In this episode, host Michael Bernzweig talks with Matt Soltau, Global Director of Strategy and Operations at IntelliPaaS, about how enterprises can make their fragmented tech stacks truly AI‑ready. Matt explains why point‑to‑point “spaghetti” integrations create data silos, security risk, and failed AI projects, and shares how a centralized integration layer can deliver real‑time, reliable data across CRM, ERP, ITSM, and other core systems. You’ll hear practical steps to audit your current stack, create a single source of truth without yet another data lake, and design scalable, low‑code/enterprise‑grade integrations that turn your existing tools into a competitive advantage.
In this episode, host Michael Bernzweig talks with Matt Soltau, Global Director of Strategy and Operations at IntelliPaaS, about how enterprises can make their fragmented tech stacks truly AI‑ready. Matt explains why point‑to‑point “spaghetti” integrations create data silos, security risk, and failed AI projects, and shares how a centralized integration layer can deliver real‑time, reliable data across CRM, ERP, ITSM, and other core systems. You’ll hear practical steps to audit your current stack, create a single source of truth without yet another data lake, and design scalable, low‑code/enterprise‑grade integrations that turn your existing tools into a competitive advantage.
Key Topics:
Timestamps:
00:00 - Introduction to Matt Soltau's personal and professional journey
02:11 - The role of curiosity and international experience in shaping leadership
04:42 - Overview of IntelliPaaS's mission and core capabilities
05:09 - How IntelliPaaS supports regulated industries in real-time data management
08:42 - Strategies for deploying on-premise, cloud, and air-gapped environments
09:34 - Building scalable integration architecture: from point-to-point to hub models
11:15 - Using data stamps and audit trails for compliance in cross-border data use
13:01 - Unlocking AI value through integrated, real-time data pipelines
16:36 - Shifting from traditional point-to-point to centralized integration platforms
18:40 - Practical example: Transitioning from spaghetti architecture to hub-and-spoke
20:31 - Enabling AI readiness with flexible deployment and high data control
22:17 - Addressing AI hallucination through robust data pipelines
23:26 - The future of digital transformation and AI integration in enterprises
24:19 - Transformations in AI-driven data management and automation
25:49 - The importance of single source of truth and data normalization
27:33 - Applying AI to unstructured data like emails for automation
29:28 - How to evolve B2B marketing with AI-informed customer insights
31:35 - Differentiating a community movement from a simple community
34:40 - Signals of waterfall thinking in QA teams and how to adapt
37:09 - Questions to uncover customer pain points through interviews
38:25 - Metrics for measuring audience engagement before fan art & tattoo stage
40:28 - Designing community journeys to increase contributor engagement
48:29 - Alleviating burnout through targeted community engagement strategies
48:48 - Aligning developers, QA, and security for organizational quality ownership
50:23 - The playbook for cleaning and transforming data for reliable AI actions
53:33 - Closing remarks and next steps for attendees
Michael Bernzweig (00:06.007)
I hope everyone enjoyed that last presentation. Coming up next, have Matt Soltau He is the Global Director, Strategy and Operations at IntelliPaaS, Inc. Shaping AI powered integrations and automation platforms that streamline data, workflows and customer experiences. With that, Matt, welcome to the event.
Matt Soltau ( IntelliPaaS ) (00:31.543)
I appreciate it. Thanks for having me on.
Yeah, IntelliPaaS is an enterprise integration platform really focusing on transforming data journeys and getting data AI ready, which is obviously a hot topic and something that's been, well, not just well discussed here now, but I'm sure there's going to be a couple of follow up conversations. I wanted to take the time to really focus on how you can unlock value in a fragment.
tech stack, really kind of using your entire tech stack as a competitive advantage and kind of, yeah, well getting integration in there to also maybe help enable your AI journey with that as well. So maybe just to quickly start with it, IntelliApp has a global enterprise integration platform. So ultimately what we do is we help connect data. I often make the analogy that we're nothing different to the local plumber.
So we help connect lots of different systems with another most of the times these are you know the larger systems of the world your Salesforce BMC software
service now, you name it, it could be like SAP of course is a big topic for us. But you could also be running through smaller sort of applications as well. So for example, we're in ongoing discussions at the moment with a point of sale provider, very innovative in its space and providing integration solution for that. So anyways, most of the times it's the global enterprise
Matt Soltau ( IntelliPaaS ) (02:16.457)
The majority of European, especially German, manufacturers use our solution. We've got customers in the financial services space, in government, etc. So it's a little bit across the board and back to the plumbing analogy, the plumber doesn't really care where the water goes to, right? So when you're opening the tap, you want to make sure real-time water arrives and ideally that water disappears in real time as well.
We want to make sure exactly that takes place with with data So we send data to any source or a target destination in in real time now What I really want to talk about is kind of getting AI ready and You can see here a quote from Gartner that really has been predicting that well
this was very timely last year, they feel that throughout this year here, about 60 % of organizations will abandon their AI projects really because data isn't AI ready. And this is a big conversation I actually had yesterday with a gentleman in the morning. We were talking about this quite a bit in terms of, know, why does...
Why do AI projects fail? It's a big AI shop. So they create a lot of different AI solutions as well. The long and short is we were very quick to agree that data is often not AI ready. So we want to make sure that we take that data, which sits in so many different fragmented systems, and we deliver that data to where it needs to be. The right data at the right place at the right time.
And in order to achieve that, you just need to have the data pipes ready in order to get that data there, but also to enable transformation of that data to really make that data that you do have actionable.
Matt Soltau ( IntelliPaaS ) (04:17.784)
Now integration isn't new. It's been around for a while but there's a traditional way of looking at integrations and there's obviously a more modern approach. In the traditional sense you're connecting in a point-to-point scenario. So you've got one point, you've got another point and another endpoint and so on and you're trying to connect them all as much as possible. Now
The average organization has, and we're talking enterprise here, has over 80 different applications.
Imagine trying to connect them all with another. It's like a huge spider web. As you can see, it's not scalable, costs a lot, not just to deploy, but also to maintain. And really as a result, you've got all sorts of data silos. Now, because of all these point to point connections as well, there's a serious security risk and also compliance risk in some cases, especially when it comes to cross border data transfer. So basically what we break all of that up,
by not having point-to-point connections but kind of moving more towards a centralized approach with that. So the point-to-point integrations we sometimes also call the spaghetti architecture because as you can see here we're trying to illustrate these are only eight applications. If you want to connect all of them with another that's obviously a lot of individual point-to-point connections you need to achieve, want to create and also to maintain.
So, and you can imagine now this times, you know, 10, for example, if we're only taking like, well, the average we said earlier around 80 applications, it's completely unmanageable. And as you can imagine, there's going to be some, some applications that are connected, some that are not, but really not across the entire organization itself.
Matt Soltau ( IntelliPaaS ) (06:10.724)
That's where IntelliPaaS comes in. We're an integration platform that really sits at the heart of it all. Rather than connecting everything with everything all the time, you basically just have to connect an application once. And then we've got a drag and drop connector that allows you to create this integration flow yourself. Or obviously we've got a professional services team that can help with that. We've got partners all over the world and to get the integration journey.
up and running. The important part is we don't store the data. So you keep the data where you need it to be. And the second part is we've got global deployment flexibility. And this is one of the unique parts because we can deploy the platform anywhere. We typically deploy on Azure. So we've got our own Azure services all around the world, but we've got customers that want to run on AWS. No problem. In fact, if, if
You're running on AWS. You obviously have a spend commitment to AWS or any other hyperscaler could be Google could be anyone So we can deploy on that which means that any of the load that kind of goes through the platform Really goes against your spend commitment at the same time So it doesn't have to be on our side can be directly on theirs. We can deploy it on
service of course we can deploy it in a containerized environment. There's we mentioned cloud infrastructure we can deploy completely on-premise. In fact we've got customers in the high security space as well that are associated to yeah well the military in the end. They end up deploying it in a air-gapped environment so no internet in no internet out it's completely kind of off the grid so to speak. Now
What does all of this enable? We were talking about lots of different data, being in lots of different places, so we suddenly have access to a lot of data. And what you heard before is that we can ensure the... Can we skip this part here? Because I realized that the Hong Kong part is in here. Sorry, team. So let me just go from here. So...
Matt Soltau ( IntelliPaaS ) (08:29.798)
How we do all of this is we've got a number of out of the box connectors. With these out of the box connectors, we've got easy out of the box intra-connectivity into the systems, which basically means we've got kind of a plug and play connection into all of that. Now, in detail, this means that once you've got a connector going, you can literally just drag and drop it into the flow itself. You connect it.
using drag and drop and then the integration is up and running. It's actually super simple to set up and again can be deployed anywhere so it's completely enterprise ready as well.
Now for those that want to go a little bit further down the rabbit hole when it comes to coding, we've got the low code approach that you just saw, but we also have the pro code part where you can customize absolutely everything. So we support Python, Java, Groovy, you can do it all.
If there's no out of the box connector available, you can just use our technology connectors, which basically means through REST or PHP or we've got flat file support, databases, et cetera, you'll be able to connect to any sort of data source in the end.
So kind of just as a summary, global deployment flexibility, we can make sure the entire data goes everywhere. And as part of that, now tying all of this to AI is making sure that your data that you want to get into AI really has like the AI system itself needs to have real time integration access to that data. We were talking about hallucination at some point.
Matt Soltau ( IntelliPaaS ) (10:19.13)
This is really the reason why AIs hallucinate. This is the reason why a number of organizations at large try to give up their data integration journey and their AI projects with that as well. So if we get the data AI ready, if we connect the entire systems, we'll make sure that, and this is still, I love this one here. This is a couple of years old, but as relevant as ever now.
because in 2024 Forbes said that there is a two-year time frame to make a significant leap ahead when it comes to digital transformation or otherwise fall behind. Well I guess that's as relevant as it is now because the years I guess 2024 was a completely different world in many respects.
Matt Soltau ( IntelliPaaS ) (11:14.616)
It's certainly interesting to see how far some customers, and I'm sure the audience has come as well, on that journey. And yeah, well, we look forward to continuing to support digital transformations and making AI much more of a reality. So, yeah. Thank you, Michael.
Michael Bernzweig (11:35.075)
Matt, I think that a lot of the audience listening today is somewhere along the continuum of that journey. And as we are into 2026, I think a lot of founders are hearing from their team, what is our digital plan? And I know that a lot of
Matt Soltau ( IntelliPaaS ) (11:46.855)
Ahem. Ahem.
Michael Bernzweig (12:01.337)
your clients that you've worked with, you know, in different aspects of AI over this past year are feeling the same thing. So what are some of the types of transformations that you're helping clients achieve in terms of some of the, you know, digital transformations that they're making related to AI with your help?
Matt Soltau ( IntelliPaaS ) (12:31.238)
Yeah, that's a good question. For the most part, it's a fragmented tech stack that companies have. So duplication of data, all data sitting in one part, very often at some point or another, either tech stacks have been upgraded or...
through acquisitions as well. You've got a lot of additional tech debt that you're taking on as well. The best recommendation that we always have is create a single source of truth.
Not in the sense of a data lake because that just means collecting everything for the day if and that can be helpful in some scenarios but making the right data available to the right tool at the right time and as a result of all of this Digital transformation has to take place right so for example, I'll give I'll give an example between two CRM systems in one CRM system, it might be
first name, last name, or maybe even the title altogether in one field. The next CRM system basically needs that broken up into two or three different fields depending on how that's structured. So that is a very simple way of obviously then having to break data apart and transform it. In some other ways, we're deploying our AI agents for that. And excuse me.
We are deploying our AI agents for that, that basically sit in the middle of the digital transformation to help make sense of the incoming data. Actually yesterday we had a workshop with a client based in Hong Kong.
Matt Soltau ( IntelliPaaS ) (14:27.186)
And the challenge there was taking incoming unstructured emails that were coming in. In this case, this was related to IT service management. But they had lots of incoming unstructured emails of various different kinds of requests.
Now, making sense of that very crowded inbox was a big problem for the team because they had to manually go through all of that. So basically what we did is we've created solution for them. where we have basically we've got an email tool that can listen to any emails. We can take that in and then we've got various different flows set up in the background depending on what type of request is coming in. One of it might be an outage request, right? So that needs to be treated as a higher priority,
else might be a password reset, something else might be whatever, something else. So this is specific to IT service management here now.
But what we then do is we take that data, we analyze it using AI, which in that case is deployed locally. And we turn that into real time tickets, which are then automatically categorized into high, medium, low urgency, depending on what's coming in. And if there's some emails that really are spam emails, for example, which happens, then they completely skip the inbox altogether.
So that's also where we can take that data and take unstructured data using our own proprietary AI solutions and pushing this directly to an output.
Michael Bernzweig (15:58.981)
Fantastic. Well, a lot of questions I can see are already coming in. For anyone that is asking a question maybe related to their organization, just type it into the Q &A that you see in front of you, and we'll get to as many of those as we can in the upcoming Q &A session. And we will take it from there.
Matt Soltau ( IntelliPaaS ) (16:24.593)
Sounds good.