Raj De Datta of Bloomreach, chief executive officer and co-founder, joins theCUBE Research at the Bank of America Private Tech Trailblazers Conference to discuss the intersection of artificial intelligence, abbreviated AI, commerce and platform strategy. John Furrier of theCUBE Research hosts the conversation with the theCUBE Research team.
De Datta outlines the industry's transition into a settling phase of AI, the emergence of vertical agents and vertical-specific models and the strategic role of proprietary vertical data stacks for ecommerce and martech. They describe Bloomreach's Loomi AI engine, Loomi Connect and platform distribution strategies.
De Datta argues that foundational models become commoditized and that differentiated value resides in proprietary vertical AI trained on domain data. They state that Bloomreach's Loomi engine, built on seven billion profiles, can outperform out-of-the-box large language models by five to ten times. Analysts and Furrier emphasize agent actionability, governance, identity and hybrid execution as critical factors for enterprise adoption.
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Bank of America: Private Tech Trailblazers Conference. If you don’t think you received an email check your
spam folder.
Sign in to Bank of America: Private Tech Trailblazers Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for Bank of America: Private Tech Trailblazers Conference
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for Bank of America: Private Tech Trailblazers Conference.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Raj De Datta, Bloomreach
Raj De Datta of Bloomreach, chief executive officer and co-founder, joins theCUBE Research at the Bank of America Private Tech Trailblazers Conference to discuss the intersection of artificial intelligence, abbreviated AI, commerce and platform strategy. John Furrier of theCUBE Research hosts the conversation with the theCUBE Research team.
De Datta outlines the industry's transition into a settling phase of AI, the emergence of vertical agents and vertical-specific models and the strategic role of proprietary vertical data stacks for ecommerce and martech. They describe Bloomreach's Loomi AI engine, Loomi Connect and platform distribution strategies.
De Datta argues that foundational models become commoditized and that differentiated value resides in proprietary vertical AI trained on domain data. They state that Bloomreach's Loomi engine, built on seven billion profiles, can outperform out-of-the-box large language models by five to ten times. Analysts and Furrier emphasize agent actionability, governance, identity and hybrid execution as critical factors for enterprise adoption.
In this interview from the Bank of America Private Tech Trailblazers Conference 2026, Raj De Datta, co-founder and chief executive officer of Bloomreach, joins theCUBE's John Furrier to discuss how AI is entering its "settling phase," where enterprises move from experimentation to repeatable, outcome-driven agentic systems. De Datta explains why value is shifting away from commoditized foundation models toward vertical-specific AI trained on proprietary data, noting Bloomreach's Loomi AI engine draws on 7 billion consumer profiles to outperform out-of-the-box...Read more
exploreKeep Exploring
What phase is AI entering, how are models and infrastructure evolving (including the emergence of heterogeneous platforms), and where will companies derive lasting value — from the models themselves or from vertical-specific stacks?add
How do you view the availability of tunable/open‑weight models from an entrepreneur’s perspective, and how is your company leveraging proprietary data and models (for example, Loomi’s dataset and models) to build a competitive moat, improve customer outcomes, and drive growth—what is your current growth strategy and positioning?add
What factors are creating tailwinds for your business, and how are you exposing your platform to support third-party agents and headless user interfaces?add
How are brands responding to the shift toward becoming their own infrastructure and going direct-to-consumer (including via AI agents like ChatGPT and Claude), which customers are winning, what opportunities and challenges (e.g., data layers, delegation, trust) does this create, and what new services or mechanisms are emerging?add
>> (INTRO)John Furrier with theCUBE. We are here in Palo Alto, heart of Silicon Valley, here for the BofA Private Tech Trailblazers event. This is where the entrepreneurs and the companies that have escape velocity are gathering. Investors are here, industry participants here. Raj De Datta is here, CEO and co-founder of Bloomreach. Great to have you on theCUBE. Appreciate it.
Raj De Datta
>> Thanks, John. It's great to be here with you.
John Furrier
>> Heart of Silicon Valley. You've done a lot of different things in your career. I was just checking out a little bit of tennis. You were on the Champions Council there, which is distinguished. Also entrepreneurial career. You've been an analyst. Before I get into the company, 'cause I want to talk about what you guys are doing, what's your take on the current market right now? Because you've seen those waves of innovation. We've kind of lived through the, okay, get a pitch, get a prototype. Now it's like, have a hackathon. Now it's like, they're not hackathons, they're build-a-company-thons. we're in a whole nother
Raj De Datta
>> era.We're in a whole nother era, but I describe this as the settling phase of AI. And what I mean by that is for the last year, 2 years, 3 years, it's felt like we've been in an AI amusement park. It's like, oh, look over here, it can be a lawyer. Look over there, it can be a banker. Look over here, it can be a coder. And now we're starting to understand what it can do, what it can't do, what the guardrails are, what the risks are, what the opportunities are. And I think you can build great companies in a settling phase because you can plan on exactly what it can do.
John Furrier
>> Yeah, a lot of the grind and toil, undifferentiated heavy lifting, whatever word you want to use, it goes away with some of the tooling. It's only getting better. The next model comes out is better. But now with kind of a realization of, okay, not just the GPU and the LLM and the foundation model, you're starting to see a systems game emerge. And people are decoupling the models from the underlying kind of, infrastructure, I guess. I don't know what to call it, but you got cloud, you got on-prem, you got supercomputing with NVIDIA. Now AMD just bought, ZT Systems. So it's like now you're starting to see heterogeneous kind of platforms emerge that we've never seen before. Yeah, cloud was great, but now you have the next thing. Explain that phenomenon, because I think this is a sleeper story.
Raj De Datta
>> I think it's a great point, John. And what we are seeing is, the narrative 6 or 9 months ago was might the model companies, basically put all the software companies out of business. My question now is are we sure there's a lot of value in the models themselves? Because at Bloomreach, we use about 100 models and we have our proprietary models. We use foundational models, we swap them in and out, we use different open source models. The model layer feels very available and increasingly commoditized. And we're starting to find all the value is in these vertical specific domains. For us, that's e-commerce and marketing and really understanding the outcome we're trying to deliver. And building a proprietary stack that delivers on that outcome using the best-in-class model and our own custom models.
John Furrier
>> I like that approach. I want to get into some of the things your company's doing, but I want to just do one more double-click on that because you kind of said a couple things that I like. You kind of need the vertical specialization in the domain.
Raj De Datta
>> Yes.
John Furrier
>> However, you need horizontal scale.
Raj De Datta
>> Yes.
John Furrier
>> Like the cloud brought in. And if you look at the data layers, the data movement is the number one value Yes. So you got to have horizontal scale for the data. But the domain knowledge.
Raj De Datta
>> Yes.
John Furrier
>> Context.
Raj De Datta
>> Yes.
John Furrier
>> Talk about that phenomenon because this is kind of the new thing. Cloud was horizontally scalable. Great. Build a SaaS app, put it in the App Store, put it on Amazon.
Raj De Datta
>> Yeah.
John Furrier
>> Great. Now we're in a whole nother level.
Raj De Datta
>> Now we're in a whole nother level. And what we can see is that you can pick your model of choice. Often it's about the data that trains the model. In our case, it's 7 billion profiles of human beings to train these unique models against an outcome. Ensure that they recursively get better and better as the outcome data sort of flows back into the system, build the associated harnesses, and then deliver more revenue for the client.
John Furrier
>> Yeah.
Raj De Datta
>> And that can only be done if you have the domain experience. And so we really see the future of the application world be vertical-specific agents trained by proprietary models on proprietary data stacks.
John Furrier
>> Yeah. I always say about the SaaSpocalypse, bad software always goes out of business. And so you guys just create better product. Let's talk about Bloomreach. Talk about, um, what you guys do, the vision, and some of the momentum you have right now?
Raj De Datta
>> Yeah, so the vision in many ways hasn't changed. And what we believe is that, there's 7 billion human beings on the web. Every one of them deserve a personalized interaction with a brand. We're gonna deliver on that future. We're gonna go after every brand in the world and give them the tech that makes it possible, first for them to reach all the human beings and increasingly reach their agents to make sure that they can put their best foot forward to make sure you get a delightful experience as a consumer and that the brands make more money serving you in the process. And so much of the AI is the enabler.
John Furrier
>> Talk about some momentum, what's going on with the product customer base. So that's where the value is.
Raj De Datta
>> So what we're seeing in the business is revenues accelerating with AI. We're seeing increased— we're doing all of that without growing headcount. So operating margins are increasing as well. Adoption of AI has skyrocketed. We're seeing, almost half of our customers already use one AI agent or model. And the rate— the number that are using 4 has 23x'd in one year. So we're seeing very substantial takeup of AI within the base.
John Furrier
>> Raj, just this week or actually yesterday, NVIDIA made an announcement on the agentic security thing. 100 companies are involved, kind of like an industry consensus. But the central insight wasn't so much about agentic security or governance or control plane. It was more that there's a bigger story than just security. Security is certainly relevant. Everyone sees what happened with Hugging Face and OpenAI. That's been beaten like a dead horse. But the real thing about agents is they're taking action.
Raj De Datta
>> Yes.
John Furrier
>> And it's not just a security thing. It's whoever can provide that actionable workflow because they're workers, right? You mentioned agents interacting with agents. So when you look at the reformation or the reformulation of an application, its job is to serve the user and get the outcome it needs. So people are talking about, oh, well, with the control plane for the security layer, we got to have all that. But the real value is the action. With agents. Talk about why that's important, how you bake that into a scalable product.
Raj De Datta
>> That's right. I totally agree with that. We're seeing that on the consumer side with Muse and other consumer agents on one side. We're seeing that on the enterprise side with agents that we build. For example, we have a conversational shopping agent that can shop for you. We have a marketing agent that can build campaigns, that can deliver outcomes, that can ensure that they're optimized, that the creative gets built. So increasingly, a lot of the tasks that people were doing in service of outcomes, the agents can increasingly do. And I'll just mention one other thing, which is it's not just about individual users anymore. It's really transforming organizations. Yeah, because these agents can do more breadth of tasks than your average employee can.
John Furrier
>> You have the investor equation, because a lot of investors are here, but you also have a customer-facing landscape that you're vectoring into. When I was looking at the NVIDIA announcement, I wrote this— I'm doing a blog post right now. I'll read you how I'm thinking about it. I want to get your comment. So you think about what's sitting above the infrastructure, the agentic things we talked about, the agentic apps. You got models, agent, runtime, identity, authorization, credentials, data, infrastructure, and audit. Those are things you got to nail down. That's like a value chain. What's indispensable is the identity, delegation, authority, credentials, governance, audit, and hybrid execution. Okay, that's under the hood. That's the platforms that are emerging. Now customers have been dealing with IT, Investors have been looking at SaaS investments. Now you have a whole new world. I might just give a little taste of inside baseball, what's going on under the hood. What's the landscape for investors and customers? Because now they're looking at essentially a generational decision.
Raj De Datta
>> That's right.
John Furrier
>> On their business. That's right. Because they got to make a technology transformation.
Raj De Datta
>> That's right.
John Furrier
>> And a business model transformation. And this is a transformational moment.
Raj De Datta
>> And I get why it's tough for investors to sort through all of this. I think what we would say is if you step back and you think about what gives the application space incredible moats and differentiation, it comes down to 3 things. One is the harness, which is a lot of what you were talking about— governance, security, repeatability, reliability. The second is proprietary vertical AI models trained by proprietary data. Bloomreach is fortunate to be in the position where we have a lot of that. And the third piece is distribution, which is— and domain, which is really understanding the use case so you can train the models against that.
John Furrier
>> So you're comfortable that you have the foundational elements to support that kind of agentic?
Raj De Datta
>> We absolutely do. what we believe is our agents get smarter with every workflow. They get more reliable. They understand the user. So they're personalized. They have memory, they have reliability. Designing for these things is the job.
John Furrier
>> You mentioned data and proprietary data. I think this is an important unlock and it's been talked about kind of in some circles. I think IBM did a study. 1% of enterprise data is being used. Yeah, huge unlock opportunity. There's been a trend from general intelligence or AGI, general intelligence, which Frontier models did. That's right, they crawled the internet.
Raj De Datta
>> That's right.
John Furrier
>> Okay, thank you very much. Got that done. Check. Now you have this specialized intelligence, the rise of companies like what you're doing.
Raj De Datta
>> Yes.
John Furrier
>> Fireworks.
Raj De Datta
>> Yes.
John Furrier
>> All these models can be tuned, whether open weights come in and help. How do you look at that as an entrepreneur? Because this is an opportunity to say, okay, I can get more intelligence and make it a moat for you and the customers.
Raj De Datta
>> Well, we're seeing it because the Loomi AI engine that we have built. That Loomi AI engine has as a dataset 7 billion profiles of actual human beings and their preferences and billions and billions of products. That trained dataset then makes specific models, specifically a lot of machine learning models that then get coupled with LLMs to deliver outcomes for clients, which is more revenue and more impact on their customer acquisition. And what we're seeing is that these proprietary models on proprietary data perform about 5 to 10x better than your out-of-the-box LLMs.
John Furrier
>> Talk about your growth strategy and where you guys are now. Obviously, there's a tailwind. You got everything teed up. You're in the middle of the fairway, as they say here in Silicon Valley. Yeah, you got a lot of work ahead of you, but now you can move fast. But so can the competition potentially.
Raj De Datta
>> Sure.
John Furrier
>> As you look at the landscape, what are you focused on now? What are some of the things you're optimizing for?
Raj De Datta
>> I think when we look at our growth, We just see that our TAM has multiplied because we were in the business of selling software. Now we're in the business of solving for outcomes, which is a labor and software problem ultimately. So the TAM has exploded on one side. The business is experiencing tailwinds because none of these brands want to be commoditized by Google and ChatGPT, they want to retain relationships with clients. So they're doing more communications which drive our business and the agents and AI are delivering better outcomes for them. And we have a business model aligned with those outcomes. All of those are tailwinds. Yeah.
John Furrier
>> And a lot of stuff you guys do in your area— search, discovery, headless— also works. This whole headless trend, no matter what the interface is, you can provide value. Talk about that piece of it, because now as the user interface expectation changes, I just want to get served from whatever platform, you've got to expose that out.
Raj De Datta
>> That's right. And we're seeing— we have the Loomi Connect offering, which exposes it out. We took this intelligence layer and opened it up. And now we are powering a lot of third-party agents. So other people are building agents and saying, I need to call the Bloomreach platform to figure out what customers, what their propensity is, what campaigns, what searches. All of that is being called. And what we've seen is an 83% month-on-month growth in calls to our core platform from third-party agents.
John Furrier
>> You know, Raj, you brought up something about the brand. You said it multiple times. I want to get into the customer because in the old days a brand would have suppliers, they'd have agencies. They would do paid campaigns, pay-per-click. They'd get some email campaigns going. They would try to make their site as merchandisable as possible and have good search. What's changed is that the fear is, wow, they could actually be their own thing. And so that's kind of a new mindset. Share your thoughts on how they are responding to that, which customers are winning, what's the mindset, because what's now happening is they could be their own infrastructure.
Raj De Datta
>> Yeah.
John Furrier
>> By using Bloomreach.
Raj De Datta
>> Yeah.
John Furrier
>> In other words, we're seeing people go direct with media. All aspects of direct-to-consumer, direct-to-customer.
Raj De Datta
>> Yes.
John Furrier
>> It's happening.
Raj De Datta
>> Oh, I think it's happening in a big way. That was a trend that preceded AI. But now you're also dealing with people and their agents. And so increasingly shopping consideration is happening on ChatGPT, happening on Claude. And so the job of Bloomreach and the Loomi AI platform is to partner with brands and say, we're going to help you put your best foot forward. If someone's looking for gear for going on a hike, we want to make sure you get the right gear in front of them at the right time. That doesn't happen automatically. You need a data layer to present that in a manner that the agent can consume. And that's a lot of what we're working on.
John Furrier
>> You know, there's always new things that come up in these waves. I mentioned pay-per-click. That was Web 1.0. You had Web 2.0 affiliate programs and other things. Are there new mechanisms coming out? You mentioned Muse earlier. Interaction is another hot thing where it's just, I can use my instant communications tool, WhatsApp and iMessage to book a flight, go shopping. Amazon shut them down. Okay, so you're starting to see now cross-domain agentic. What is the opportunity there and what needs to be solved? Obviously delegation, one of the things I mentioned earlier, is important, but trust is important. What does that look like? What new services have come out of this wave? Because the game is still the same. I still want to find what I'm looking for. I want to buy it. Do I like it? Do my friends like
Raj De Datta
>> it?Are there new things emerging around that?
John Furrier
>> Around this?
Raj De Datta
>> Yeah, I think the first thing to observe is that it's a great opportunity for brands because what we know from the history of commerce in 25 years is when you reduce friction to purchase, people buy more. And agents do exactly that. So now the brands are really coming to us saying, how do I be present in serving those? How do I make sure? And how do I make sure that the experience I've curated for 20 years on my website actually is harnessed into the agent that's shopping on the consumer's behalf? And we see that as a big opportunity for the brand, a big opportunity for us. And certainly consumers are going to have a better experience.
John Furrier
>> I love that kind of first principle. If you make things— reduce the steps it takes to do something, make it easy to use and simple to understand. That's a business model home run.
Raj De Datta
>> That's right.
John Furrier
>> Talk about the investment side of your company. A lot of investors here at the BofA conference. It's a private company event, even for the ones that have escape velocity. You guys have got a great architecture doing well. What's on your mind on funding growth? IPO? What's the— what's the trajectory look like for
Raj De Datta
>> you?I think we see all of those in our future. fortunately, the company is free cash flow positive and doesn't need capital, but we're always thinking about the
John Furrier
>> data.Databricks said that too. That's right. On K, a version, version K, Series K, whatever letter they're
Raj De Datta
>> on.That is the point is there's always step function opportunities and step function opportunities need capital, whether they be for M&A or organic growth. And we're focused on doing that and we'll tap markets at the right
John Furrier
>> time.Talk about the organic versus inorganic growth, because I think there is a huge capital opportunity for you. Obviously, you're cash flow positive and your TAM's expanded with the action agent side of it. It's not just software as a service. You got a lot there. What's— what would be your strategy on the organic and inorganic?
Raj De Datta
>> Yeah, I think we're doing both. I mean, our organic business is growing fast, but we're also looking at a lot of startups in the AI space. I think we can be a great home for a lot of that tech. We do have distribution. We do have a common platform. Our perspective with M&A, and we've done 5 deals over the course of Bloomreach's history is it's got to be core to the product strategy and the problem we're trying to solve for the client. If it is, we're open to anything, but we're not in the business of being a financial buyer where we say, hey, let's just buy low, sell high and amalgamate a collection of assets. That's not the business that we're in.
John Furrier
>> Well, you're certainly enabling a great user experience. Final question for you. Obviously, whether you're doing the core competency work, whether you do the inorganic and organic partnerships and ecosystem becomes huge. Where are the white spaces for folks to partner with you? What would be your message there for folks watching and saying, hey, I love this company, got a platform, it's on the right trajectory. How do I partner? Is there white space? I can maybe build a tool. What are you— what's your thoughts on this?
Raj De Datta
>> Yeah, we certainly have a set of core strategic partnerships with folks like Databricks and Snowflake and Google and Shopify and others. But then there's the large group of innovators and ISVs that want to build tech. And what I would say is the Loomi Connect platform, one of the best things we've done is open it up. And so anybody can build an agent if they're focused on the domain for marketing and commerce. We expect to be the central nervous system for that domain. Come to the Loomi Connect platform, build it, go sell it to our clients.
John Furrier
>> Well, certainly the commerce, martech stack, all these previous things are being abstracted away. Some go away, some stay. But system of record is a great thing. Systems of intelligence brings a whole nother opportunity for agents in action. That's right. And results. Congratulations.
Raj De Datta
>> Thank you, appreciate it.
John Furrier
>> I'm John Furrier. We are here in Palo Alto again, the escape velocity of the new generation of companies that have hit the formula of good software, good architecture, good platforms, using intelligence to create new expectations and experience for users. And ultimately outcome, tokens equal revenue, good intelligence plugged into a product makes it better. We're doing our part bringing that to you. Thanks for watching.