Fiona (00:08)
Welcome to UnDUBBED the podcast that's unscripted, uncensored, and undeniably data. I'm Fi
Sarah (00:14)
And I'm Sarah. TC26 has wrapped and the conversations coming out of it are genuinely exciting. Not just product announcements, but a real shift in how Tableau is thinking about the role of data in the age of AI.
Fiona (00:31)
Today we're talking to someone who is involved in that conversation. Will Pitzler is a director of product management at Tableau, and he works on the integrations that get Tableau insights out of the platform and into places where people actually work: Teams, Slack, Google Workspace, and now Excel spreadsheets.
Sarah (00:53)
But to understand why that last mile matters so much, you need to understand what's being built underneath it. And will is across all of it. So today we're going end-to-end the architecture, the AI story, and the very practical question of how insights actually reach the people who need them.
Fiona (01:13)
There is a lot to get into. Before we dive in, please like, subscribe, and share this episode today with a data leader who's trying to figure out what AI and analytics actually means for their team. Will, welcome to UnDUBBED
Will Pitzler (01:29)
Sarah, Fiona, thank you so much for having me. It's a pleasure to be on. I've seen the podcast out in the wild and I'm glad I get to finally be on with you both. So thanks for having me.
Sarah (01:38)
We're excited to have you here too, Will. Before we get into the thick of Tableau, tell us who you are and what you actually do day to day. Because your title is Director of Product Management, but you clearly go very deep into the craft of this. What does your world look like?
Will Pitzler (01:54)
Yeah.
It's a good question. We wear a lot of hats. I guess I would say that I've been at Tableau for almost thirteen years now, so a long time, and I've been through a several different roles, both on the go to market side, the strategy side, now on the product side. And really what our job is is we sit at the intersection of our product and our roadmap with our technology partners.
So our team predominantly any any of our tech partners that you might have heard of across the data stack, whether you're working with DBT or Microsoft or Google or Databricks, how do we deliver products with our partners to the benefit of both of our our joint customers? And so that's really what we do. And there's lot of work that goes on underneath the hood, but that's sort of our primary responsibility.
Fiona (02:35)
Yeah, I I mean I've always loved this the sound of a job in product management. It sounds really cool to go deep into different areas and researching with people and talking to people about what's really important, but being able to translate things. Is that is that the right kind of take in product management? Or have I got that a little bit off kilter?
Will Pitzler (02:54)
I think that there's different there's a spectrum of product managers, right? You on one side you have technical PMs, you have people that work strictly in architecture, behind the scenes. You've got on the other side of the equation, maybe a go-to-market PM who's predominantly responsible for not just the product itself, but how do you take the product to market and and how do you make sure that it's being adopted by customers? We're probably right in the middle, but essentially it's you know, where's the market going? You know, what are customers asking for and how does that
Fiona (02:57)
Mm-hmm.
Will Pitzler (03:22)
triangulate with the strategy of our platform and how do we make all those things work. So it's you work cross-functionally. again, I mentioned earlier you wear a lot of different hats, but at the end of the day, it's just how do we make our customers successful? And that's really what excited me about the role and kind of why I'm in the position that I'm in.
Fiona (03:38)
Yeah, nice. well, there will be a number of leaders listening to this episode today who have heard the term data driven for 20 or more years and they've sat through a lot of conferences and a lot of announcements. What's your honest pitch for why Tableau Conference 2026 felt different and not just a new product cycle?
Will Pitzler (04:03)
Yeah, I think that's a that's a difficult question. At the end of the day, I think that, you know, a data driven company, a data-driven organisation or becoming data driven really takes people, process, and technology, right? It's the intersection of all three. How it relates to Tableau Conference, I really think that it kind of went back to what makes Tableau Tableau.
Right. We're appealing directly to the data people, the practitioners, the people that are on the ground. again, as you scale as an organisation, you kind of go from that product-led growth, getting in the deep in the weeds to that executive level presence. It's kind of finding that balance. At least from my perspective, I think that's where maybe the conference took a little bit more of a pivot than what it was the last couple of years. Is we really went back to the practitioner. We went back to the people on the ground.
talking about real tangible things that are in production or being released and really focusing our efforts there, which in my perspective and I'd love to get your opinions, is sort of where it took a little bit of a turn from prior years.
Sarah (04:59)
Yeah, and I definitely think, after Devs on Stage and things like that came out, it was like there was a buzz amongst the data fan community of yes, they're they've gone back to that problem that we've been trying to solve for a while, with composable data sources. And then one of my favorites, which is the the Viz chart layers, looking forward to some of that stuff coming in because it does help solve real problems that analysts have working in Tableau.
Will Pitzler (05:25)
Yeah, and I think that goes back to the heart of what we do and what Tableau, like people love to use Tableau. There's a lot of different tools out there if you just distill analytics down to building a visualisation or building a chart, you're missing the point. It's all those granular, iterative pieces that work together. And I think that that's the beauty of the VizQL engine that Tableau is built upon that continues to excite, practitioners like yourselves. You have deep expertise in Tableau, but these little additions really just provide such a benefit to the people that use it.
And make their lives better. And ultimately I think that's what we strive for. And I think that's what we've really gotten back to from a from an R&D standpoint.
Sarah (05:58)
Yeah. And it has a real knock on effect to the executives that are reaping the benefits of of those developers being able to do more and potentially build faster as well.
Will Pitzler (06:08)
Yeah, absolutely. And I think that the developer term that you used, I I love that term because it spans such a wide gamut of of of user persona, right? Are you appealing to like a traditional analytics engineer that's becoming more modern today? Or are you appealing to someone that has no data background that becomes an analyst because they love Tableau and they can make sense of data, right? And I think that's the beauty of Tableau is just it spans quite a spectrum of users and user personas. And I'm sure that, you know, you two both
have worked for a long time in Tableau and find this in the companies that you work with as well.
Fiona (06:37)
Yeah, I and it's quite interesting hearing you talk as well around that span because you've obviously been around at Tableau for such a long time. You've seen the shift you would have been coming on, I guess, when Christian was still leading the company
he was around obviously when you started out and then you've gone through the massive shift with Adam, getting the company ready for acquisition and then going through through the sales force side of things as well. And Sarah and I haven't been at conference for many years. We took a break with other work priorities, but now that we're back in the fold and really focusing in the Tableau space, what we'd heard.
in the past few years was there'd been quite a shift. But when we went in, we're like, wow, they're they're doing so well. I love, the direction that everything's taking. We both came out feeling really energized about the direction.
Will Pitzler (07:32)
Yeah, that's exciting to hear. And I think there's a lot of noise in the market. I think that it's such a different place than it was two years ago, five years ago, ten years ago. Any acquisition's hard. I think that there's just growing pains and figuring out where you fit. But I think that leadership has done a really good job of listening to the customers at the end of the day and figuring out where Tableau sits as part of Salesforce. But ultimately, again, we stay true to our mission, see, understand, and then act on data, right? I think that's sort of where we're going. But I think
the the primary focus is our customers and making sure that we continue to delight. And at the core of it is Tableau's user experience, which to this day remains just different to everything else in the market. And so I think the most refreshing thing, and again, to your perspective, it kind of echoes what you're saying, but I still just, you know, it just blows me away that every time I go, there's the fandom around a technology product. And like
I I still just am pretty blown away, but about how excited our customers get. There's such a sense of community and a bonding that goes beyond just a technology. But ultimately, I think that's rare in our space, right? I mean, everyone has a preferred vendor you use a product because your executive buys a product, but at the end of the day, it's it's not just good enough. It's it's hey, I really love to use this product. This has elevated myself and my career, and ultimately the organisations that take advantage of Tableau I can be ultimately become, going back to your term.
More data driven.
Sarah (08:53)
And it really feels like Tableau got its groove back in TC26 we were, you know, we hadn't been there since TC19 And it feels like, you know, a lot of people were saying it was similar to that. So it it's it's gone through some hurdles and and kind of bounced bounced back. Okay.
Will Pitzler (09:01)
Yeah.
Yeah, and I
think we had we had the COVID pause there for a while. And then we got further embedded into the Salesforce ecosystem and we're speaking to a much broader audience. If you think about, the size of Salesforce is 80,000 people, Tableau is, five thousand. So it's a it's a small piece of the bigger puzzle. And so, you know, think about addressable market. Where does Tableau sit? Where should it sit? Who are we appealing to? What customer conversations are we having? And I think again, we've pivoted back to we know who we are, how we benefit.
both the traditional analytics or line of business, but also the Salesforce customer and finding that balance between two very disparate user personas and different audiences of the products.
Fiona (09:50)
And it was really exciting to see you bring some of the insights into Sydney last week and sharing them with the local community who didn't have the pleasure of traveling to TC. And I wanted to actually go really deep into the weeds on something.
The integrations with LLMs where they looked at the accuracy of pointing LLMs and creating SQL. And there was a 6% accuracy when you pointed a raw LLM at a complex enterprise database. Now I know that obviously there's a lot that comes in underneath that context, etc., but more.
Broadly, for a leader who's been actively sold AI analytics tools right now, what should they take from that kind of number?
Will Pitzler (10:39)
Yeah, I think that's a that's a big question, Fiona. but first of all, it was good to be in Sydney. I'm glad that we're able to bring the insights locally. I think that's a big part of it. I think it just benefits so much to being in the room at TC. And so we do our best to take TC on tour. And I'm glad you you you found the value and customers were excited about it. I think we're still, again, just in the infancy of this whole AI game. And it's changing so quickly that candidly even myself in the product org.
Right, we kind of lose track on what's next and what how is it evolving. And you've got every vendor talking about AI, you've got, you know, every solution offering AI, everyone has different models, different approaches, semantic layer, no semantic layer. But really back to your point and what I was speaking to was just the a lot of customers are are getting their hands on these LLMs and getting these hands on these these AI tools, whether it's you know, Claude Code or it's you know OpenAI or any of the cursor.
And they said, well, why do I need something like Tableau? Let's just go straight to the database. And so this kind of the number that I was talking about.
And here I've actually got the slide here. I can actually pull this up. this is sort of going back to what I was speaking about and just kind of analyzing the challenges that customers have. And so this is the Spider 1.0 β Text to SQL benchmark. It was a test that was founded from a group of from Yale that they came up with and they tested the original.
Frontier models back in 2024 against this Spider 1.0 benchmark. It was essentially how well do these large language models reason and accurately answer questions in natural language, which is exciting, 87%. Again, for data people, we're saying that's not very good. But if you're looking at some of the complexity that was introduced, it scored really high. And so what they did is they went back and they said, β let's create spider 2.0. Right? So this is kind of, let's address these large language models and let's throw real-world solutions.
or real world data environments at these models. And if you just go raw LLM to a database, right, it exposes kind of this foundational flaw where it's great for prototypes. It's great, you know, they all these tools, it seems like they know what they're talking about, but ultimately it's just you're predicting the next word in the chain of events. And it really struggles with enterprise complexity, as we all do, right? You think about your customers and everybody likes to say it's as simple as just, you know, pointing to data and visualizing.
It's so much more complex as we all know. So the the real crux of the argument here was just if you're just going to use a raw large language model and point it at a database, you're going to struggle. And so ultimately, some of the things that they found hampered the reasoning or the accuracy of these models was really just the missing element of context. Is the LLMs, they see the data, they don't understand how it all works together. They don't understand truly the meaning behind the the word or the table or the
column and how it maps to everything else and not just that but how do you know how does it relate to other elements in your data ecosystem. And so that was kind of the the the gist of the story that we were telling and the importance of ultimately again analyst at the center. How do we not just provide data but define data, bring meaning to data, add context to data so that both humans and AI can β you know accurately develop insights. And so that's kind of the evolution of where we're going.
And Tableau's kind of on that journey as well to help our customers as we strive for that.
Sarah (13:57)
Yeah, it's it's so interesting, right? And I think people do forget that context is so important, like and how that's stored, how the memory is accessed. I was listening to a podcast the other day and now it's all about, how you set the goals for the AI and it's just that constant evolution of it's as good as the data that you put into it and the context that it has from outside, but when you start looking inside, that's when you've really got to hone it.
down and and explain all the idiosyncrasies you have inside your business. Otherwise it's gonna come up at something like six percent accuracy.
Will Pitzler (14:35)
Yeah, and I mean even think of the different models, right? If you want to bring your own LLM, sounds great, but they all operate very differently, right? So how do you create that shared language that works across different model types? There's still a lot of deviation how the different models operate, and you find disparities in that lane as well. So β again, I'm sure you're probably feeling the same. This this space is changing so fast, right? And new models coming out all the time, new techniques.
new ways to handle this. I think every organisation, you know, has a a North Star, but we're still learning along with our customers of how do we bring this vision to life. And so it's just a really interesting domain. It's kind of changing the way that we think about data and analytics, β entirely new possibilities, which is exciting, you know, both for you and your customers. but again, like we're we're there with you and these are some of the things that we're working to figure out as well.
Fiona (15:26)
Definitely. I mean, one of the things that always comes into my mind when we're when we're talking about pointing an LLM at a database environment, whether that's going to be, your normal enterprise data warehouse or if it's going to be something more like your Tableau environment and using MCP as the middleware. The thing that frightens me about this is about the
PII, data sovereignty, where things are getting processed. And there's literally nothing written about it. Like I I I I had to write a paper for one of our clients that explained, well, if you're going to use the Tableau agent, you know, this is this is what you can point back to and how everything, how your data is actually handled appropriately when it comes to using an MCP client.
There's good results you can get using MCP and querying through Claude, for instance, it really pops things up. But I'm not certain that I would want if I was the s chief data officer of a bank, that I would want to be pointing my banking data at Claude.
Will Pitzler (16:37)
Yeah, I mean I think this is sort of the kind of goes back to that original premise is we think of, you know, every executive sees this. Hey, we can talk to databases in natural language and we get insights. Like that opens up a whole new wave of self-service. Or we hear this a lot, we use cloud for prototyping where we build an HTML file and then everyone shares it around the organisation, which sounds great. But as you mentioned, there's really no governance there. And it's not repeatable. And there's no
There's no, you know, you'll built in lineage or even an audit trail. There's really nothing there. And so, when we think about the Tableau platform at least, that's really where we start, is sort of at the governance layer more backwards. So at least from the first party AI services, right, you have that level of trust, of compliance, and that you know your data is not being trained, et cetera. When you start to traverse cross-boundary and start using third-party models, and as you mentioned, MCP, there are whole new
think elements to consider. And again, we're investing pretty heavily in the MCP space and our APIs in general. And so you're going to see this continue to evolve as well. But you bring up a great point. And so it's it's how do you a lot of organisations think, how do we go quick? But at the same time, IT is pulling everyone back and saying, all right, well hold on. Like what is our governance processes? What are our you know, what's the PII regulations, regulatory compliance, all these other things that we have to think about. And so it is that pu you know tug of war.
We always, you know, business and IT and moving quickly, but I think that it's a whole new, whole new shape to that tug of war. And I think that you bring up an awesome point and something that again, you're right. Like the space is just so nascent in terms of just agents and and humans and how it all works together. And so what is the process?
Fiona (18:00)
Ha ha
Sarah (18:16)
well, most execs that we talk to know they have a data problem, but they can't always articulate it. And you described in your session as the pain of the last 22 years, semantic models that can't talk to each other, definitions that vary everywhere.
duplicated effort and you know that happens up and down the whole stack. Does that resonate with what you hear from customers?
Will Pitzler (18:45)
Every day. I think back we were in Sydney and went on site to I think there were three different customers that I went on site to the same week that we had the the the TC review in Sydney. And every single one had a different way to describe how they're applying a semantic layer. Everyone's choosing a different vendor, everybody is using a different that backend database, and everybody while pointing to a a very familiar kind of the nirvana that they want to get to, totally different processes.
Right? You're starting at the semantic. Are you doing ontology? Are you doing a graph technology? where do you start? And so I think that that's the big challenge that everyone has right now is that we know that especially AI agents, they need context. They need data that's grounded. But the ecosystems that all of our customers have are quite different. It's heterogeneous and heterogeneous in nature and it's never so simple. And so that's really the topic that we
we were talking about earlier was that open semantic interchange. And again I have another slide here. I'll just to something to speak to, I find that easier. β yeah.
Fiona (19:46)
Actually while you're bringing that up, so if
you're watching that if you're listening to this on Apple, you might want to go onto Spotify or YouTube so you can see Will's Slides.
Will Pitzler (19:55)
okay. I'll have to make sure that I speak to those listening. β but essentially what it is, β the idea of we β creating com between a bunch of different vendors, whether that's analytics vendors, that's semantic layer vendors, that's β data warehousing, data lakehouse vendors, kind of an industry neutral standard for how semantics are defined and shared across platforms. Right? Organisations are making such an investment in ensuring that
Again, I feel like date like the pipeline's being pushed even further left, but taking all of your data, making sure that semantics and meaning are prescribed and defined correctly to be able to be consumed by both humans and agents. But how do we make sure that that meaning traverses as data starts moving across platforms and doesn't end at the platform boundary? And so that's really kind of the impetus for this open semantic interchange. You can see all the vendors there, Salesforce there, others in the space.
β but you know, we're talking about a common SQL language that we all support and use. It sounds great, right? And I think we're working towards it. There's a council that's working on this problem. As you can imagine though, it's not simple. There's all sorts of discussions that need to be have. You know, what SQL variant do we use, what data models are supported. And so again, we're hopeful that this is a nice step forward to make the lives of our customers so much easier and less of the data duplication and improved governance.
But it's a long roadmap. We have a long ways to go.
Fiona (21:25)
And I know that we talked after your session in quite a lot of detail about this because it's something that I'm quite passionate about, trying to help people to be able to understand how they can leverage the investments that they've made in their semantic layers and other tools, be able to reuse that within Tableau. And trying to find answers in places proves quite difficult. And I suppose because it's
so new as well. But in the ongoing conversations that I've had, there's there's two schools of thought around how Tableau might be doing that. β One is that everything will be able to be reconsumed through through just a I think the way that you described it when we spoke about it was
Each vendor has a different dialect and I really love that. It's, you know, when you think about the the different dialects that even, you know, they might have in Chinese, for instance, there's lots of different ways of actually saying things for different regions. And it's similar for for data organisations as well. And how does that then get translated through a a solution like Tableau?
So one prevailing thought is everyone will use the same language and we will narrow it straight down. I think as much as I would love that to be the solution, I think that will be really difficult to get, in particular the organisations that are at the database level, because what's the what's the reason why they would do that? And in fact, the thing that you and I talked about was that the connectors will need to be able to
translate in a different way to bring that through into Tableau, which I think is probably more likely. β what I understand is that they're in great discussions on this council of trying to work out the best way to do things. Is that your understanding?
Will Pitzler (23:30)
Exactly right. β and so I I think I have this built into this slide here. Yeah. So I mean this is kind of the work in progress and I touched on this earlier, but to get to this point, β you have to get consensus across the board, which yes, twelve people in a room, getting consensus on anything this day and age is difficult to do. β there's a framework, there is a path forward. I think that there's there's a pretty good there's a working group, a council that continues to meet and have these discussions. But as you mentioned, I think there are both explicit and implicit
reasons for I don't want to say dragging your feet, but it's hard to do. Right. And you think about as Tableau, we've always said we're the Switzerland of data. We don't care where your data is if your data is in Databricks or Snowflake or DB2 or β you know, wherever you want to in prem on premise in the cloud, we're gonna work with you. The challenge is just this, right? If you have 50, 60, 70, 80 different connectors, they've all got different variants of SQL. They've all got different database drivers. And so there's a lot of work that needs to be done.
β and that goes for us also inheriting the existing semantic layers that we have today. And to your second question, if we have this uniform standard, awesome. We just update our existing connectors and it recognizes these metrics or these predefined semantic models and it can be inherited directly into the Tableau data model, no problem. Ultimately though, we're having to build one off point solutions. We describe it as delegated semantics where
We let the database vendor enforce the semantics, enforce the aggregation, but then you can still bring that as a predefined metric into Tableau. And that's really the path we're going in the meantime as this open semantic interchange standard comes to fruition.
Sarah (25:08)
and doing it that way and really, I guess, listening to your customers, you must be getting a real sense of, you like an eighty twenty rule with some of the things so that you can then take back to this open semantic interchange and say, we see this is happening a lot, this not so much.
Will Pitzler (25:25)
Yeah, and I I'll preface I'm not sitting on the council, so I I I can't be privy to the exact conversations, but that's exactly right. I mean, what what are the common data models that are being used? What should be the the the the standard that should be applied? And then the the answer also goes to all right, well how do we how does an existing vendor so for I'll just take for example, I have another slide here, but β let me just fast forward. But kind of the way that we think about it, maybe your looker and your DBT.
Right. Those are two very different flavors of a semantic layer. They have very different you know, you think Looker has Look ML, DBT uses YAML files. And so again, we're gonna try to respect what these upstream semantic layers provide, but getting to a common standard becomes really difficult for that reason.
Fiona (26:14)
it's really interesting seeing how the data market has evolved over the last 10 years, for a a hot minute, everyone was, interested in being as far to the right as possible in the visualisation. But as we've shifted into this AI world.
Will Pitzler (26:27)
Yeah.
Fiona (26:31)
that data foundation, the data warehouse, the semantics, the governance has become so critical to the solution, whereas previously it could be more sort of tested. we're moving at such a pace at the moment that, you know, the real focus is to get that right in order to make the the visualisations, for instance,
Will Pitzler (26:54)
you both see it every day with your customers, but it we're kind of doing circles. You think about the late nineties, early two thousands, it was the monolithic stacks, right? It was your business objects and cognos, right? These these monolithic systems that you had all the pre processing work that had to be done. You then go back to, as you mentioned, the rise of the likes of Tableau, the the modern you call self service tools where
Yeah, you can do extract or live connection, but again, it's all about getting data into the hands quickly. That really iterative analytics forward. Then we had kind of the modern data stack approach where everything is lightweight and interchangeable. and now we're kind of going back to the stack where these kind of lake out cloud lakehouse vendors are sort of building this this uniform stack and everything has to land there first, you know, at the sacrifice of maybe kind of that more iterative experimental analytics that you see in the likes of Tableau. And I would even say.
Going a step further, now it's starting to go back again quickly as we see these AI tools. And we talked about it earlier, but for prototyping and things, customers are just working in these third party AI tools. And for now at least, they're accessible. Generally, they're not too expensive for the end user. And so everybody just kind of builds their own analysis. But there's all sorts of issues with that as well. So the pendulum swings back and forth quickly. I guess the good news for Tableau is is we kind of keep our mission the same and we'll work with you regardless of where you want to go.
Sarah (28:13)
And I think underlyingly it's always rubbish data in, rubbish data out, which always comes back to to the kind of first couple of questions that we ask our clients around their data.
Will Pitzler (28:21)
It's always been
Fiona (28:22)
Ha ha ha.
Will Pitzler (28:23)
the same, hasn't it? Right. Th that's what I was saying earlier, like, you know, tools, technology, people, and process, right? Without those three things, you don't get it right. And that's why the work that you both do is so important, is that technology doesn't do it on its own. So yeah, wholeheartedly agree.
Sarah (28:38)
Composable data sources were a massive part of the TC26 story. In plain terms, what does it actually unlock for organisations that have been living with siloed data for years?
Will Pitzler (28:54)
Yeah, I think it's the number one, you both can keep me honest, I think it's the number one feature request in the history of Tableau. Right. You think of Tableau published data sources, we we you can call them a semantic model, you can call it a data set, but it's essentially the metadata attached to your data. It describes meaning and really brings to life how you how you enforce self service at scale, right? How do you make sure that every analyst has the same understanding of what a field is?
The problem with them historically was that you build all your logic into a published data source. It's a distinct modular snapshot of that data, whether it's a live connection or an extract, but they can't be joined together. And so what happens is maybe you get a slightly separate question that you need to provide. You need to go take another published data source, even if it might be the same field. So you end up having sales in 15 different places as part of 15 different data sources. They might all be different extracts. It requires you to extract it 15 different times.
So the beauty of composable data sources is now we're finally able to compose those published data sources together. So it's going to provide all new data modeling flexibility. So you think about entirely new use cases. We have a lot of customers that use Tableau Prep to prepare data and publish to a published data source. Now you can join them together. So you kind of close loop that analysis. You think about extract efficiency, right? No longer are you having to refresh the same cut of data 15 different times, perhaps for just a slightly different segment.
Now it's a single published data source that you can reuse multiple places in a single extract. And then you also think about things like governance optimization, where you want to define your data policies in the database and a live connection, go ahead. If you want to use maybe real-level security in Tableau and an extract, and then you can bring them together as part of the same published data source. So you can have a live connection and an extract together. So drastically simplifying your data ecosystem and just ultimately.
up leveling your data governance and the ease of which you can manage your tableau estate.
Fiona (30:48)
Yeah, I mean, it is and has been for such a long time been the number one request. I believe that Francois, the first time that he saw a Tableau published data source, he said, So can I join them together? And that's what I want next. And so it's taken a fair while to get there, but I think that this is gonna be an absolute game changer. And obviously not for necessarily for
The data engineers who have all of the ability to write back into their data warehouse. There's still so many organisations that don't allow the visualisation analyst to write back tables into their data warehouse. There's also definitely a huge amount of business users who don't even know what a data warehouse is. And this is going to give them so much more flexibility because.
Will Pitzler (31:29)
Yeah, wait.
And it will be that.
Fiona (31:42)
It's gonna be so efficient
in the way that they can bring things together. Like couple prep is gonna become extremely powerful for those business users. You know, they've brought together the data, whether it's from an Excel spreadsheet or they're bringing it down from an API and they're able to push things out, but without creating one big tape.
Will Pitzler (32:04)
Yeah. And I think it again kind of goes similar to the conversation we had about where self-service analytics is going, but like brings up a whole new governance question. So now that you can join all these things together, how do you make sure their performance is still optimal? How do you make sure that you join this data on the right granularity? How do you make sure that the people that are composing these sources you still arrive at the same answer? I think that's gonna be
Fiona (32:21)
Mm-hmm.
Will Pitzler (32:26)
I think like there's gonna open up whole new possibilities, which is awesome. And we're excited about it. At the same time, there's entirely new considerations that are at play. but I think at TC, this is kind of in terms of where we're going is how do we create a single plane for humans and agents to work together on? And this is sort of the start of it. You kind of have that, those composable foundations where you can much better
capture, define, and store logic in a single place, whether that's from a database or it's an Excel file, as you said, or a statistical file, but you have that single point of representation. That's kind of what you would strive for as opposed to redundant copies that exist all over the place. And so the question back to you is how do you help customers do some of the cleanup? Right? Now that you can join these together, I would I would be curious in your take, like how do you even go about tackling this problem of where can we find commonalities and how do we reduce
the amount of published data sources now that are scattered all over the place.
Sarah (33:18)
Yeah. And and I think for me it always comes down to governance and ways of working because we want people to move fast. If you're at the tail end and you need to calculate something on the fly, you're just doing, a bit of a what if analysis, we want we don't want to take that ability away. We'd like them to do it a little bit more governed. doesn't always work. And it's nice with composable data sources that you can use an existing framework and maybe add a benchmark on the side that you're playing around with.
But think what is important as well is when that becomes productionised, there's a process that puts that back through the proper channels. And that's how I talk about ways of working. It's like, let me run fast when I need to, but make sure we've also got in practice the way that we can go back and productionise that.
Fiona (34:04)
For me, I think just thinking off the cuff with it and adding to what Sarah said, I think there's an opportunity using Tableau Prep to hook into the TS events and different things that from Tableau Cloud admin insights and being able to understand a bit more about what's actually being used or perhaps look at the metadata API as well and see what's actually being used. And it gets a little bit technical, so you might need some help with it.
But you could see, is there crossover? Are there things that are occurring? So whilst people may have moved off the Tableau server environment, there may be some people that are back there still, but while most people may be off there, making sure that you're getting the most out of it, really understanding how to bring these views together and and using that administrative time to actually help people understand where they can gain better efficiencies. Cause at the end of the day,
you're still all churning data up there. You're you're gonna be hitting somewhere. You're gonna be having, you know, different information getting pushed up. Either, you know, if you've got a bridge or you've got different things that are going on, it's it's important to be able to track it and trying to use every tool that's in the toolbox to get there is where I would start.
Will Pitzler (35:20)
Yeah, a hundred percent. I think as we, you know, composable data source now you can bring two prep flows together and join them together, right? Like as a published data source, which is awesome. so wholeheartedly agree with what you're saying there.
Fiona (35:32)
We're gonna switch gears a little bit and focus a little bit more in the space that you sit in. and this is the part that I wanted to get most into it with you, is because you're personally involved in it. The whole idea that insights shouldn't be locked inside Tableau and they should reach wherever people work is at the heart of what I try and do. And I really love how Tableau's starting to embrace this. What's the core problem?
that you're trying to solve.
Will Pitzler (36:02)
I think if you go back to where Tableau started, right, our whole MO has been let's keep people out of spreadsheets, right? It it's not governed. you have data all over the place. We believe that ultimately you do see and understand better with visualisations. But at the end of the day, spreadsheets have a place.
And spreadsheets are used by every organisation. I think if you know, we you know, the end of the world happens, spreadsheets will still be sitting out there somewhere, right? They're used prolifically. And so I think for us, it's it was time to just say, all right, let's take Tableau and let's bring it into these tools where every customer works, whether you're an executive, whether you're an analyst, whether you're an end user, everyone is using these office productivity tools, specifically sheets and and Excel docs. And so
For us, it's no different, right? The mission remains the same. but how do we translate Tableau's visualisation layer into a spreadsheet tool? You're working in teams, it's a very different user experience to just working in Tableau. And so going back to kind of the problem we're trying to solve is is we're just trying to make Tableau that context engine and
provide more ways for people to access their trusted data that exists in their tableau state.
I think the bigger challenge is how do we provide a spreadsheet experience that's performant, scalable, and that enforces a level of governance. I think that's the biggest challenge that we had when we first started this project. and it's something just a little bit different to what a Tableau user's used to.
Fiona (37:24)
Yeah, and look, I don't think that the story about letting people access Tableau inside Microsoft products is the end of it. And it's not the end of it in Google products either. I think that what I'm seeing and hearing from Salesforce more directly is
People want access to data to be frictionless and in the place that they're working. And if that's opening Excel because they're working in there, that's where they should access Tableau. If they're in Salesforce, that's where they should be accessing, in CRM, they should be accessing Tableau in there. If they're in Slack, they should be accessing it in there. you should be able to get access to your governed data wherever you are.
Will Pitzler (38:08)
Yeah, and I think composability and extensibility are the common themes that we keep touching on. All these topics that we've addressed, and this is just another example. So when you think about published data sources and how we're really making the published data source as sort of the the I'll use the term because everyone uses it, the semantic layer, semantic model of your organisation, that should be your entry point for for anything, right? Whether you're using Tableau MCP, whether you're creating a dashboard.
Whether you're doing, you know, using a pulse metric, whether you're you know, accessing data in Excel, right? That is your that is your version of the truth that you get to work off of. And so we're we're funneling more of these experiences down to the published data source. And since we're enforcing more composability, it just expands the breadth of questions that you can ask.
Sarah (38:51)
Mm. And you know, we've spoken for God, maybe five to ten years now about keeping in the flow, right? Staying in the flow. But I truly believe it's now that it's actually becoming available. And I and I love to see the demos and so forth of, asking Slack a question and and getting something from Tableau that makes sense back is incredible.
Will Pitzler (39:14)
if we we talk about visualisations and I think that change is hard, right? You've gone into organisations, I'm sure, you can speak to it of like how do we get people onto this this center of excellence around Tableau that we create insights and we share visualisations and we how we visual best practices and how do we communicate. And at the same time tell customers, yeah, sure, just just go on and ask a question in natural language and you get some bullet points back.
Right. It's just very different. Yes, you can get it, you can get a a visualisation, but how do we marry those two experiences together? How should we think about applying the same visual principles as we expand on these AI experiences? And so these are all things that we're thinking about. but yeah, back to your original point, I think that at the end of the day, it's it's still a form of self service. How do we get people the data in their hands where they work? And I think you you said that, you said that really nicely.
Fiona (39:39)
Mm.
And it it's interesting you raise the point around an an agent spitting out more text than visualisations. because that's been something that's on my mind as well. That it's that tension point between can people actually read a data visualisation? And there's a lot of there's a lot of executives in particular who can't read data visualisations. They don't like to say that they can't, but that's just not their strength. And so
But I also know that sometimes a visualisation can speak much more to perhaps the urgency or the issues that are actually occurring when you can see really strong outliers, for instance, on a chart. So I think there's definitely the need for a bit of both in there. And but I do see that the role of the data professional is.
perhaps changing and evolving so that it does include a lot more around how do we give the end users the context,
Will Pitzler (41:03)
it's so true. And I think as solutions like Tableau make it easier than ever for people to to experiment and analyse and and visualize data and as you communicate with best practices, like the importance of the data analysts and creating the foundations is the important thing, right?
And you can expose, you still have those, you know, those dashboards are always gonna be there, right? I I hear all the time, dashboards are dead. It's like, well, I don't go to my car, look at the car panel, and keep asking it how fast am I going, right? Like there's always going to be operational use cases for dashboards. but I think that the surface area, and this is kind of back to the original premise I talked about, it's gonna make analysts busier than ever because you're gonna have more projects than ever, more asks than ever, and the of you know the the tools that you can provide to your end users.
are becoming smarter and sharper, but ultimately you're the ones that are gonna have to prepare the data effectively, make sure that there's context, make sure that they can use, you know, visual best practices. And so I think that it's exciting. Like there's all new opportunities for the analyst or for the the data engineer that that didn't exist a two years ago. but I think that's where it's going. But I don't think the the art of visualisation's ever going on.
Fiona (42:10)
Do you think we can get a bit of a demo on the things that are coming?
Will Pitzler (42:14)
Yeah, let's do it. All right. I, you know, I will I will provide a product demo, I promise. but just for those that aren't familiar, I just kinda wanna tell the story of where we're at with some of these these third party integrations. but as we mentioned, we've kind of been placing more of an emphasis on this. So we've had our Slack integration now for a few years and we're continuing to invest in that space. I think you spoke earlier about how do we bring
Fiona (42:22)
Sure.
Will Pitzler (42:36)
you know, use agents in Slack and use Slack as kind of that agentic interface at least the Salesforce organisation. I think that's where it's going and bringing Tableau insights natively into Slack. β we then made an investment with Microsoft Teams. So that was kind of phase one. So how do we work with that team to make Tableau a first class citizen within arguably the most used office productivity tool out on the market? And so we released that back in 2024.3 and I should mention server and cloud. So we're investing for for both sets of customers.
we then took a similar approach to Google Workspace. And so in slides and docs, the ability to authenticate, provide a level of governance, but let your consumers embed Tableau dashboards or Tableau Pulse metrics and refresh on demand. And then we did the same thing recently in 2026.1. I don't know if you've both gotten hands on yet, but we expanded this to PowerPoint and Word. And so from just an office productivity tooling,
providing easy ways to to bring in Tableau insights. But as I mentioned, right, the bigger problem is spreadsheets. Right. So it's just a very different, we're not just taking an image of a dashboard and embedding it as part of a presentation. We're kind of creating entirely new experiences and and all sorts of different considerations to how do we bring these to life. And so we just released the Tableau app in Google Sheets, which I will show.
β and the Tableau app for Microsoft Excel is in development as we speak and will be coming later this year. And so we're pretty excited β about these new experiences, like totally different for our Tableau customer base. But ultimately, how do we create a flexible data experience that allows customers to bring data in from Tableau Cloud or Server into Google Sheets, but also push data back into Tableau? How do we incentivise our customers to start using the product? And so we wanted to bring both of those experiences together and
That's what I'll show here today.
β let's start in PowerPoint. And again, there's all sorts of really creative ways that customers are using the Google Slides and PowerPoint app. And I'll walk you through first β how it works, and then maybe I'll show you a conceptual just how we're using this app at Salesforce.
kind give you a real world example. So β here is the here's the PowerPoint deck that I'm working with. β And the Tableau app is going to exist as an added. So again this will be installed. I have a demo But your customer can go in, install the app, Tableau Server or Tableau Cloud. You can install, configure via intra, and then you can provision your users. Once it's set up and installed,
Fiona (44:45)
Awesome.
Will Pitzler (45:10)
You can see that I'm logged into my Tableau site here, this MS Team site. It's you can add multiple sites to any tenant. And there's a few different options. So first, let's go ahead and embed a visualisation. So maybe I'll go down to slide four here. And I want to embed my FY27 overview. I can come down and you can search for content. You can embed a URL, keyword with the author, however you want to find it. But the easiest way is that we just used custom views.
So, any custom view that you have, if you want to kind of create a repository of slides or analysis that you maybe have a weekly board presentation and you just want to make it easy to find, use custom views. I'll click on a particular dashboard, and it gives you kind of a preview thumbnail of that source, a hyperlink if you want to go back to Tableau to look at it, and then you can insert the image. And there we go. Now, you'll notice I didn't size this directly to the PowerPoint β
dimensions, but you can do that on the Tableau Cloud or Tableau Server site to make it easy to just one click and embed directly into the view.
So that's kind of phase one, super simple. β
Sarah (46:16)
Nice and I like
how it's got that timestamp at the bottom as well. That will save a lot of arguments across the the board.
Will Pitzler (46:23)
Yeah, exactly. And so you've got this timestamp, it tells you when it was updated, by whom, β and to that point at any time. Like the next step is just how do we keep this data fresh? And so I don't your customers, I'm sure you see this, but these get pretty unwieldy. You might get 50, 70, 80, 100 different slides that customers are updating every week, every month. And so again, we wanted to make sure that this process was easy to automate. And so we also have the ability just to refresh.
So if I come down to the bottom here, you can find all of the different slides in your slide deck.
It sees that I have three different slides. and then just one click, you can refresh it on demand. And so, yes, you have to take the original lift to embed the views, but once you do that, you should be able to automate this pretty simply with a single click.
Sarah (47:03)
Nice.
Fiona (47:12)
if people want to be updating those header titles with the main call outs or putting commentary down the side there, it's so easy to bring together. I mean, imagine that board reporting. It's like blink of an eye, done.
Will Pitzler (47:26)
Yeah. And well, I'm I'm always surprised at how many customers, including ourselves here at Salesforce, you know, we do this weekly, monthly reporting, and a lot of it takes a lot of manual work historically in Tableau. β so to that point, I kind of love to show you, if you're okay with it, as conceptual of just kind of Salesforce on Salesforce, like how do we use the app? Because it's fairly simple on the surface, but you can do some pretty sophisticated things. Now, β this, let me just go back to the right slide deck here. β
Fiona (47:44)
Mm.
Sarah (47:45)
For sure.
Will Pitzler (47:54)
We showed this at Tableau Conference. β again, this is Google Slides, but it's the same mechanism as PowerPoint. β but our internal operations team, there's β six different products, but ultimately expands out to 15 different CMOs at the company. And each has their own operations unit that has to churn out, all of their different metrics that they're responsible for. And so it became a really tedious process for this ops team. So a hundred plus slide decks.
150 different employees and they estimated 200 hours of data entry. So a whole lot of time spent. And I know that deep down, β Sarah, if you and I'm sure you're like, why not just use Tableau? And you know, we push them that direction, but again, change is hard. And a lot of executives want to work in slide deck. So they want to work in Excel. That's what they're used to. And so we accommodate. So this was the historical β summary deck. Don't worry, this is not this is not real production data.
But these are the types of scorecards that we were producing, or the ops team was producing. And so this would require manual effort to build this in Tableau and then enter the data and report it. What they've done is they've created an automated process that allows our ops team to set up and configure a Tableau dashboard with all of this data live, up to date, connected to our back-end database, so that it's ready for real-time reporting at any time. They do it one time, they pick the slide template. This is actually a Tableau image behind the scene.
This is actual Tableau data. And they use this Google Slides integration to auto update this. And so what this has done is they estimate, I I asked him, like, what how many hours are you talking? He 2,000. And I was like, that's a pretty big number. β but again, it brought just entirely new credibility, both for them as Tableau users. but they're starting to roll this out across, again, an 80,000 person organisation that's just driving a huge impact and saving.
a whole heck of a lot of time on just processes that should be manual. I know everyone wants to talk about AI, but ultimately it's just how do we give people time back? How do we make it easier for them to do their job? And so what they've done is they've built this deck builder concept and I'll expand this. And I'll let me just show you how it works. Again, this is all a prototype. β but if you can look at what's here, it's again it's very complicated in Tableau. You think about all the different lines of business, all the different date filters, all the different products that we're responsible for.
How do we make sure that operations analysts can do it? So they go and set their filters. They then have pre-built images in Tableau of the slide layout that they need to use for their particular CMO or their particular line of business. They use dynamic zone availability to be able to select their filters and auto-populate all of the data on the slide in Tableau natively. And then from there, what they're able to do is they go in, they size it for Google Sheets, they embed it, and then they have a living document that they can refresh on demand.
So providing the ability for all of these operations analysts to self-serve and so there's not a single point of content anymore, validating all these decks, making sure it's correct, right? A much more streamlined approach.
Sarah (50:47)
Yeah, wow. Like looking at those hours saved and you know, we've all been in these cycles and you're trying to get the report out and then someone goes and changes something or someone says, Hey, we forgot to put something in October and there's like fifty pages that need to be updated again. Amazing.
Will Pitzler (51:04)
Yeah.
Yeah. And I think the first question that every customer asks is why aren't these live dashboards? And we evaluated that solution. There's some API limitations as well, but the the main qu the main decision we made was just customers don't want this to update, right? You don't want to walk into your executive deck and realize that your data's changed. And so that's why we made that decision. β but again, the whole goal here is really just keep it simple, but just make our customers' lives easier.
Fiona (51:16)
Mm-hmm.
Sarah (51:24)
Yeah.
Fiona (51:30)
I'm so sure that any leaders watching are just like, this is gonna save me and my team so much time to get this done. And thinking about the steps that are required in order to bring those dashboards together, it's not that big, you know, as long as the data exists, it's not actually that big a step. And gee, that it looks clean and slick.
Will Pitzler (51:53)
Yeah, and again, the using the creative skills of Tableau to do it, right? I think that if you just import the image, yeah, that's great. but you can customize it and appropriate it for your organisation with your custom themes and fonts and templates. And so to really make it look like a an a a more, you know, custom built solution, I was just really impressed how they did that.
Sarah (52:13)
For sure. It's the first time I've seen that. So thanks for sharing. Will it looks amazing.
Fiona (52:18)
having spent a lot of my life in PowerPoint, it makes me super excited to see this this happening and knowing just how many people are going to be able to save countless hours instead of doing copies and pastes and downloads of images and then not having anything that's
the wrong resolution or not actually coming through, it's gonna be awesome. or forgetting to do something like copying and pasting like an image and then forgetting to do one on one slide. And then of course that's the thing somebody notices and then they say, this is all wrong. I don't trust it.
Will Pitzler (52:51)
And well, we're just getting started too. So I mean Sarah, Fiona, you're, you know, we appreciate all the work you all do. And so as you're talking to customers and new ideas come up, new features, like we rely on your ideas too. So please keep the feedback coming. I think we just want to get the word out and get more customers using it.
Sarah (53:07)
good, 'cause I've thought about ten other things that I'd like to see in that.
Will Pitzler (53:11)
Awesome. Come find me.
Fiona (53:12)
I also want to see what's going on in Google Sheets if possible.
Will Pitzler (53:14)
Yeah, let's
do it. Let's absolutely do that. So I'm gonna come back over here. Let's go to Google Sheets. and again, same same workflow. We showed PowerPoint. It's gonna be the a similar user experience, we similar user interface. You can get this from the Google Workspace Marketplace, Tableau Cloud, or Tableau Server Customers for both. once I've installed the app, you'll see on the right hand side that I get the Tableau Sparkle.
So I'm going to share my screen again and let's jump into Google Sheets. And again, β very similar workflow, very similar user experience, very similar user interface to what you saw in PowerPoint. β Like the Microsoft 365 app, the Google Workspace app is available for both Tableau Cloud and Tableau Server customers. Get it from the marketplace. Once it's installed and configured, I've been provisioned as a user. You'll look on the right hand side and I've got the Sparkle here.
And you'll see that I have my Tableau site that is far too long here at the top. I've already set up the authentication. I've built a connected app so that we have a secure translation between Tableau and Google Sheets. Couple of different options. So again, we can import data from a view. So we'll show we can show what that looks like. But essentially, you're just looking at an existing, I'll just walk through it first. Let's do that. So let's say I know the name or I know the keyword, or maybe I have a custom view I want to save as a favorite. You can identify that data source.
I can look at the what if forecast if I want to look at the data. Let's look at the individual sheet.
And again, so now I'm looking at an individual sheet, not a dashboard. This is already a text table, a highlight table, I guess you would say. And then I can import that data. And what it does is we're using the REST API under the hood to programmatically query that view. And it brings your data into a nicely formatted table here in Google Sheets. So that's one way that you can turn your visuals into row-level data. Now the second
Way you can do this, and again, this is probably a more scalable and more flexible method, is you can use published data sources. So our common theme of the the session here is that we're talking about published data sources. That's your foundations. β now the first challenge we had here again when we talk about governance and permissioning, the first thing to say is that this inherits your tableau permissions. So if you establish those permissions and your roles, it's going to be inherited through the sheets app. So if you've got viewers, you don't want them downloading data, they're not gonna be able to.
Fiona (55:16)
Mm-hmm.
Will Pitzler (55:35)
But if you want to facilitate this, you can. If you tag a data source with Google Sheets, this is how you expose these published data sources. We didn't want any customer to come in and see all the published data sources, right? How many? I'd be curious, how many do your published data sources do your customers have? Tens, hundreds, thousands? there's too many. So you have to explicitly tag the published data source. And from there, you come into this query builder experience. And so this is where we can β choose to completely disaggregate data.
Or we can aggregate similar in Tableau to the level of detail in your view. So you could say account name, account timing. We'll just select a few fields. We'll take our sales. We'll next configure import. And again, I could come down here, I could filter data, I could β change the level of aggregation. So you have some autonomy and how these fields are represented.
I'll leave that out for now. And then again you'll hit this import data button.
And this one was again successful.
So a couple of things is that at any time you can come up to the comments. We talked about governance and how do we know the data's up to date, who updated them. So you do get this nice audit trail saying, was this refresh successful? Again, there's a finite amount of data you can bring in, right? Google APIs, I think we have up to five million cells you can use. So this is how we would communicate, whether you're it failed or succeeded, or it's too big, or it's cut off. This is the transcript layer of of the app. But again,
Pretty simple to just bring data into the view through a view or through a published data source.
Fiona (57:05)
What licensing would somebody need in order to get access into the data in here?
Will Pitzler (57:11)
as long as you have a tableau license, viewer, explorer, creator, as long as you have the appropriate permissions assigned. So again, your site role and then the permission to to use the API, you can have access to this. So that again, it's any user type as long as you have access to the the Tableau Cloud or Tableau Server site.
Fiona (57:29)
Did see one extra thing on that β right-hand side menu, which was you can upload data back to Tableau Cloud.
Will Pitzler (57:39)
Absolutely. So if I go back in here, β we talked earlier about how do we create kind of a flexible experience, not just pulling data out of Tableau, but what if you've got a big spreadsheet that you've modeled? Maybe you want to use a pivot table and you want to go explore that visually, you can do that. So if we hit the export button, now I have not logged into my site, so we are going to authenticate live on air. but I will select, let's say, this this sheet that I just imported. Maybe I augmented it with something else outside the platform.
Fiona (57:53)
Yeah.
Will Pitzler (58:07)
You can then hit explore to Tableau Cloud. Now I will say this feature specifically, the export to Tableau, is Tableau Cloud only. That is the only thing that is exclusive to Tableau Cloud, which makes a lot of sense if you think about how this operates. But what this does is this creates kind of an ephemeral web authoring experience where you're not actually creating a published data source.
Fiona (58:27)
I just wanted to point out that it's possible because one of the things that I think β people have been etching for for a long time is write back. β and and so I know that there's some things on the horizon. There's both an already an extension that you can use for write back, but there's also β some things that they presented at TC26 in the lab's Tableau Solve.
But this could also be another way for people to get around it and to really have that data flowing in different directions because not everything's gonna make it into your Snowflake environment or into your Databricks environment. And people love Excel. People love Google Sheets, it's something that they're very, very familiar with. And it feels like you've kept open.
a whole new world of possibilities here for people to be able to extend their data.
Will Pitzler (59:21)
Yeah, and I think that was the the biggest challenge is we heard we I think we interviewed something like fifty different customers before we built this app. And we had fifty different responses to how they want to use it. But the ability to go back into Tableau and to you the way you eloquently stated that is sort of another way to β help customers self serve. And you know, whether it's Tableau prep writing to a published data source, whether it's publishing data to your your your sandbox environment or publishing
Sarah (59:29)
Okay.
Will Pitzler (59:49)
directly back that you can then join as part of a a published data source, just expanding options for customers is really what we're going for.
Fiona (59:56)
Yeah, and that's the kind of innovation that I think that a lot of people have been looking for. You know, that old adage, you know, if Henry Ford had asked everyone, about the car and they would have said, I just need a horse. I mean the same thing happened with dub dub. We went out and asked a lot of people about what we should be doing. We got some great feedback on our ideas and on our idea for a marketplace and we tanked it.
You know, like it it didn't it it didn't work. Part of that reason would be just around l I guess the sales and doing everything like that. There are some very successful marketplaces out there, but it just wasn't for us. And I suppose we did all the research in the world before we built it is my point. And like you say, everyone gives different answers or positive responses, and it may maybe that it just doesn't land that way. But in this one, I think it will.
Will Pitzler (1:00:46)
Yeah.
yeah, and I think that we're excited about the possibilities. And again, we're just getting started here. So for customers that are already using it, this is version one. We've got a a backlog that's quite long. So we're just kind of getting started. And really like the best part of this, and again I'll go back to a Salesforce example, β you'd be shocked that our finance team still uses Google Sheets for certain use cases.
And so what they do is they build a big pivot table that's far bigger than this simple one that I built. What they do is they download these data from disparate sources. So think of it like another connectivity outlet, right? You've got data in your CRM, you've got data in your Snowflake environment, you've got data on-premise database. Ways to just kind of this provides a lens to that data that they can export into different sheets, build the pivot table, and then you can easily just hit update data. You can select your data that you want to refresh.
And you can override it. And with one click, you refresh that data with the current data and it updates your pivot table on the
Sarah (1:01:44)
That looks amazing. My governance senses are going a little bit wild, but again, with the right, ways of working and governance and processes in place, I can see it it it really allows an environment where people that need to do stuff fast can really fly with it.
Will Pitzler (1:02:04)
Yeah, and I think again, just respecting the permissions, it has to be there's only certain people to get access to this use case and it's for, you know, again, it's locked down and there's all sorts of governance procedures, of course, before before they use it. But that's the whole point is just giving flexibility, is that if you wanna give access to a data source, you can. You don't have to. If you wanna give it to one user or all your users, you have that flexibility and you don't have to duplicate your efforts in terms of governance or data policies. It's gonna respect your level security, it's gonna respect your site role and respect your permissions.
Sarah (1:02:33)
That's what we like to hear.
Fiona (1:02:35)
Love that.
Will this has been a genuinely great conversation. But before we wrap up, I want to give you the floor. What's the one thing you want a data leader walking away from this conversation to actually do? Not just think about, but do.
Will Pitzler (1:02:52)
that's a big question. But before I answer, Sarah Fiona, thank you so much for the work that you do. Thank you for having me. It's been nothing but a joy. for a data leader, again, going back to kind of your original premise, I would really think more broadly beyond just technology. Think about the processes you have in place, think about the people that you have in your organisation and really come forward with a data strategy and an understanding of where you want to go before you just jump in and start. I think that's kind of goes back to just have a process in place, have a vision of where you want to go.
And then go execute and get help from people like Sarah and Fiona.
Sarah (1:03:24)
And now, well, where can people find you and follow your work? Because I imagine there's gonna be a lot of people listening that are gonna want a lot more of this.
Will Pitzler (1:03:32)
Yeah, absolutely. I'm admittedly not an active LinkedIn user, but that's probably the best place to find me, at least initially, or if you have a contact at Tableau or Salesforce that you want to ask to speak, happy to chat with any customer. but you can find me on LinkedIn, probably the easiest way.
Fiona (1:03:45)
Thank you so much, Will. You came with a point of view and you didn't let go of it, and your storytelling is on point. I reckon you should have your own podcast.
Will Pitzler (1:03:55)
No, I'm gonna leave it to you. But
no one needs to hear any more from me. But really appreciate it. Thank you both and for all the work that you do on behalf of the customers as well. We we appreciate it. So thanks for having me and always good to chat.
Sarah (1:04:06)
thanks, Will. And if you loved this one, hit follow, leave us a review, and share it with a data leader who needs to hear it.
Fiona (1:04:14)
What a conversation. maybe take a few of these things away today. Your foundation is everything. composable, governed, well modeled data is what every AI promise is actually built on. and trusted AI is not a feature. It's actually the architecture.
The last thing you should remember is that insights need to travel. That last mile really matters. And your data's only gonna be as useful as the places where the decisions get made. And spoiler alert, for most organisations that still includes a spreadsheet.
Sarah (1:04:49)
Huge thanks to Will. Thanks for being so open about what is really hard, what's real, and what's still being figured out. You'll find the links to his work and everything we mentioned in the show notes.
Fiona (1:05:04)
This has been undubbed where we're unscripted, uncensored, and undeniably data.
