Databricks Data + AI Summit 2026 | Bob Pisani, Addepar

Clips
More from Databricks Data + AI Summit 2026

Bob Pisani

CTO

Addepar

play_circle_outline Addepar: data and AI platform for investment professionals (family offices, RIAs, private banks)

play_circle_outline LLMs and Agents: Extracting Private Market Positions from Documents to Deliver Total Portfolio Views Across Alternatives and Public Assets

play_circle_outline Scaling Alternatives Data with Databricks Lakehouse and Unity Catalog: A Strategic Partnership for AI-Powered Platform Innovation

play_circle_outline Graph-Native Ontology Reveals Network Effects Across $9 Trillion in Assets: Anonymized Cashflow and Investment Patterns

Info
Transcript

Bob Pisani, Addepar

Bob Pisani

CTO Addepar

In this interview from Databricks Data + AI Summit 2026, Bob Pisani, chief technology officer of Addepar, joins theCUBE's John Furrier to discuss how AI and the Databricks lakehouse are transforming portfolio intelligence and unlocking the largely untapped data potential of private markets. Pisani explains how the platform, which oversees more than $9 trillion in assets across family offices, registered investment advisors and private banks, made an architectural bet on Databricks roughly four years ago. By adopting the Lakehouse and Unity Catalog, Addepar bu... Read more

explore Keep Exploring

Is Addepar a modern wealth management platform? add

How can Databricks and AI (e.g., LLMs and agents) be used to extract and structure unstructured private‑market data (PDFs, documents, spreadsheets) to create integrated, transparent portfolio views for alternative investments? add

What was your architectural bet in choosing Databricks (the Lakehouse), and how did that architecture enable your platform for processing alternative investments, document processing, and using LLMs/agents? add

How does Addepar use graph-native architecture, models, and agents to represent and analyze investment and client data across its global asset base and support products like Navigator for scenario analysis and cashflow forecasting? add

bolt Powered by CUBE AI

Bob Pisani, Addepar

search

John Furrier

>> Welcome back to theCUBE's live stream here on the show floor at Databricks' summit. It's the Data and AI summit. theCUBE has been here since the foundation and the founding of data. We've followed all the events. I can't remember how many events we've been to, but this has really kind of crossed over into the mainstream. 31,000 people are here. If this is breaking out the seams, if it gets any bigger, it's going to go to Vegas and probably look like a re:Invent. If it hits 50,000, then it's going to be too big. But the big story here is injecting intelligence into businesses, injecting intelligence into process, injecting intelligence into domains. Our next guest has got an excellent venture around how they're using AI to change a very specific domain in financial service. Bob Pisani, CTO at Addepar's here. Thanks for coming on. I really appreciate it.

Bob Pisani

>> Thanks for having me.

John Furrier

>> One of the things I love about AI over the past couple years and following it coming from the cloud days is cloud was horizontally scalable. Great. You could get it. But the SaaS apps were just very domain-specific. You want file sharing, you get Dropbox or box.net. The early days now is called Box. So now we're seeing the domains have AI in it because there's data there. So you guys are in an area where you have a very unique position and I would say, I guess financial services. Explain what you guys do because you're using AI to change the game on getting intelligence fast, data sources, many data sources. Explain what you guys do.

Bob Pisani

>> Yeah. So Addepar is itself a data and AI platform for investment professionals. And we serve family offices, registered investment advisors, private banks and institutions. And ultimately they're using us for reporting, performance analytics, and really driving insights and decision making through intelligence on all of that portfolio data and all of their clients' assets.

John Furrier

>> So would you say you're a modern wealth management platform?

Bob Pisani

>> Absolutely. And we have over 9 trillion in assets on the platform and it's split probably about 40% in alternative investments and 60% in marketable securities.

John Furrier

>> All right. CTO, you got to look at the tech, so I want to get your thoughts on the AI because we hear workflows, automating workflows. Where's the real value creation with the tech besides workflow automation? Because workflow automation is a process. That's not true value enablement.

Bob Pisani

>> Yeah. The way our clients think about value is in terms of their clients. So we have to think about how do we drive outcomes all the way from back office to front office. So this is where you're enabling an advisor or an investor to make a decision either to change a portfolio, to make an investment, to think about a cash flow, to think about a life decision that affects their family or them personally. And so the value comes. And can you take that workflow all the way from, yes, the back office work, the operations, the risk, the data quality all the way through investment decision making where you're having a portfolio decision made or an investment decision made?

John Furrier

>> One of the things about private markets is... And I'm curious to get your reaction because there's not a lot of data in private markets. It's not a market, it's not public. So a lot of the family offices and wealth management evolves around private markets. Companies are staying private longer. There's secondaries happening. There's a liquidity. It's not yet understood, but there's plenty of unstructured data out there. How do you view that? I think this is a big opportunity where you can apply some of the Databricks tools and inject intelligence in. What's your view on that?

Bob Pisani

>> Yeah, exactly to that point. We use Databricks to build an alternatives data management solution. So we end up doing document processing. We apply LLMs and agents to process those documents into positions and transactions for alternative investments, which then allows us to provide a total portfolio view across alternatives and publics to our clients. And so now we're bringing transparency and the same equal footing to investment decision making for alternative investments which were trapped in PDFs and spreadsheets and bringing them to be at the same level as public investment feeds and public data that they have with a typical set of custodians and banks.

John Furrier

>> So take me through that because... Talk about the architectural bet. Was it their Lakehouse? What was your key thinking around it? Because that's smart. You're basically collecting data and putting it into a system. What's that architecture look like? Share your thoughts.

Bob Pisani

>> Yeah. So we started with Databricks about four years ago with the Lakehouse and we started really replatforming our infrastructure and thinking about how do we modernize because we needed scale, we needed a tool set, and also a set of capabilities that was platform-oriented. So I guess it was quite fortuitous that it set us up to have Unity Catalog and all this wonderfully curated data to then add alternative investments to and bring alternatives data to bear. That then set us up to then leverage all of the AI capabilities and the agentic capabilities that Databricks has brought to bear. So it fit in perfectly with our platform and we built it in at depth natively.

John Furrier

>> I was just talking to Jonathan Franklin before you came on to the air. And we were riffing about AI and I asked him a question about how the model's going to work. He said, "I'm the chief . I don't want to answer Reynolds question. That's his answer." And I said, "Okay, that's fair." And he made a good point that a lot of the models you can branch off with agents and like almost fork data sets within Databricks, kind of like a GitHub repo.

Bob Pisani

>> Okay.

John Furrier

>> But the idea was there's a lot of data that you can get and connect that's not obvious. So when you think about what you guys are doing, I want to ask you, you must be seeing some network effects developing. Could you share your thoughts on this? Because the more you have out there, the more that's coming back in, because ontologies is just to me like a graph mindset, like concept. So ontologies are hot, okay.

Bob Pisani

>> They're back.

John Furrier

>> I mean, I used to do ontologies in the 80s by hand, but now you've got agents and super computing. So we're seeing a compute level and a data relationship that's opening up new patterns and new ways to ... What's your thoughts on these network effects?

Bob Pisani

>> To me, it's the most natural way to represent data. Addepar has been graph-native for the last 15 plus years and we've been working on our ontology and finance and investment for that same amount of time. So now you're absolutely right. These agents... And we use lots of models, lots of different size models and agents to basically process data. So we do have different views. So we'll have a client and portfolio view at depth. We'll also have a cross-sectional view where we'll look across $9 trillion in assets and look for patterns in an anonymized and aggregated way to understand cashflow movements, to understand different types of investments on an industry or sector basis, to look at different segments and how they invest. What are family offices doing? What are advisors doing? How are the private banking movements looking? And also apply a global lens because Addepar has clients in over 60 countries. So how is data moving around the world? And we have wonderful insight into all of this and then we bring that to bear in different types of products, whether it's through our navigator product, which is all for around a scenario analysis, and forecast for cashflow modeling, or it's in our private fund benchmarks.

John Furrier

>> And you get engagement data coming back.

Bob Pisani

>> Absolutely. We see what the clients are doing and what their end clients are doing for us.

John Furrier

>> It's almost a fourth pillar because you sit-

Bob Pisani

>> It is.

John Furrier

>> Because you sit between institutions, advisors, and asset owners. And now you have this core thing called what are people doing?

Bob Pisani

>> behavior data.

John Furrier

>> So it's almost like passive monitoring and active data coming together in one.

Bob Pisani

>> And our clients are so eager to understand how the platform itself is being used, which creates a whole nother data domain to explore and apply.

John Furrier

>> All right. Let's talk about what that translates into and what's the requirements because again, Jonathan and I were talking about trust. If the data sucks, you don't trust the data, you don't like the outcome and it's not even worth it. And they even said on stage in the keynote, they were showing agents failing. And they were transparent about it. Finally, someone said it. Now the hype is real. I get that on agents, but there's progression, it's evolution. The trust piece with agent becomes huge. How do you look at that? Because you're building a network, it's not just software, because you got coding now is easier to do. You've got the network effects, you've got the observation space. Trust, how do you square that circle?

Bob Pisani

>> So trust is multidimensional. I think it's in terms of data quality. Can you trust the data? Is it operationally viable? Is it analytics-ready and research-ready? Then there's trust around, can you secure the data? Can you ensure that access is controlled, whether it's by people or by AI and by agents? So our clients have certainly come to trust us as stewards of their data and the same level of trust and the same permissioning and authentication schemes we apply at the data level is inherited by our products and is inherited by our AI and agents. So we ensure that clients know that whether it's all the way down at the data tier, call it the context tier and the application tier and the agentic tier, that same permission authentication model is applied.

John Furrier

>> Can you share, Bob, an example that illustrates some of the insights you guys are getting that's unique to you guys?

Bob Pisani

>> Oh yeah. So we did a lot of research. We do it every year. We publish it on a quarterly basis as well. We were seeing that as the global markets were retracting from kind of this bet on America, tariffs and the wars and all the rest, what does it mean? And there was this desire to diversify away from the US. But we were seeing in the data tell a different story in that we were seeing all this data actually flowing back into the US were staying there because massive VC bets, massive PE bets. You also had the big AI play and everything that was happening was keeping dollars in investment in the US and denominating the USD. So that was counterintuitive and it was very ... Everybody wanted to tell the story of how dollars were leaving the US, but it wasn't actually happening.

John Furrier

>> I love what you guys do. Just want to say I'm a fan. I'm asking some probing questions I want people to understand. I think that you are a template for the future because when you look at the big companies that move the needle... I'll name a few just to illustrate the point. Uber, they pioneered multiple moving parts in one system. Palantir, they pioneered how to do complex ontologies and all kinds of systems, what you guys are doing. The question is that unique or is that a path that all companies will see? Because what Databricks is showing here and what we're seeing with the cloud scale and these neoclouds, certainly sovereignty is going to be a huge issue. I mean, damn, that's going to open up a boatload of data for you when people say, "Hey, how do I run a sovereign cloud and a sovereign AI economy in X, Y, Z country?" I mean, it's mind-blowing.

Bob Pisani

>> It's amazing to partner with Databricks, but it's also amazing to see how they're going to change the industry just as AI is. And we certainly recognize that there are opportunities and threats. And we have to be cognizant of the fact that we need to stay ahead of all of that, and that means that we have to continually innovate. We invest a ton in R&D. We work closely with Databricks. We love to have them as a partner, but it also means that what do we do well and what differentiates us? And it is our data, it's our network like you called out. It's the client intersection that happens with the data, and that's a pretty defensible mode.

John Furrier

>> All right. I have to ask you this question because you're an expert. Since you're here, I might as well get a pound of flesh on this one question. I've been researching this and I'm trying to find an answer. We're seeing a lot of these industries get disrupted. Research traditionally has been collect data and report it.

Bob Pisani

>> Yeah.

John Furrier

>> Kind of a yawner. Some people throw AI at things. What's the modern research look like? Because it seems to me that it's trending towards data acquisition, proprietary intelligence, platform based systems, not manual linear scale services. What's changing in research from your ... It could be all research, just generically, but you actually, you're disrupting traditional research.

Bob Pisani

>> We are and I think it's not just someone what either Ali or Arsalan said in one of their talks where the jobs are still there. You still have to do research, but the tasks and the way you do it are changing. So absolutely. I don't think anyone is manually doing research anymore. Or if they are, it's more of a final step. I think ultimately we and our clients are all using AI for sourcing information. The better that information is in terms of quality, but also I think uniqueness, and therefore the kind of access you have adds value even regardless of what the content may be. So once you bring those together, AI is great at synthesis, AI is great at obviously creating reports and written language, but to me it's going to come down to how do you provide access to disparate data sets that otherwise would have required, frankly, getting the manual effort to go find it, which now AI can do for you, but now you still need that access and almost the permission into otherwise .

John Furrier

>> Great example. A platform that's horizontally available, but could be tailored with AI. All right, final question for you, market structure. What is the biggest change in the wealth management sector that people underestimate?

Bob Pisani

>> Acquisition. It is absolutely changing the face of wealth. And you see massive consolidation in private equity dollars flowing into the RIA space. You see massive consolidation globally of advisories. You also see massive competition as a result for both retention and attraction of advisors and clients. So now you have this huge war for talent and war for clientele. That's always happened, but now you're actually in as much a war for talent, for your advisors, and your investment talent as you had been for clients. And it's all being fueled by massive investment from VCs and PE.

John Furrier

>> And when you say acquisitions, you mean M&A or acquisition in the consolidation way?

Bob Pisani

>> M&A.

John Furrier

>> Okay.

Bob Pisani

>> M&A, yeah.

John Furrier

>> Because I mean, think about a private equity playbook. AI is a perfect storm for them.

Bob Pisani

>> Absolutely.

John Furrier

>> I mean, they can move so fast. They have the money, they're resetting, they can replatform. The old days is like, "Okay, change management, get the new management team, go to market, blah, blah, blah."

Bob Pisani

>> Yeah. It's interesting. I think it's beyond the financial engineering and kind of the deal making, to now where you're using AI to actually reshape the operating model of the business, which allows you to go much bigger, much faster. And we're seeing that play out in our clients. We're facing this every day.

John Furrier

>> All right. Final, final question since you're here. Do you think the private market is underserved? Because the public market's been around for a long, long time. I mean, the private market has a new dynamic. Yeah, there used to be some secondaries for the biggest companies and then you go public.

Bob Pisani

>> Yeah.

John Furrier

>> So there seems to be like a bigger later stage market and there's talk of tokenization of equities. So a future could converge where I'm going to need a lot of data to make decisions. What's your take on this whole private equity, private late stage, pre-public?

Bob Pisani

>> Yeah. So one of the things our clients are desperate for is sort of these different levels of data where normally, sure, you might be able to see the investment in the fund that you have, but can you see the look through data? Can you see the underlying investments? Of those underlying investments, can you see the underlying financials? You want to get the level of depth that gives you the kind of equity grade investment and the thesis and the analysis, but in the private markets. And it has been underserved in that it remains a very disjoint ecosystem. You have GPs and LPs with fund administrators in the middle and technology players that often have been in the mix for decades. It's a lot of friction.

John Furrier

>> It's like your example about the spreadsheets. There's data there, but you put it in one spot and you get all that data. The private market, some would argue like, "Well, there's no one's really trading. It's not really your market." Well, kind of a market. There's a lot of data laying around that no one was paying attention to in the past.

Bob Pisani

>> And we're doing our best to start to leave the glue for that because we sit at the intersection of all the parties. Like you said, it's the network.

John Furrier

>> You sit between a lot of people. All right. I have to ask one more question because the next guest is not yet here. What's your goals this year? You got the CTO view. What's your focus? What are you optimizing for?

Bob Pisani

>> I'd say it's certainly AI transformation internally as well as the product experience. We're seeing an inversion where all of our products that have been built over the last decade around reporting and analytics and dashboarding and mobile experience are really morphing to become AI-first. So there's an inversion where AI will lead to the other products and the rest of the surface. Internally, it's about using AI to transform how we build and how we deliver. So it's both internal, external. So it creates a lot of pressure to deliver so much faster and deliver so much more value sooner.

John Furrier

>> If you can handle the pace of play, you're going to do well in this one.

Bob Pisani

>> Yeah. Absolutely.

John Furrier

>> Bob, thank you for coming on. Really appreciate it. Congratulations. Love your venture, love your architecture. Thanks for sharing.

Bob Pisani

>> Thanks for having me.

John Furrier

>> All right. I'm John Furrier of theCUBE. This is a great example of the inversion. You're seeing these AI factories. Now you see these PCs come out. It was never like that. It was the PC, then the server. Now you've got the AI factories. And the PCs you're seeing disruption at the domain level impacting market structure. And again, acquisition is a big part of it. Again, the data doesn't lie. You get the data right. A lot of good things happen. We're doing our part to share the data here with you. I'm John Furrier, your host of theCUBE. Thanks for watching.