How Tugger Turns Scattered Business Systems Into Trusted AI Answers

Your AI assistant is only as smart as the mess behind your dashboards. When business data lives across CRM, accounting, HR, and job systems, “connect ChatGPT to our data” quickly turns into rate limits, broken joins, confusing IDs, and answers nobody trusts.

We sit down with Craig Morrall, co-founder of Tugger, to unpack a practical architecture for enterprise AI that actually holds up in the real world: pulling data from many platforms into a warehouse, then layering on a semantic model that explains what the data means and how records connect across systems. That extra context is what turns a chatbot into something you can rely on for revenue questions, profitability analysis, and cross-platform reporting without spending months on custom pipelines.

Craig also shares what customers are doing once the foundation is in place, including building interactive dashboards in minutes and generating repeatable board packs that used to take finance teams hours. We dig into time to value, early ROI stories, and how Tugger approaches security and governance with ring-fenced data storage, ISO 27001 certification, and guidance on using business-grade LLM plans to reduce training risk.

If you’re evaluating enterprise AI, data warehousing, semantic layers, or secure analytics with Claude or ChatGPT, this conversation will help you separate real capability from hype. Subscribe for more practical AI stories, share this with a friend building on enterprise data, and leave a review with the biggest data problem you want AI to solve.

Everyday AI: Your daily guide to grown with Generative AI
Can't keep up with AI? We've got you. Everyday AI helps you keep up and get ahead.

Listen on: Apple Podcasts   Spotify

Transcript

Meet Tugger And The Pivot

SPEAKER_00 0:00

Hey everyone. Fascinating chat today with Tugger, who is tackling one of the biggest challenges in enterprise AI right now, turning scattered business data into something AI assistants can actually use in a meaningful, secure, and reliable way. Craig, how are you?

SPEAKER_01 0:18

I'm great, Evan. Yeah. Yeah. Great to be here today and uh looking forward to the chat today. Yeah.

SPEAKER_00 0:23

Yeah, really excited for this conversation. The more I learn about the work you're doing, the more intriguing it becomes. Before that, maybe introduce yourself and the team and how do you describe Tugger and the Tugger app, uh, TuggerApp.com uh to the folks viewing?

SPEAKER_01 0:40

Yeah, yeah, sure. So yeah, um, I'm one of the co-founders here at Tugger. Um, started a business a number of years ago, but in recent, I'd say weeks rather than months, probably, we sort of shifted the business quite a bit. Uh historically, we sort of focused around reporting, particularly around sort of Power BI reports and giving people the ability to get insights that way. But what we've done since then is open effectively open up our platform so that people can then connect AI tools to it. So things like Claude and ChatGPT so that can then use these to get insights out of the businesses, the business data they couldn't do before, effectively.

AI Assistants Become The New Workspace

SPEAKER_00 1:21

Yeah, it's an intriguing time, and uh you're right at the center of it. And you know, as you know, AI assistants are becoming the front door for how people work with information and data. How do you see yourself fitting into this new workflow with Claude and ChatGPT and Gemini and more?

SPEAKER_01 1:39

Yeah, well, the way we sort of see it is that you know, tools like Claude, Chat GPT, they're almost sort of like the nucleus, if you like. They're starting to sit on people's desktops, people's phones, and people are using them all the time. A bit like sort of like, you know, when people started using web browsers for the first time. They're all you know, we're all in we're all in web browsers every day. And the way we see it is that people are gonna start you start working with their business data in different ways. So historically, you'd log into a platform, whether that be Salesforce, Hotspot Zero, you do your work in there. A lot of the work is going to start to shift into tools like Claude, tools like Chat GPT, things like that. And that's effectively where we sort of fit in. Rather than going down the route of we could have, we could have easily built an AI layer in our own platform, but actually, we'd want to be part of that ecosystem. We want to be bringing that data into the hands of people in the tools they're going to use every single day, which means that they can go into Claude, Chat GPT or whatever, and say, well, actually, tell me, um, you know, tell me what our sales revenues were last month, you know, chart that for me, you know, build that into a dashboard and let and give basically give them just so much more control over that data than what they've ever had before.

Why Plug And Play Fails

SPEAKER_00 2:59

Brilliant. And a lot of companies think connecting AI to enterprise data is, you know, kind of plug and play. Um, even folks who are familiar with manipulating data and Excel and tools, that turns out to be a lot harder in reality and in practice than is uh in theory. So, what what are some of the challenges there?

SPEAKER_01 3:20

Yeah, so you probably see you probably see a lot of um you know platforms out there, particularly SaaS platforms, they're building what they call MCP servers on top of what they've got. Those MCP servers, they're a bit like a a USB-C cable, if you like, that can interchange between charging one phone and another or charging whatever device it likes. These MPC servers are giving tools like Claude and Chat GPT or Gemini that ability to talk to that data, if you like. But what all these sort of all these tools do are allowing them to connect to it and talk to it. But what they're not sort of doing at the minute is making it and you know, you can't do things like you know, cross-platform queries and things like that. These MCP tools and stuff that are in there, they're not geared up in that way at the minute, and they probably won't ever be geared up in that way. They're effectively geared up so that you can, you know, do things like, you know, right, create me a deal or whatever, or go and re go and do something with this, you know, sale or whatever it may be, but they're not designed around sort of reporting, if you like. That's not where these things sit right now. And so the problem you've got is if you go and say, Okay, well, I want to go and report on the HobSpot data or my zero data, you hook up to that, but actually, what you'll find very quickly is that you're hitting things like rate limits on APIs, and those APIs could change if you like. Our platform effectively has its own MCP that that sits in between these, and we've got a data warehouse behind it. So, what we're effectively doing is looking at bringing that data from all these platforms into our warehouse. We've then effectively got a semantic layer, if you like, that sits on top of that. And by that I mean what that's doing is that's showing how all this data is linked together. So that's saying, you know, this job is linked to that quote, this quote is linked to that sort of salesperson. But not only that, we're also saying, you know, we're actually hinting and telling the AI that actually this might be a record in zero, which has got, you know, it's an invoice, it's linked to this customer, but that actually also links to your HubSpot customer record. It refers back to the deal that was won by this salesperson, and it's this interlinking of all this data together to get to the point where you can then analyse it and get all the insights out that you want that's sort of key. And you know, there's a lot of people building MCP servers out there, but they're geared at one thing, they're geared at solving things just within zero or sold in solving things just within HubSpot or whatever. And we're going, well, actually, you want to solve things across the board, you want everything joined together, you want to get one source of the truth and really get really delve into that data so that you can run your business more efficiently, you know, get those time savings in there, you know, increase your GP, whatever it may be.

SPEAKER_00 6:19

Brilliant. Uh the intriguing approach. What led you there to this warehouse-centric kind of model, maybe hard experience?

SPEAKER_01 6:28

It is, it is, it is hard experience. So um, I mean, when we originally built Tugger before we sort of made this pivot into the AI space, we the whole point in Tugger was to kind of make it easier for SMEs, if you like, to work with their data and build out reports. It's very, very, you know, costly, very, very expensive, time consuming to set up data warehouses, if you like. You know, where wherever you host them, you need DevOps engineers, you need people to look after the security, all that needs to be ordered and all the rest. So, what we thought some time ago is actually let's simplify all the process, let's handle all that data warehousing for people. And that means that then, and then we're talking back in the day now, people could then connect up to RBI and or Tableau or even Excel up into that warehouse and start to build out the Rose reports, do all that analysis without any of those complications. And then what we realized earlier this year is actually now we've got all this data that our customers are holding, that's even better because now the next problem is in this world of AI is you know, people want to have that AI interact with that data in different ways. But we've already got that data, so we can then build a layer on top of that that makes that data easy for the AI to understand. I mean, the problem you've got is if you say, right, okay, well, I'm gonna plug, I'm just gonna go and store my data in this warehouse. There it is, it's thrown in there. If you then go and plug Claude or whatever straight at that warehouse, yes, it'll be able to write database queries, yes, it'll be able to do stuff, but it won't have a clue that you know, job status ID seven, what does that even mean? It doesn't know that might be a completed job, it might be a job that's in progress or or whatever, it doesn't sort of know all that. And so we thought, well, surely the right thing here to do is to kind of go, well, we've got the warehouse, let's build a layer on top of that, and let's make it really, really easy for people to use.

Surprising Use Cases From Customers

SPEAKER_00 8:29

Brilliant. And so, you know, when you launched uh AI Insights, uh I you know, the idea of you know, ask your business data anything sounds so intriguing. Um, what are customers actually asking and any surprise use cases or usage that you haven't thought about?

SPEAKER_01 8:48

I mean, weirdly enough, I mean, like you say, I mean, our strap line is like you say, you know, ask your business data or anything. But weirdly enough, that's not necessarily what people are doing with it. So yes, yes, they are doing that, but they're doing a lot more than that. There's still a concept, uh, particularly outside of the tech world, that you know, tools like ChatGPT, that they're there to help you write emails ahead, there to write market and copy it, but actually you can do so much more with them. So, what we're actually finding is that people have, you know, a particular sort of when they've got Claude hooked up to it, they're not just asking questions, you know, you know, how profitable was my job last week or whatever. They're actually using Claude then to build out interactive dashboards that sat on top of the data that our platforms then servicing. So those those reports could be stuff that they're dashboards that then share out with other people in your organization, and you know, those dashboards are typically interactive as well. So they've almost done what we would have historically done just with Power BI, they've gone and done that themselves, but they've done it in like minutes as opposed to days or weeks in terms of building those out. And the other things that we've seen as well is because you know, as I said, as I mentioned before, Claude and you know, these tools will be the nucleus, but then it they're also doing things like you know, let's let's build a board pack out of this, let's build presentations. You know, I was literally just before we did this call, I was talking to someone who's actually trialing the product at the minute. I said, you know, I said, Oh, what are you doing with it? I said, Oh, literally. I have said I have somebody working for us, you know, our financial controllers, they're typically putting me board cap packs together, you know, on a weekly basis to build out uh, you know, building out on the financials what it's looking like, and that will take them hours and hours to do that. So he said, I don't even need to get them to do that now. I said I can literally just click a button because I've done it before, and out comes out the board pack again, and that's it's done. So it's it's becoming a way, it's be it's they're using it for more as almost like a process tool, I guess, in some ways as well. But yeah, it's honestly it's it's shocking what people are doing. And literally, we just keep talking to you know customers and people on Transit, you what are you actually doing? Because we're trying to understand more and more what you know where people are going with it.

SPEAKER_00 11:45

Yeah, so so exciting. That's the CFO's dream, what you're describing there are so uh and crazy interesting use cases. And um, you know, one thing I'm experiencing many are sort of tool overload. Um you know, so many trying so many interesting tools, some work out, some don't. How how quickly can a customer go from you know signing up with Tugger at tuggerapp.com to getting a useful answer or insight? What's that time horizon look like?

SPEAKER_01 12:15

It's pretty quick in reality. So they basically they can be connected up to a platform within minutes. So that could be you know Xero, QuickBooks, Sage, whatever it may be, can connect that up. And then, of course, obviously we're warehounds in the data, it can take a little bit longer to get the data across. Sometimes they can, you know, it's ready within an hour, sometimes it might be the following day. Um, but if you look at the fact that if it takes you a day to get to the point where you can start asking questions, historically, you would have to go and actually get someone to go and write some code against a s a ton of APIs, whether that be APIs for QuickBooks or whatever, you'd say you'd need that. You'd also then need somebody to go and build you out a data warehouse, you'd have to secure that down. So you're weeks and months into the point before you even get to that, and then you've got no AI on top of that either. So you so when you compare it to where you're going, it's literally within a day, you're pretty much good to go.

SPEAKER_00 13:14

That's fantastic. And you know, everyone's talking about productivity gains and asking tough ROI questions that are not simple to answer. Um are you starting to see any hard numbers or or feedback from customers yet?

SPEAKER_01 13:30

Yeah, I mean, as I've touched on the on that before, in terms of the hours and hours of saving in putting together like, you know, uh board packs and things like that. Um, you know, there's other customers that are seeing increases in GP and stuff. Um one particular use case where a customer um they, you know, you know, in the SME world, they're a four million pound turnover business. Um this this would be they were they're an electrical contractor, if you like. And what they'd see on that four million is an increase after using our platform um within like a few months of literally it was a half million GP. So it's a half a million increase in GP they got from it. So they're very happy.

Scaling Integrations Across Verticals

SPEAKER_00 14:15

That is that's fantastic, and it's great to see a nice real-world blue-collar uh uh tangible example of a company getting real uh value.

SPEAKER_01 14:29

Honestly, it's crazy. I mean, literally when we launched it, I had literally people ringing me up on the phone and actually going, Oh, well done, guys. So this is just they were just that's they were just so shocked and they'd hooked you up and were like, I said, I can't believe you've done this. I don't know how you've done it, but yeah, it's brilliant. That was the sort. So we were we were we were more shocked than anyone, I think, at the reactions we got.

SPEAKER_00 14:51

That's great. Well, it must be very gratifying and satisfying to hear. And you so you you integrate into a broad variety of systems uh pretty elegantly. What about going deeper into specific verticals or becoming the kind of connective tissue across everything? You mentioned building contractors, construction. I I imagine there are many other verticals that could use your technology.

Security Governance And Hallucinations

SPEAKER_01 15:15

There are. There are a number of verticals. I mean, obviously there's the you know, there's we're we're into you know accountancy type spaces, so anything from you know anything, I mean, anyone who's using any sort of accountancy product, you know, so CFOs, you know, financial controllers, you know, everyone's pretty much got them in their businesses. So you've instantly got you've instantly got that sort of covered, if you like. Um historically we've done a lot in the sort of the field service businesses, so like you say, you've the the blue collar sort of stuff. But now because we effectively opened up to a number of platforms, at the start of the year we were about 24 platforms we integrated into. Now, as of today, we're up to 52 platforms, so we've you know, so we've increased, you know, we've more than doubled with it within the space, you know, in less than half a year, if you like, and that we sort of accelerate, and if you like, so we can effectively go into so many different verticals, but where it really becomes powerful is the fact that because we're connected into different platforms, there's different elements of different businesses that can use us. You know, we're hooked up into HR platforms, we're hooked up into CRM platforms, um, you know, hooked up into job management platforms, accountancy platforms. So you've got different areas of different businesses that can sort of move in on this. And yeah, I mean, it's we've got a huge roadmap of different platforms that we're going to build out. So, and what's great is we can do it, you know, relatively quickly in terms of launching out, you know, new integrations with new platforms. And we, you know, our customers are even saying, can you connect into this? Can you connect into that? And then we're going and exploring that. And as long as there's a way to connect in, as long as that platform's got APIs or whatever, we can connect them in and bring them into our ecosystem if you like.

SPEAKER_00 17:03

Brilliant. So security and and governance are still major questions around AI for enterprise buyers, particularly large enterprise, lots of concerns about hallucinations and trust and all the usual important topics. You know, how do you help customers feel confident in, you know, the answers they're getting, but also uh build trust without slowing all the good innovation you're you're doing down?

Roadmap Fundraise And Adoption Barriers

SPEAKER_01 17:31

Yeah, so I mean, it's obviously something that comes up in our game quite a lot, is the security questions. And I mean, most calls we will, most conversations will get that, typically from CFOs and CTOs, they'll want to know how secure the data is. So, in terms of everybody's data that we store in our platform, they effectively are all ring-fenced databases. So every single connection we create is in a separate database, so everything's kept ring-fenced in that way. All our platform as well. So we're an ISO 2700001 certified business. Um, that means that we've got all how we handle the data, how we secure that data, that's all externally audited. So that's all covered off from that perspective. And in terms of what actually goes into these AI tools, whether it be Gemini or whatever, what we do is effectively we never really send in all the data at any point in time. It's the AI is effectively delving into our platform and getting out what it needs at the time. And what we're always recommending people do is you know, if the as we're always pushing people towards using, you know, business versions of claw, business versions of chat GP. Don't use the free versions, use the business versions because by default you are not sharing any of that, any of that data with these LLMs for them to train them. And we don't do any, we don't pass anything to the LLMs for training either on our side, we're not doing any of that. So effectively, it's you know how you want to how they want to manage that security with you know Claude or Chat GPT is you know, it can be controlled themselves so they can you know be comfortable in the fact that they've you know made sure that nothing is going to these LLMs for training.

SPEAKER_00 19:14

Brilliant. Um, I'm sure you have a very aggressive, interesting product roadmap and feature roadmap that we can't get in in uh in detail today. But um, can you give a peek into the future in terms of things you're working on or or planning over the next uh 12 months?

SPEAKER_01 19:32

Yeah, so we're sort of typically growing out sort of our connectors at the minute, anywhere between one and three new connectors launching sort of every week, if you like. So that's by connector, I mean that'd be a new platform that we'd integrate within. So that's ever evolving, if you like. You know, we're evolving our semantic layer, if you like, the one that's you know making it easier and easier for the AI to understand what's going on with the data as well. That's evolving all the world. But what we're also looking to do um in the next quarter, we're gonna do a fundraise as well. So we're gonna go out for investment and do a raise there to bring that up because what we want to do is effectively speed up that connector roadmap um and you know, you know, put more money into build that semantic layer, if you like.

SPEAKER_00 20:19

Brilliant. Well, great opportunity for any investors who might be watching or listening. You're you're also busting a lot of myths about AI and how difficult or easy it is to deploy. Um What are some of the biggest roadblocks do you think to really driving enterprise AI adoption beyond the uh you know the flashy news headlines? Uh, you're you're really you have quite a practical approach. What needs to happen to to really make this industry uh achieve its potential?

Free Trial Onboarding And Wrap Up

SPEAKER_01 20:49

Yeah, I mean it's it's it's strange at the minute because when you talk within tech communities, everybody's adopting it, everybody is using it, everybody is jumping on board with it. When you go outside of that community, there's a lot more fear. Fear is one thing. Um, there are certain industries that are absolutely terrified of what's going on with the data, which is obviously why secure is a big thing that always could we always get asked about where's my data going, you know, is it secure and all the rest of it? So that's so governance has got to be something that sort of like is is really sort of brought on as well. People are scared of hallucinations as well, because there's a lot of you know, there's a lot of track, you know, a lot of traffic out there if you look on you know platforms like LinkedIn and news, you know, it's it's come up with this number and all the rest of it. Um so there's there's fears around that as well. Um, the way we sort of alleviate those sort of fears in our platform is the fact that we've got this, we're only talking to data that you have, real data. It's not going out to the web to go and try and guess this stuff, it's talking to data. And we're obviously evolving the layer all the time, this semantic layer, if you like, all the time, which tells it how you're supposed to join this and and all the rest of it. Um but yeah, it's uh it it is a it is a strange world, but it's and it's at the minute things are evolving quickly, very, very quickly. It's it's bizarre. I mean, one day something else is happening, you know, the next day something else something else has come out, and it's it it's so fast. I think from an enterprise perspective as well, there's a lot of um there's a lot of there's a lot of hurdles that have to be sort of jumped, if you like, in large scale enterprises. I think SMEs are quicker to adapt because they don't have all the all the tape. They've got you know the red. Tape they've got to get through. But I think in those enterprises there's a lot more red tape. Um certain industries, whether it you know, you know, energy industries, uh, when you're talking like if you're talking about like nuclear power or whatever, they're just not going near it at all. I mean, they can't some people can't even connect Wi-Fi into the so accident sent out to AI is gonna be is gonna be a sort of a problem for those guys at the minute.

SPEAKER_00 23:05

Yeah, lots of uh uh roadblocks that need to be navigated. So if you're uh you know, maybe a mid-sized enterprise looking to dive in, how do they get started? Uh how do they, you know, go to uh tugarap.com and you know reach out, maybe I see you have book a demo, but even get started for free. What what does that mean exactly?

SPEAKER_01 23:27

So, yeah, so basically you can literally get started for free. Um, our platform comes with a 10-day free trial anyway. That gives you plenty of time to get the data out and integrate it. Um, but yeah, you can literally just go through, click up sign up for free. Nobody has to put any card details, there's no tie-ins or anything like that. Um, and it'll basically walk you through. So we've got an onboarding process, if you like, that just takes you through step by step how to connect everything up. And then, yeah, once your data's there, you can start to produce reports just like you sort of see in scrolling up and down the screen there from the sort of claude example there, if you like.

SPEAKER_00 24:03

Brilliant. Well, I can't wait to touch base in 12 months and hear more incredible stories and anecdotes from customers and see where you are with the platform. Thanks so much for joining and sharing the vision.

SPEAKER_01 24:15

Yeah, no problem at all. It's been great, Evan.

SPEAKER_00 24:18

Uh yeah, thanks everyone uh as well for listening, watching, sharing the episode, and uh check out our TV show, techimpact.tv, now in business uh channels like Fox and Bloomberg Television. Thanks, everyone. Thanks, Greg.

SPEAKER_01 24:32

Thanks, Evan.

SPEAKER_00 24:33

Cheers, everyone. Bye bye.