setup.

50:31Tim, we went through it on Friday.

50:33And yeah, let us know what you think about it.

50:36And we're really happy to get your feedback in.

50:39Amazing.

50:39Amazing.

50:40Thank you for coming to the channel and sharing people the product.

50:44I guess maybe I'll give you a bit of time just to set the expectations on when it's going to be available, what people can do, when, and then yeah, how to find out more

50:53Sure, so we are planning to launch the trial.

50:58So let's say that you can play and

51:00play with it this month.

51:02So it could be next tomorrow, could be end of the month.

51:05That's still a surprise.

51:06Yeah.

51:06But we need to have a solid, stable and secure platform.

51:10So the moment we reach that

51:12It's going to be there.

51:13So SAS first on premises comes after.

51:16Okay.

51:17And then after we we talk about a a month after launch of the SAS version.

51:21Yeah.

51:21That's cool that you're thinking about on premise actually, just to add, like, does that suggest that someone could take Trillip and self-host and then manage it within their enterprise session?

51:31That's incredible.

51:31Yeah, and then that's a big of a big part of a company still do this because they are in these heavily regulated environments.

51:37They only want on-premises software.

51:40And that's actually why we do this.

51:43Yeah.

51:43To also solve that question.

51:45And then that requires quite a lot of engineering, as you can imagine, besides just AS platform.

51:50It's quite a lot of engineering going into an good version, on-premises version.

51:55Yeah, and that's why we also allow people to connect to their local LLMs.

51:59So that part is also not a black fox that you can of course use.

52:03open AI or anthropic, but you can also use your local LLMs, also addressing this need in regulated industries to keep things in the organization.

52:13Yeah.

52:15Amazing.

52:16Great.

52:17Thank you.

52:18Thank you for coming on the channel and showing us the product.

52:21Once again, every time you come on, you always show something great, I think.

52:25And this is yet another.

52:26repeat of that.

52:27I'm looking forward to trying it.

52:28I'll make some videos obviously as I start to get into the part for myself.

52:32And yeah, if you've got questions, people who are watching, if you've got questions in the comments.

52:36Please put them in the comments and let us know and I'm sure we'll we'll get these out to Klaus and Melin and they'll be watching the video as well.

52:42So they'll they'll catch some of those and maybe respond.

52:45All right.

52:47Thanks for having us, Tim.

52:48No worries.

52:49Absolute pleasure.

52:50Oh yeah.

52:51And by the way, if you want to know when we launch, join the wait list on truelip.

52:57com

52:58Amazing.

52:59Oh nice.

53:00Nice splash page.

53:01Amazing.

53:02Yeah.

53:03Good.

53:03I'll put that in the description as well.

53:05I'll put it as a QR code throughout the video so it's easy to just scan on your phone and join as well.

53:10So yeah.

53:11Good.

53:12Thank you for having us.

53:13No worries.

53:14Absolute pleasure.