Inside Neo4j’s Platform for AI-Driven Decision Making

Hey everyone. Welcome back here to Tech Drunk tv. I'm happy to invite you or introduce you to my next guest from Neo four J and know it's not Steven.

Uh, Steven Chin, of course, is with Neil four J. We've had the pleasure of interviewing Steven all the time, but sometimes you wanna get a different view. Maybe someone who's a little smarter, maybe, uh, probably not as much hair, but let me introduce you to Su Sudir.

Hospi. Sudir is the Chief product officer, CPO at Neo four J, and we are lucky enough to have him join us today. Sudir, it's a pleasure to have you on here.

Thank you for joining us. It's great to be here, Alan, and looking forward to the conversation today. Absolutely, and no disrespect to my Fred Steven Chin.

We love him. But, uh, UD dear, as I mentioned, you're the CPO at at Neo four J. Why don't we maybe start off with your story before we jump into everything else?

Yeah, so I, I've been with Neo four J for almost two and a half plus years now. Um, I am responsible for all our product strategy, product execution, roadmap, everything. I work with customers across the globe trying to understand what challenges they have and how do we build our graph intelligence platform that enables them to solve problems.

Uh, before coming to Neo four JI actually ran a product for all of data analytics services at Google Cloud. So this meant anything that allowed you to bring your large scale data assets into Google Cloud, how do you process it, how do you actually analyze it? Uh, and I also was responsible for multiple acquisitions, including Looker.

So basically the bi side of the house, so everything. So BigQuery, which is one of the largest products there, that one plus another 10 odd services. So I ran that for five plus years, uh, at Google before this.

Uh, but no, I think my opportunity I saw was with large language models just getting ready to take off, I saw there was a huge opportunity for knowledge graphs to provide that intelligence layer or knowledge layer for these systems, and that's why I came to Neo four J. Absolutely. Um, it's a great story and they're lucky to have you there.

You know, we, we, we mentioned Neo four JA bunch of times and, and people who watch text drug TV all the time. Thank you. Um, but beyond that, you, you've probably seen us interview Neil four J and we cover them on our various websites as well.

But sudir, I'm sure there are people out here who don't know Neo four J or maybe they know a little bit about Neo four J, but you know, not the whole story. If, if I, if I can bother you, can you kind of, you know, just lay that as a foundation. Give us the Neo four J background.

Yeah, so Neo four J the company was started almost 16 plus years back. Uh, we are in the, we are the category creators for the graph database category. Uh, we now are the graph intelligence platform, uh, that enables, uh, developers to build, uh, agent applications by converting their data into knowledge, right?

So all these AI systems need to access enterprise knowledge. How else are they going to make better decisions? How are they going to be accurate in their, their information and all?

And so that's what we enable, uh, customers to do. Uh, as part of our graph intelligence platform. The core components are our graph database.

That's what we have built over 16 plus years. We have pioneers in, in the graph space from that perspective. We also have graph algorithms.

So we have 65 plus algorithms that you can use to run intelligent, uh, analysis intelligent, uh, decisions on top of the graph data that you may have. But our new capability allows you to run these algorithms on any data anywhere in the enterprise. So you may have data and Databricks or Snowflake or BigQuery.

You can still run these algorithms on top of it. So that's the core part of our graph intelligence platform. The layer above that is AI power tools.

So you can literally, I have this theme for our product portfolio where I want people to be able to, in five seconds sign up for our service in five minutes, actually use their data and get wowed. And then in five days you should get to value. And so in this case, you can literally take your data from Databricks, convert it into a graph data model in like three clicks, and then from there build agents on top of it within few minutes.

So that's the, the AI power tools that we have built out for that. So automated, uh, graph model generation and all. And then at the top of it is our AI stack, which is our, or our agents.

How do you create your own agents and how do you go ahead and, and like, you know, we, we are backbone for many of the memory companies, uh, in the, in the space. So that's the graph intelligence platform. That's what we provide.

Many of our customers use us for various use cases in, uh, agent ai, primarily as the core knowledge layer that can power these AI systems. But also in financial services. We are big in fraud detection, anti-money laundering use cases.

Every supply chain company uses graphs to go ahead and manage the supply chain risk assessment and all, um, all the healthcare life sciences companies use us for the whole, uh, knowledge graph for all of their r and d graphs. Like, hey, what drugs have been like, you know, uh, identified, what are the things they solve, what enzymes, all of that kind of, uh, actual knowledge graph. Um, US Army uses us for all of their supply chain.

So a lot of intelligence analytics and all is done, uh, by me, various intelligence agencies on top of our, uh, our software. So that's, that's roughly the space we are in. I love it.

And, and thank you for taking the time and explaining that, you know, look, we, we live in an, this whole LLM and AI and rag and vector databases, right? Graph has really kind of found its place, if you will, 'cause it it, it's great technology and it, you know, as you mentioned, NEO four j's around 16 years. We've had graph databases all these years and everyone knew what, you know, there were so many great use cases, but this may in fact be the killer use case Yes.

For, for graph database. Um, just before we jump into the topic of discussion, for people who maybe wanna find out more about Neo four J, where, where was, what's the best? Just go to the website or Yeah.

com is a good place to go start. Uh, there are multiple videos we run. One of the things Steven Chin, our friend runs is a devel for US developer relationships and his team manages this Graph Academy.

So if you want to learn about graphs and what graph technology can do, we have Graph Academy that has lots of courses for audiences. They can go ahead and learn everything. How do we build an application getting started to more advanced courses?

Uh, so that's another good resource. Another resources, if you go on YouTube and search for NEO four J, you'll find tons of content from various of our conferences that we run, including a lot of customer, uh, driven content, customer stories and all that. Love it.

Good stuff. Alright, let, let's pivot a little bit. You guys recently announced GA of NEO ga, meaning general availability.

For those out there who may not be familiar, uh, for the NEO four J fleet manager, which is advertised as the industry's first unified control plane for graph databases. Give us the scoop here. Sudir.

Yeah, so we have tons of customers that actually use Neo four J large enterprises, almost I think 80 plus percent of, uh, fortune hundred users. Uh, and then a lot of these customers have various use cases, like they have been using us for let's say fraud or anti-money laundering or supply chain and other things that we talked about. And as the agent AI is becoming more and more popular, they have newer use cases, they want to go ahead and use us for, for, for like just the knowledge layer to power these assets and especially knowledge layer.

What I mean by that is you may have data and disparate systems, your LLMs are not going to get access to all these systems and be able to make sense of it. So what we allow people to do is take this data and create a semantic layer on top of that data as the knowledge layer and then your LLMs can use graph rack to go get access to that information in a very secure, trusted manner. And that improves the accuracy, reduces hallucinations.

So as the new use cases were coming, we saw a lot of modern new use cases were getting deployed in our cloud offering, which is Aura. It's a fully managed offering, zero operations. We take care of all the infrastructure, but lot of existing use cases may be running, they run it themselves in one of the clouds or like, you know, 15, 20% of our customers still run their own data centers.

And so they're running in all of these, uh, environments. Additionally, we also have databases that are, are community edition, which is completely free edition that anybody wants to build on graph. They can take our open source database and just start building applications.

We have a lot of adoption of that, but one of the challenges for IT leadership and CIOs is okay, when you have that much deployment in an organization, how do you monitor it? How do you manage it? How do you know what is happening across the fleet?

So the fleet manager gives you a single pane of glass to look at all your deployments, whether they are on-prem, on-prem, run by yourself in cloud or our fully managed aura offering. Whether it is our enterprise offering or whether it's our community edition, which is like completely free open source offering, it doesn't matter. You go, you can actually visualize and see all your deployments in one single place.

You can look at them, you can monitor them, you can operationally figure out what actions you want to take if something goes wrong. It also will tell you any security risks if something is going wrong, like, you know, last time when the lock four J issue happens, something like, if that happens in future, we will be able to alert the CIOs about what the deployments look like. So I think the most important thing is single pane of glass.

Any deployment anywhere in any platform, we can give you complete visibility into it. I love it. And of course, general availability means it's generally available right now.

People can, it's available Now. Any customer who wants to use it, especially if you are in our Aura, uh, database user, you will by default start using it. If you're self-managed, you can register your self-managed databases, start using it in production scale.

I love it. So dear. You know, I'm doing a few more interviews this week.

Won't be doing any next week as we, you know, break here for the holidays and the new year. You know, it's been a heck of a year right? By any, no matter how you wanna measure it.

Yes. Um, as you look ahead to 2026 with this, is it more of the same? Is it accelerating?

Is it a black hole? We don't know what, you know, as part of your role there as CPO, what, what is your, what's your gut telling you? I think I work with lot of customers as I said.

Right? One of the things, and just one interesting data point in last, in this year, 2025, I was just doing my measurement. I have been on road for 24 weeks of this, like Out of 52, this goal, Yeah.

Out of 24, majority of that is actually going out, meeting customers, spending time talking to them and all. And what I have seen is 2025 has been a pivotal year in the switch from lot of experimentation with Agen ai or AI in general, LLMs rag, all that, that happened in 20 23, 20 24, 20 25. We have seen real use cases being deployed by large customers.

Like our recent, uh, graph, like, you know, uh, we, we do these, um, events with our customers. In the recent one we had Walmart, we had Uber, we have had, uh, uh, Qualis and Brady, which is like an law firm. Like we have had so many different customers actually putting things in production, seeing value from what they're, what they're doing.

No, nor disk has been there. So I think the thing is, I've seen like in 2025, lot of these customers actually getting value and deploying things to production in 2026. I believe that's only going to accelerate because I still believe a lot of our customers have been the early adopters that have transitioned from experimentation to real business value.

But there's this majority that hasn't actually seen this. So 2026, I think it's, uh, it's the acceleration of like, you know, value driven, use case driven implementations of AI that I think will happen. And I think we can really help these customers, especially developers, get value of a from the AI systems.

I love it. Udia, thanks for coming up here on Techstrong tv. You know, I do interview Steven a lot, but anytime you want to come back, you let us know there's a place here for you.

Okay. Thank you Alan. I really enjoyed talking to you and thank you, uh, for bringing me here.

Also, happy holidays to all your audiences and I'm looking forward to coming back and sharing more as we go into 2026. Thank you. Well Spend a few weeks at home during the holidays at least.

Then it will, if I don't see you on here, we'll see you on the road. Udia, CPO Neil, four J here on Textron tv. We're gonna take a break.

We'll be back.