#133: Mike Fitzgerald, Serial entrepreneur and Founder of Senttr, on Digital Workers and the Future of Enterprise AI

0:00If we think that our employee satisfaction, let's say our human satisfaction, is going to get better by having people type things into systems, I think we have a problem as leaders. I think we do.

0:17Today's guest is Mike Fitzgerald, founder of Center. Mike has built his career at the intersection of technology, business and innovation, eventually pursuing entrepreneurial ventures across software, reselling, professional services, before founding Center, which is a digital worker platform for operational teams with a motif, Gruntworks Sorted, which I think perfectly sums up what they do. In recent years, the conversation around AI has largely focused on co-pilots, assistants, and content generation. But what happens when AI moves beyond simply helping people and starts carrying out the work on their behalf?

0:55That's exactly what we're going to be exploring today. We'll discuss the rise of digital workers, what AI means for enterprise software, licensing, governance, cost management, and whether businesses are prepared to manage an estate of AI agents and how all of this could reshape the way organizations operate over the next decade. Mike also shares his views on the future of work, how AI is likely to reshape professional services, the skills tomorrow's leaders will need, and why organizations that gain the greatest advantage won't necessarily be the ones with the most AI technology in their estate or the most powerful AI, but those that integrate it into their operations most effectively with the right governance.

1:39So this episode is a real eye-opener and really is a taste of things to come. This is Mike Fitzgerald.

1:53Mike Fitzgerald, lovely to meet you. Thank you so much for agreeing to do this. How are you today in sunny London? Doing really well, thanks, Gareth. Thanks for inviting me. Very happy to be here. Brilliant. Well, look, we've been busting to get someone on who's trying to scale an AI company at the moment for various reasons. There's so much going on. I did an interview recently with a couple of interviews with executives at the GCP AI Summit in London. Some of the stuff you guys are doing looks really interesting in that space. So I can't wait to dive in. But before we get going, I wanted to start with a question we always ask our guests to kick off, which is what does good leadership mean to you, Mike?

2:33It's an interesting one, actually. I've been very lucky throughout my career to have good leaders, let's say, I think. I've been blessed with being around what I think has been very good leaders to maybe help me shape what I've become today and what my leadership style is. I think you're always trying to improve that. But for me, I think number one is kind of guidance first, guidance and coaching first. because we have this ethos and I have this ethos that I think if you think about, you want to bring someone into your organization that can help you do something that you can't. So if you bring in a very specialist skill set or someone that's an expert somewhere, then you really shouldn't be telling them what to do in their field, right?

3:17So guiding them to be successful and building the right environment for them to use their skills, their knowledge and their experience best. and then I think the coaching piece is helping them take that to the next level so how can they be impactful and and then building on that is building the right environment and the right organization culture the right performance culture for people to be successful so for me what throughout my career I always thought I was good at something specifically and maybe sometimes it was technical and now it becomes a little bit more of a kind of broad management subject people have always been able to provide me with the right environment and coaching and definitely communication and feedback that that allows you to operate at your best skill set let's say and I think I think for me I was very lucky to be around those people I try and replicate those things and learn from those things but yeah I think if it if you really want to be a great leader I'm not proclaiming that I am yet I think you've also got to be learning and listening and that that's the the style on the side of the good communication as well yeah for sure no that's a great answer I think you covered a lot there I mean have you ever had a mentor or anything like that Mike Yeah, early in my career, I had a couple of quite senior mentors that really helped me think a bit differently about what I might do with my career.

4:31And then as I've gone on throughout the career, I've actually had some incredible management coaches. I'm going to say that they weren't my choice. They were always maybe implemented as good guidance and good coaching from my good leaders that you could do with some help in these areas. And that's been very good. I had two very, very good leadership coaches in the US over the last 10 years that have been really, I think, taken my leadership style and capabilities to a level I didn't think I would ever get to. Could you give us a little bit of a high-level overview of your background, why you sort of went into the technology software world to start with, and a little bit about the sort of journey you've been on ever since, right up until forming Centre back in, is it, early 2023?

5:17A bit of personal background probably helps. So from an Irish family, originally grew up in Ireland. My mum was from the north of England, from Newcastle. so we ended up moving up north and to north of the uk where that's where i went to school it's where i learned to play football it's where i have my largest group of friends i'd say the closest group of friends and and then i also did a lot of boxing and i always thought that i would either end up in one of those two things did a lot of kind of amateur boxing really enjoyed it oh wow i love boxing by the way yeah i but yeah that was that was really where i wanted to be and I enjoyed it so much.

5:53Although I had part-time jobs, I went to a local sixth form college and I'm not, the academic side, it never really drove me and I really struggled to see success in it. But I think boxing and training in football, it taught me discipline and it taught me hard work. And actually I put down, when we talked about leadership coaching, I put down my original boxing trainer as probably one of the best leadership coaches I've had that taught me hard work and discipline and controlling your emotions and things like that as you grow up and then that didn't work out for for many reasons and so i went back to night college and i i did a placement time locally in my local town doing what was called an mvq and installing supporting it systems at the time i learned to code so i had two jobs i'd one during the day at a local college supporting it i'd go to college one day a week and then on a night time i'd do some it work basically i'd take whatever contracts i could get and very quickly then i I started to build a career in solutions architecture and software, and I built small software packages for small companies, and that became bigger.

6:58I worked in local governments and councils. And the opportunities would start when you were looking for contracts. Contracts was the way to go. It would pay a little bit better, and you could get a lot of experience in a short period of time. Absolutely. So I ended up taking contracts a little further afield. So I'd drive an hour maybe from where I was in Durham. I could drive then to Leeds an hour and a half, and then eventually I'd be driving into the Midlands. and a good contract came up in London. I always wanted to kind of try to live the kind of corporate life and the better jobs and the better opportunities.

7:29So I moved to the city in my early 20s and I met a gentleman, a good friend of mine still, Ian, and his friend Tracy. We decided let's go and build a tech company. They were very, very experienced in healthcare and I'd been helping them with some projects in healthcare. And we decided to start our first tech company, Innovate. I think I was in my mid-20s at the time. And we built a very good company. We were kind of nationally renowned. We won some international awards. First time I'd traveled to the US, I had to fly to the US for a few award ceremonies and conferences. And I was really enjoying being part of that.

8:02But I was still immature and I didn't really understand business or management culture. I didn't really understand how to scale an organization, but we did very well. And I left there after four years and I went off and I worked with a friend of mine in the industry who was the CEO of a very large managed service and tech company based out of Switzerland. And there were a few thousand employees. I'd never worked anywhere that kind of size or scale or experience. And so they wanted someone that could help them move from kind of reselling to services and solutions. And so very quickly, I had a corporate job leading large teams, and I had to grow up very quickly.

8:37But after being somewhere for seven years, the kind of entrepreneurial itch started to come to me again. And as much as I love a challenge and building tech and building things, at that job, there was always something to do, always something to build. I'd been building a marketplace there and I was very interested in AI machine learning. I saw how it could make a big difference to us, but I didn't think I could take it in a big company at the speed that I needed to. So I took some time off. I had a little bit of a rest for six to eight months. And then I actually met up with some ex-colleagues from that company.

9:11And we said, we've always wanted to do something together. maybe we should so just late in 2023 we decided we'll formulate the idea that became center over time we initially got it wrong and that's what most people will say oh we pivoted from one idea to another no we got it wrong and then we edited it till we got it right and then about 18 months ago we commercially launched what was originally called ai center now called center and we can talk a little bit about why that changed and and now we've been successfully scaling center for 18 months and we've got we're now at a point where we've got customers in six countries where we've got a growing team based out of london in the uk i believe we are the leader in what we do and now what we call a digital worker segment so it's been been an excellent time of building and joining friends and doing something fun but one of the one of the best things about doing things with center is probably the stuff that we get to do with customers as well so yeah a bit of a whistle stop tour of my career but really i've spent most of my time in kind of solutions architecture through the product management and then leadership in technology all the time.

10:12Fantastic. Well, that's a great intro. So what I wanted to ask you about is you've worked obviously across vendors, resellers, services, businesses. You've been a CTO, entrepreneur. Looking back, which is what you've achieved up until now, what experiences or companies or industries that you've worked in have had the biggest influence on how you are building Centre today? Yeah, it's actually, it's something that we try and revisit a lot. like I have this concept that actually a leadership coach talked to me about. My team get bored of me saying this, but I call it revert to manual or return to manual.

10:46So we usually write down the principles of the things we're trying to achieve and then always go back to that and say, are we achieving them? Are we achieving them? So one of the things that we set out to do is try and solve some of the problems that we had when we were trying to use AI to solve business problems. And the same principles apply. So the first one is in most technology solutions that are launched to industry today, the customer is not the first thing that everyone thinks about. It's usually the return on investment, how much money can we make, how much impact can it have, but people are not thinking about the ease of deployment, how easy it is to understand the cost of consumption.

11:23And so fundamentally for us, customer first is a very, very important thing. And so even down to our pricing model, we want it to be predictable, simple, easy to manage from a customer's perspective. And I think if I'm going back to your question, the biggest impact for me and still continues to be the biggest impact to me is get in front of the customer, spend time with the customer, understand the business and then create the solution or then edit the solution of fit. And there's too, too much of our industry spends time on probably over complicating solutions that creates a false economy for the customer.

12:00Right. Yeah. I understand exactly what you mean. We've got this whole economy in our business, in our business, business segment of licensing, of software licensing. And software licensing has never been simple for the customer. It's one of the most difficult things for customers to understand. And there's almost this economy, an industry we created just to understand software licensing. Well, actually, all the customer wants to do is consume some technology that makes a business impact for themselves and potentially their customers and their people, right? And that's really the thing that for me we are responsible for in our industry is solving that conundrum let's get in front of the customer let's make it easy for them to consume things let's make it easy for them to have an impact and that's the value step that i think that you've got to be driving yeah absolutely can you give me an example in terms of software licensing then of how you would help businesses manage their estate from a licensed commercial licensing standpoint yeah i think when go back go to back to i was like i said i was at a very large service provider reseller and one of our services was we will look at your estate and help you keep control of that spend and make sure you're not overspending, underspending, et cetera.

13:08When we look at that, when I was designing the pricing model for the platform, I say I actually, when we were as a team, one of the things we said is we want it to fundamentally be self-sufficient as a model. So let's take common AI tools that exist out there and we don't have to list them, but we're all using at least one, I think now, especially in the industry that we're in. You basically pay a fee to use the platform. At some point, that platform will probably send you a message in the app, and it will say you've run out of credits. So you've now got to build some credits. And then you start testing those type of things, and you say, okay, this is great, but I'm spending$100 a month.

13:48And now I've been spending$160,$170 a month on credits to get this thing to work. What I want to do is I want to roll this out to my business now because I can see it as an impact for me. I want to take it, I want to roll it out to 100 users from small to medium-sized business. What you then need to do is work out, is it$260 per user? What's the variable low rate cost, high rate cost? And then, okay, when I put that out there, what else do I need? So I need some infrastructure for it to sit on. I need some security wrap around it. I need a management cost to make sure it's rolled out properly.

14:17I need to make sure that people are using it. Otherwise I turn it off. And then there's a whole set of services in the industry defined from very good service providers. I did work at one that can help you manage that. But that economy is being created by technology companies creating a licensing and business model that actually, for me, shouldn't exist for the customer. Because if you can't predict how much something costs, I'm pretty sure you probably won't consume it in most cases. And in other industries, mature industries, let's take the motor industry to keep it super simple. If you're going to go and buy a car, there's 1 ,000 calculators out there in the car while the hours and similar, you can work out how much you want to spend per month, how many miles per gallon it does, or the electric range, and how much the consumption of the electric range is.

15:02You can work out how much that is plus the fixed and variable cost as an estimate. You can't even really do that sensibly in AI technologies today. So that was one of the things I wanted to solve in the building center, was make it predictable, make it easy, and then somebody can build a very quick business case on top of it. Will it have an impact? Will it have an impact, basically? looking back across your career you've obviously you said obviously being led by the customer which is absolutely right you generally seem to build businesses around emerging technologies or around problems that you felt a lot being solved properly is that is that generally what you've done with all of the businesses you've tried to create is that is that fair to say mike yeah i think if you i think it's exactly right carth to answer your question and just give some context i think I kind of grew up into this industry as a solutions architect and so you're a consultant and your job is here's my problem or help me find my problem and then help me find a solution so being a kind of solution oriented mindset in by design and also by solution sales right because that's how you build a business on that that's fundamentally what's happened if I saw something that maybe I couldn't find an off-the-shelf solution for I think we could build a good entrepreneurial business around that and that that's really what we what we did and and they were probably problems i thought i just had uniquely in my own vacuum in my last business or in the business that i represented and now in the future what i've realized is spending more time with customers these are common repeatable everyday problems that most businesses have as well absolutely yeah so coming on to center i noticed you describe it as a digital worker platform rather than an AI agent platform what is the distinction then can you maybe just talk us through the problem you're trying to solve and what makes you guys your platform unique and all value to business owners or business leaders if we go back to that customer problem or conundrum and then take it forward to how a let's say AI platforms and agentic platforms and what we're seeing now is a kind of probably a generation of agentic platform which is agentic workflow platforms and then what I think is the kind of digital worker platform so there's this concept of you use AI to do tasks for you you drive it it does tasks in a lot of cases that's what we've been doing with most of the AI the kind of common LLM based or model based AI clients and they're very good I use them every day I'm sure you use them also if you if you need to create something quickly or moderate something quickly take some notes look at transcripts they're fantastic and then there is versions of those now that allow you to do this kind of agentic workflow, right?

17:44So you can basically build out common tasks and each agent might take care of that task and he could join them together into workflow to chain those tasks into kind of a common business workflow. What we see as a, let's say, a development beyond that and as a future beyond that is digital workers. And digital workers are a combination of, let's say, AI agents of workflows with some common thinking around a particular business problem and some learning and training around a business problem. I'll take a simple one just to explain it. So if I have an invoice management or an invoice processing digital worker, I want to be able to trust it as a member of my team to process my invoices.

18:24It knows what invoices look like. It knows how to process them in my systems. It knows where to find them. And it knows the commonality and the business rules around how I process invoices. So the idea is very quickly you can use an invoice management digital worker to process your invoices. And what the idea of the digital worker is, take out common repeatable kind of grunt work or back office core business type tasks that most people, I'm not going to say humans because I hate that we use this new term inside of our industry, but people do not like data entry and execution in systems. And we have multiple systems that are disconnected.

19:03And what we say in the way that we'll create APIs, workflows, other things, and fundamentally they're unreliable because they're not specifically designed for that task. A digital worker that's designed for invoice management that has seen one million invoices tends to know what the invoice structure looks like, even if it's written on the back of a napkin, right? And so fundamentally, what we've built is a platform that has pre-built digital workers. so you can say an invoice management is one of them i use that as an example and then you can also build your own so you can basically add join the dots very quickly build your own worker that that covers your task let's say gareth you want to you want to create a lead management one so every time a lead comes to your system you want to take it put it in crm and assign it to someone and then you want to report on what's happening with that lead you can build your own lead management worker and the idea is you can do it without being a technologist so in our platform you don't need to be tech savvy.

19:55You do need to understand your own systems. So if you're a salesperson, you should know how to log into CRM. If you're a finance person, you should know how to log into your finance system or your ERP. But fundamentally, what's happening, I believe, in this industry with digital workers is they will become actual parts of teams driven by the team, the person, the human worker. They will be controlled by the human worker, but they will take over those grunt work tasks and allow us to scale businesses in a much more linear fashion, in a much more simple fashion, in a reliable way, fundamentally.

20:28This episode was brought to you by Be Digital. Be Digital support leadership teams to optimize cost and get more out of technology investments. Be Digital and the team have unrivaled expertise with technology license management and data remediation and are therefore perfectly positioned to help prepare organizations for AI technology capability. And on the last point, BDigital have developed a cutting edge AI readiness assessment, which provides tech leaders with a platform they need to make well-informed decisions about AI adoption strategy. Go to BDigitalUK.com to find out more and get in touch.

21:16I've tried to create agentic workflows i've created one which posts for me on twitter every now and again and that wasn't too bad but i've tried to create them for other purposes and it is quite painful i used make.com and i've looked at n8n is it that one as well i'm not technical so i just use a youtube walkthrough to create it but it is still not a work involved at the moment it's it is not many platforms that enable you to create agentic workflows through natural language prompts is that sort of part obviously i know this is for the purpose of operational work and finance and back office and stuff like that but is is that what you essentially the problem you're trying to solve you enable people non-technical people businesses to create like workflows that are going to deliver grunt work i noticed your slogan grunt work sorted which i thought was great by the way That appeals to me a lot.

22:11Yes. So I'll go back to kind of your explanation. So very powerful tools like nannmake.com, they're actually excellent, and they're excellent at specific things. If you're a technologist, maybe like I have some technology skills. I'm not as deep technically as I used to be, and I will definitely run out of skills when it comes to those two platforms. There'll be parts I would have to ask some of my friends, technical friends or technical colleagues to be able to help me with. And that's the idea is if I'm running accounts payable, accounts receivable at a medium-sized company in the uk it's likely i've got so many invoices on my desk or inbox or my team has that i do not have time to go youtube and it's not that it's not the initial setup that's the problem it's the maintenance and reliability of that because you're building your own models and you're pointing at your own things you're not sure about data privacy you can't cover the security and in most cases your it team won't let you do it anyway right so if i can go to a pre-built platform and i know this thing is reliable and it's built specifically for the purpose.

23:08I can consume it. I know exactly how much it's going to cost to do the work. I can go do that and set it up. And that's exactly what we set Center up to do. And we have two ways you can do it. So you can basically go on our platform and choose from a template. Let's say we choose the invoice management template. It'll just say, Gareth, where do your invoices come in? They come in email. Okay, and what's your ERP system or finance system? I'm going to choose QuickBooks. Okay, log in at QuickBooks. Okay, we're not going to process some invoices for you. When they're processed, where do you want them to go?

23:34Send me a Teams message, please. Just let me know to approve them. and I'll approve them and then put them in the system, please. And you'll get four prompts and you're ready to go. Or if you don't know what you're looking for, we're bringing, it's actually coming to the platform in the next few weeks. You can actually type in and say, I need something to process my invoices, please. And it will take you through that workflow in the same way. And the idea is we want to solve the small to medium business conundrum of scalability. That's something I've been challenged with myself over the years.

24:04So building small and medium businesses is very, very hard because of that technology conundrum. You're trying to learn everything at the same time. You're trying to learn bookkeeping. You're trying to keep your eyes on operating your business. You're trying to learn technology. I'm not so sure the majority of us have enough time to go and watch eight hours of NHN videos and sign up and all the things to go around the arts. Absolutely, yeah. Yeah, no, of course. And that is very appealing, what you described there. And I think that is definitely the future, isn't it? In terms of that human versus digital worker conversation, do you ever have to say to customers that a human should stay firmly in control of that?

24:40Or when do you think it makes sense to give work to a digital worker and when should a human stay in control? Yeah, that's actually a good point. So we're designing, I talked a little bit about, we designed the platform that it's used by the user that usually controls that function. So the finance process will be very familiar for invoice management to a finance person, that they would be very capable of setting up the finance worker. I would say maybe the lead gen agent or the lead gen worker would be much harder for someone in finance to set up because they're not familiar with the systems and the processes.

25:13By design, the platform and the workers are built to be able to be understood by the owner, particularly in that part of the business. And then if you look kind of on top of that is we've even changed some of our vernacular in the platform, right? So most of the platforms in the industry call the kind of human process, they call it human in the loop. We call it human and purival or human in control, right? So it starts with the human, ends with the human fundamentally. The idea is that you need to be in control of these platforms if the digital worker works for you. The idea is it helps you scale your team, not it replaces your team for time.

25:49because we're not, and that's our focus of Grunt work is, it's impactful of the things that we don't like doing. I haven't met anyone yet, and I'm sure there's someone out there that does, but I haven't met anyone that likes typing in invoices into a ERP system. It's not a fun task. It's prone to error. It's repeatable. And in most cases, when a lot of people that we meet from finance see that process, they cannot wait to automate it, even if that's taking a significant amount of their role away so they can go focus on something else. Exactly. And that's definitely the situation with myself. And yeah, that makes total sense.

26:27So I wanted to talk to you about governance and managing AI at scale then. So this is obviously deploying hundreds, perhaps thousands of AI agents over the coming years. What new governance challenges do you think companies are going to face and what should they do about that in terms of making sure that these agents are doing things in the interest of the company? There's probably a range of things that are happening in our industry right now. Right at the top level, including ourselves, there's the software vendors that are now having to look at new legislation and governance associated with deploying AI technologies and their ability to be able to do things with company data, for example, personally identifiable data, as well as just make a lot of sensible business decisions.

27:09So this is something that we wanted to do as part of being customer focused was we wanted to be very particular about making sure our platform was private. We have things like data encryption in place and it's controlled by the administrator or the human that's responsible for that worker. So we saw it initially as building workers that would live in your team. You would be a leader. And we talked about leadership earlier. You're going to lead the digital worker. You're going to give it governance and guidance. You can control. and we built in things like auditability and accountability into the platform to be able to do that.

27:42What's going to happen from a customer perspective is making sure they choose a platform where those controls are in place, but also add some other levels of controls around the rest of the organization. So we're all using our, as I've said about this, and we all talk about this. We used to have this IT conundrum 10 or 15 years ago. We talked about shadow IT that just blew up. Shadow IT is the new IT, in my opinion, because if you stop your users from being productive using tools, let's say like a cloud or an open AI, you might actually slow them down today. But you still have to be responsible for what company data is going in outside my organization.

28:16Am I users well-trained or at least understand the threats of using AI for certain things? Can they potentially connect it to systems and data that are important to our organization that they shouldn't be connected to? So IT departments have another layer of governance and another layer of controls that they now need to implement, but without throttling the productivity of our end users. We must say that I've always been talking about this, about you wouldn't implement something on your own laptop that would stop you from working. So if you work in IT, you should not implement that on any of your users' laptop that should stop them from working or being productive either.