AI is increasingly moving beyond tools that simply help people work faster.

The next step could be AI-powered digital workers capable of taking responsibility for entire business processes.

In this episode of ThinkData, Alex Hutchings is joined by Dries De Coster, Founder & CEO of meet DWIGHT, a company building AI-powered digital workers to automate middle and back-office processes.

After around 20 years in the HCM industry, including leadership roles at SAP and The Access Group, Dries launched meet DWIGHT just over three years ago.

The conversation explores why he believed the timing was right for a new approach to automation, how the company identified genuine product-market fit, and what it takes to build and position an AI company in one of the noisiest technology markets we have seen.

In this episode

• Why Dries left The Access Group to build meet DWIGHT
• The problem meet DWIGHT is trying to solve with AI-powered digital workers
• Why the timing was right for a new generation of enterprise automation
• How early-stage AI companies identify genuine product-market fit
• How to position an AI company when virtually everyone is talking about AI
• The biggest GTM lessons from scaling meet DWIGHT
• Hiring lessons from building an early-stage AI company
• What Dries looks for when hiring into a startup
• What success looks like for meet DWIGHT over the next 12–24 months
• How the digital workforce could evolve as AI technology matures

Welcome to the Think Data podcast brought to you in partnership with Mydataworks. If you want to stay up to date with the latest breakthroughs and trends in the world of data and artificial intelligence, and if you're curious about some of the strategies that companies and founders use to launch data and AI products, then you're in the right place. Our aim is to bring together a diverse lineup of fantastic guests from the founders, through to accomplished leaders and product owners at some of the most fascinating data and AI companies worldwide. They will each offer you their own unique insight into what it takes to launch and scale a great data business. Thanks for tuning in, and I hope you enjoy the episode. Welcome back to the Think Data podcast, and today I'm joined by Dries de Costa. He is the founder and CEO of Meet Dwight. They are a company that are building AI-powered digital workers to automate middle and back office processes, specifically within the recruitment and staffing space. Dries has spent around 20 years in the HCM space. He held leadership roles at the like of SAP and the Access Group, and there's no better person, in my opinion, to get a really honest take on where AI and automation and this digital workforce are heading. Dries, it's really good to have you on today. As I mentioned in my intro, you've got a really kind of interesting background with respect to kind of this staffing space and kind of HR as a service, software as a service. I know you spent six years at Access, went on a crazy journey there, I think in terms of kind of the point you joined, the investment, then obviously growing that business to kind of a $9, $10 billion organization. After kind of six years, what made you leave ultimately and set up Meet Dwight? Because obviously I'm a recruiter by trade and I think we talked off air about kind of all those kind of backer processes, systems. It does become a bit painful. But for people who don't know, who are Meet Dwight and what made you ultimately join and set the company up?

Yeah, so I mean, very briefly to touch on the access journey, I was brought on board there to help start a brand new division as it happened into hospital. hospitality software did also then go into the recruitment side but um we did five acquisitions in nine months to start that division and i really enjoyed the kind of forming storming figuring out you know who are we in this market and then executing on the go-to-market piece so obviously i mentioned there the growth story was insane um when i joined we were 225 million market cap took that up to 9.24 billion at the end of obviously in the space of about five and a half year so I would dare say yeah absolutely insane but i really enjoy that kind of earlier part of the journey and had always really wanted to do my own thing just never really had the right idea or wasn't the right time you know kids were too young yada yada different reasons but i also identified that you know there's a huge gap particularly in middle and back office for recruitment and staffing firms because we operate uk us similar to yourselves and yeah lots of disparate systems that don't talk to each other and then particularly for kind of higher volume recruitment and what that leads to when you're talking about onboarding compliance credentialing going in and out of ems systems payroll finance all those kind of processes typically people were throwing people at it because again they weren't talking to each other so um also a big focus certainly historically in the sector on driving top line revenue and make the i think in part because a lot of great recruiters were sales people set up their own business and really similar to myself by the way And naturally, you know, my forte is not on the operational side. That part of the business often gets neglected, but it's a massive sap on profitability. And that's really why we started Meet Dwight. So we build digital workers. Hence, Dwight stands for digital worker interface for generic human tasks.

Okay, interesting.

Yeah, automate a lot of that. Works 24-7, never goes on holiday. It doesn't give you any HR headaches. And, you know, has all those kind of benefits of particularly, again, in high volume kind of work when you talk about onboarding, for example. Lots of workers upload their documents when their shifts are finished. Dwight processes it overnight. They're ready to go on the payroll the very next morning. That's where we started is really to fill that gap, help recruitment staffing firms increase their profitability or at least not have to hire as they grow by putting more people into that kind of part of the business. And then more recently as well, we've released a Dwight Studio, which essentially enables our... um our clients to go and build their own digital workers or ai agents whatever you want to call them across all of these different systems where again typically they would would have had to throw people out

Yeah, it's really interesting. I was going to go back to my mind, I think about when I was at a staffing firm, we grew from like 25 heads up to time leaving what 300. I remember we had like literally in operations, we had one person doing the payroll, one person in the timesheets, one person doing, you know, back office credentialing, background checks. So in terms of obviously for anyone working in a kind of middle and back office now within recruitment or staffing, their first thought is, oh, I'm out of a job here. You know, a lot of people I speak to in the kind of AI agent space on this podcast is there to enable people to free them up to do better work. Is that the kind of line that you tow as well? Is there ultimately going to be a repositioning of that back office function recruitment companies as a whole because of Dwight?

I think it's probably a bit of both. So, you know, we started the company just three years ago, basically.

Yeah.

For the first kind of year, spent building stuff. And then the last two years we started billing. And the... growth rate has been phenomenal but certainly in the early days we were really kind of creating the sector in the space no one was talking about digital workers so much also why we identified the sector right because it was unpenetrated by some of these more advanced technologies in part because it was completely unaffordable so that's the other thing is we made it affordable for recruitment companies um yeah but initially it was very much kind of having to explain get people comfortable with it etc we're now in a totally different um you know Thank you. It's funny to say, right, because it's only been two years, but we're already in a totally different market now where people are now a lot more comfortable with what it can do. Certainly wider usage wise, you know, I use Claude a heck of a lot day to day, but there's still a bit of a gap oftentimes between we want to execute an AI because we know it has benefits, but we're not quite sure exactly how that works for a company like ours. And that's really the role that we often bridge. Yes, there's the affordability side. But there's also very practical use cases of, you know, if you're a recruitment business that looks like this, you're probably struggling with X, Y, Z. And hey, we've already built these agents to go and solve that for you.

Yeah, and I really like that. I know a few kind of founders of recruitment companies who are very all in on Claude. They're very public about their kind of automation of all their various workflows. And that's great because they're a sub 10 person business and they kind of got there. kind of fingers in most of the pies so where where would you say meet dwight's flourishes in terms of that size because i know it's from what i'm seeing online and reading about you guys obviously if you look at a contractor heavy business obviously there's a huge amount of laborious data entry and workflows needed but in in your examples and some of the use cases and maybe some of the that

customer feedback where where do you really where do you really sit yeah i mean typically it is that kind of higher volume higher complexity right because you you'll know So, you know, if you're running 10... you'll have different rates agree with different clients all those kind of things and again you we've we talked to recruitment agencies that have got maybe 30 different vms portals that they that they have to interface with for various clients as an example so typically it's that space but we've got smaller customers by by turnover by number of heads that they've got in their business where you know in one example they were going to go and hire a credit controller and actually decided could dwight do this the answer was yes and then you end up with someone that costs all. someone dwight um you know something that costs a fraction of a full-time person's costs but doing that kind of 24 7 work and they've not had to hire that person they've saved a fortune and have now got a more scalable engine because dwight's doing you know things like doing the invoices doing the chasing doing reconciliation all those kind of things for them so typically it is that kind of high volume but again also with the rise now of the dwight studio where people can go and build their own automations. opened up a lot of smaller bits that people can go and automate so again if you're if you're not one of those kind of big or upper mid-market kind of organizations you often have people that do lots of little things and that's really what video is aimed for is you can still go and automate that whereas typically before where we were building bots for people we would say it needs to be replacing one full-time person or more or 40 hours per week or more of a task to really make it worth while so it's you know we we kind of feel like we now go across the gamut with people will often now buy the studio because they have all the little things they want to automate or it's project-based because they've got a big contract that they're onboarding but then it might still be like a really chunky process that they still want us to build on the studio for them and we can do both yeah

no no i like that and i think it's probably In this world of kind of, I know we talked off air about this, kind of there's so many new shiny tools, AI agents are kind of the biggest buzzword in the world right now. And actually where you focus is a niche, which I'm also, I mean, I've seen firsthand the challenges with in a bullish market. You know, you just hire all these ops folk and you don't necessarily look at the bottom line. But when obviously we've seen the market post, you know, COVID, we've seen market in the last few years, it's been really tricky. But for you, obviously, you're kind of giving that very clear, this is the cost base. And I'm guessing with Dwight itself, is it aligned to different parts of that recruitment workflow? Or have you got different Dwights based on, say, the credit controller one? Or is it learning from your business and picking up all parts?

Yeah, so a couple of things to come back to on there. So one is that kind of linear alignment of revenues going up, therefore, back office staff and cost goes up and we help break. Yeah. Right. So. I think people in the post-COVID world, they know they have to do that. Like their competition are doing it. They're competitive. You have to kind of do that. In terms of does Dwight kind of self-learn? Not really, but can do. What I mean by that is within the platform. So the challenge with a lot of AI functionality out there at the moment is, one, it gets pretty expensive. And I think people have now started to realize that these token costs, they add up pretty quickly. So instances where the token cost is costing more than the people that were doing the work. So technology is out there. It can do some things. It gets quite expensive and it can't do the really kind of advanced, complicated stuff. And that's really where we come in. So, you know, to your question of are there different Dwights, there are different Dwights and different automations that do different things. But by and large, we combine agentic AI just like a Claude has, right? We have that within our platform, but you don't have to. consume those kind of expensive tokens. A Gentic AI always needs good connectivity into systems, so an API or something akin to that to work over systems. The reality is when you look at a recruitment business, they still have other platforms that don't have API connectivity. And the beauty of what we have is we've got deterministic kind of automation. It's called robotic process automation, also in our engine. They can combine the two. But with that piece, what we can do is we can go into legacy software, we can go into VMSs. by using, for example, logging credentials. We can go into government platforms to do background checks and whatever else needs doing. So we are able to combine the two, and then we have a very strong governance layer on top of that. So when you talk about things like payroll, invoicing, getting timesheets out of VMS systems, doing onboarding and compliance, those kind of tasks, you simply cannot afford for bots to go rogue or hallucinate. You can go into chat GPT today, ask a question. Ask the same question tomorrow, you'll get two different answers. So what we are, and I think the whole market is now moving back towards this. And when I talk to big technology firms, they are now also coming to that realization is use automation as much as possible. Have that deterministic because then the bots don't go rogue. And as I say, you can connect pretty much anything, but then infuse AI where it makes sense to, again, to try and keep the copy. And right. So a very practical example of that would be. We, for example, use intelligent document processing, which is a form of AI where essentially we will read things like people's timesheets. And to do that, you have to interpret the data that's being pulled from, say, the PDF and what you want to go and do with it. And then it hands over to the deterministic, if this, then that kind of workflow process. So that's really the best practice in the market. And we were fortunate enough through our advisors that we were well ahead of that from a product perspective. We bring the two pieces together.

Yeah, and that kind of leads on to the next point about kind of that product market fit, because there's so many companies who have raised a huge amount of money and they might have had that rapid ARR growth. They've gone from, you know, zero to two million pretty quickly, but actually they're stuck because there's so many more companies coming on to market. And your point about companies trying to do things internally, being a bit smarter about the adoption of ARR and not just going all in on that. When you're selling... into that market and you're selling into recruitment companies ultimately which i can't imagine how that is it must be uh well it must be pretty tricky but how are you positioning that in a way that the founders the ceos there's the heads of ops they're kind of listening because obviously building a go-to-market team i know you've got a team in in florida you've obviously got one in one in the uk as well but how are you positioning that in a way where it stands out and you're getting noticed yeah very simply we go use case led so

We don't go broad attack, if you like, onto the market. We go specific. Each of the team identify about 50 accounts per week. We do quite a lot of research on that. And then we go, we lead with, hey, Alex, I noticed your recruitment company that does XYZ. That probably means you've got ABC problems, right? or you know you probably have more people here than you would like and the ratio should look like this and it's probably more like that so and you're right in in how you phrase the question a lot of time actually we are sometimes selling to it and cios but a lot of time we're actually talking to the business owners and really again going after the the problem the use case Whether it's speed to the pound or it is just we've got a lot of costs sat doing that kind of process and it's not scalable for us. That's really what we lead with. And that's, you know, you say nought to two million. We've gone over that already. So we've got the high growth, but then that kind of sustainable penetration, if you like, both UK and US, because we really understand the sector and the use cases.

Well, there's a lot to be said for just niching down in a vertical, right? I think there's a lot of companies that are trying to attack everything and everyone, whereas actually you really know the space. You're building Dwight specifically around recruitment, the various workflows. So smart. And I think if you look back to kind of when you launched 300-bit years, I know you said obviously the beginning was very much building the product, getting the message out there. What's the one bit of feedback that stands out for maybe those initial customers that really shaped the future Dwight?

Well, there was a lot of, I always say to business owners, you know, as we were talking about offline, I'm involved in some other businesses too. One, I like the verticalization, like you mentioned, because you have a high degree of repeatability. So one, you focus from a go-to-market perspective, but actually your product, like in our case, our integrations into, we've got 350 plus different platforms we integrate with. We've already got the integrations, right? And then you've got an element of repeatability because if you need to go in and out of field glass. it's typically a bit different for each client but not hugely so we can copy paste make the editor and so can our customers so it just helps with how quickly you can stand up the digital workers um so so i yeah i like that kind of vertical um approach what

was the second part of your question sorry the kind of bit of feedback that you've had from yeah any customer that's maybe stood out and you've gone ah that's really interesting i didn't think of it that way

And so the second bit I often say to, you know, tech company owners is be prepared to pivot. You know, however you start is not probably where you're going to end up. So that kind of listening to customers, getting feedback is really critical because actually, yeah, we're called Meet Dwight, but our trading name's Chatbot. So that's our trading name, but our actual name of the company on Companies House is Chat Automation. Because we start chatbot technology into these kind of high volume recruiters to try and help them see the wood from the trees and qualify candidates. And then it very quickly became apparent that actually, no, what people really wanted was this automation piece. And we were a bit clever in how we took it to market, giving it a name, all those kind of things. You know, you pay Dwight a monthly wage, et cetera. But outside of that, it was really listening to what the people really want. And we identified that kind of, you know, middle back office piece as being something where we could double down into. And then even when we named that piece of technology, Dwight, again, we then had...

feedback from the market i remember being at an si conference in florida where one of our prospective clients said you know this dwight just sticks in my head like because of the office and everything else right um yeah it was on your video wasn't it for the sia i think it was you the first picture you showed on the video was dwight yeah

because we do get asked right is it is that where it came from actually we're just playing around with acronyms that worked and Dwight's quite an unusual name. And yeah, of course, we've played it, but... officially it's not linked because i don't want to get sued just to be clear just to be really clear one of the points of feedback from this customer was you know i always remember dwight i can never remember the freaking name of your business automations i remember just literally going on godaddy and finding the domain meet dwight.com like within hours of that comment and 150 bucks later and here we are

Yeah, that's amazing, isn't it? You kind of the strength and the pivot is what you're obviously practicing what you preach, right? You're telling people that you're working with and investing in and saying, actually, especially in this market, I think it's really, and I see it from a hiring standpoint, on one hand, you know, two years ago, the type of profile we were hiring is drastically different to the type of profile we're hiring now. And I think those companies equally on that journey have pivoted. doubled down on something they've kind of adjusted obviously got news yesterday about chat gpt with a healthcare integration to epics that's a huge impact in to health tech because now they're thinking oh actually all those health tech companies like well actually our usp is no longer that unique or actually they can get stuff for free and what i'm really interested to touch on now is the the kind of joys of growing your business in terms of not the revenue numbers but the the headcount because obviously you're a startup a bit more mature now obviously but you When you look, recruitment companies have notoriously been really bad at hiring internally. It's just something we always joke about. We can't hire for toffee, but we can find really good people for our customers. Yeah, exactly. So for you specifically, what kind of traits and backgrounds do you look for? I know a few of the people that are working at Dwight that we mentioned when we met up. But what works for you and what is the message to people joining a company like Dwight?

It's probably a bit different now to what it was about a year ago, I would say. So then that evolves too, right? Because, you know, I've always been quite a big fan of homeworking and I am at home today for the purposes of recording this. But, you know, typically we're quite heavily office based. And that's really because, you know, we've been figuring stuff out in terms of I've touched on some big pivots, but there's like almost on a daily basis, there's little pivots. And a lot of that comes because we talk to each other. come off a call you know the water cooler type conversation where it's like i've just had this conversation i think maybe we ought to do this and so that you know to come back to your question that's been one of the key things that we've looked at is you know one thing we can absolutely still promise people that want to join us is is going to be constant change so if you don't do well in that kind of environment um then we're definitely not the business for you right so i often say to people i can't tell you what the business looks like in 12 months time but the one content I can promise you is that is the change part so that's a key component for us culture is really really important um it's not to say we haven't had a few missteps we have but generally speaking we figure things out pretty quickly and that's the other thing as well throughout our business i'd rather make mistakes on the way and keep moving forward than you know being overly cautious and and not really getting anywhere and i've given everyone the license to do that too so we absolutely do make mistakes but it's it's that old cliche of It's how quickly you learn from them and how quickly you adjust. So equally, we had some people that came into the business that just for whatever reason, whether it's them to us or us to them, wasn't the right fit. And we have those conversations pretty quickly because I'm, you know, when it comes to recruiting, I'm a firm believer that it doesn't do anyone any favors if you just kind of pretend everything's going to be okay when you're already in your gut, you know, it's not going to work out. It doesn't help the team member and it doesn't help the business. So, you know, having those kind of difficult conversations early on. is really important and we that's the other thing in our business we are brutally honest with each other good and bad um because again that openness allows us to develop and and make sure we have good relationships in you know like you said is still a relatively small business right yeah

and i think that the best leaders i've worked with or had on this are the ones that make those decisions quickly and equally people are joining her and she's relatively earlier stage growth business especially in AI they need to be fine with change because as you say every single day there's a bit of news there's a change the new frontier model released there's you know you need to be pretty resilient in terms of you now obviously you're you this run race you know rapid you're obviously growing it's actually got this international offices next 12-24 months what if you look back in say a year or so what What would you be happy with?

I just told you, I don't know what it's going to look like in 12 months. No, listen. Yeah,

well, that's it.

We have, of course, we have plans, right, that we want to execute on. And, you know, our revenues typically are, sorry, currently are about 60% UK, 40% US. We feel like there's a heck of a lot more runway to go, particularly in the US market in staffing, because on the face of it, both markets have 30,000 staffing agencies, recruitment companies, roughly, right? But the reality is in the US, they're typically a little bit bigger. And, you know, there's a lot more market for us to go out there. So it's definitely growth in terms of go-to-market teams doing more things like, you know, different events. In the UK as well, healthcare staffing is actually our biggest recruitment vertical because of the high volume, the high complexity, et cetera. And so we've also dipped our toe into healthcare itself, privately healthcare, both UK and US. So we've done a few trade shows this week uh sorry not this week this year and that's had some success or else as well so starting to maybe look at that kind of second vertical um and then the other key thing is to just keep moving forward so the other big pivot um was opening up the studio because people said hey we've got all these little automations that you can't help us with so you know this came into being in like six months um and the innovation in that's pretty insane like we you can record you doing a task on a web application or even even an as400 for people that remember that you know legacy system and it automatically goes and builds um what you've just done into the dwight studio and automates that and then you can kind of tweak it the edges so that that's the other thing is just keeping up with the innovation we've got a highly affordable platform that has some phenomenal functionality but a bit like sales you're only as good as your last month in terms of keeping up with what other people are doing. And, you know, but for me, I said at our last... board meeting that's actually keeping up with what others are doing isn't a concern of mine really because we've been ahead of the curve in most things Again, we're very fortunate that we've got quite a broad spectrum in terms of seeing what other people are using through different advisors that we have. And that's then very quickly put into the product. So the big thing there is more how long it will take, how much will it cost us to keep doing that. But with the growth we've got, we've got no concerns around that.

Yeah, and I really like the branding, the positioning. I know we met at an event ourselves, but I've come across Dwight before. You know, I think I said to you, you seem to be everywhere, which I think is the biggest compliment I can give you. You guys are obviously not relevant, unfortunately, for our business, but I know a lot of companies out there that would be crying out for that. And my final question to you, which I know you've not actually talked to us beforehand. If you're a founder of a recruitment company or you're at a larger recruitment company, now you're going to find, you know, launch your own. recruitment company at what stage do you think they need to be looking at a meet dwight and say is it when they get to a point where they need it or should this be one of the first things they look at when they're conceiving the business saying actually get that right now it's just going to make scale easier it's

at the point of scale up really not not immediately so it's good to have an awareness of what we do there's lots of other fantastic providers in the sector that provide you know more top of funnel type automation that may be more

relevant earlier on the process but no it's it's really when you're when you're thinking about multiple people in that kind of middle and back office it's really where we uh where we come into play yeah no interesting i implore and obviously i've got a lot of recruiter contacts and obviously hopefully hopefully listen to this but uh do check me dwight out because honestly i met treason an event i've seen you guys i know you've got some really good people working for you and check them out because they were generally if i had this as a founder and as a director of a business with thousands of contractors years ago that made my life far easier so uh i almost misfortunely weren't around before covid but yeah thanks so much for coming on this morning it's been really interesting be sure to share the website there's obviously a really interesting video that drew stood at sia was that the one in florida a couple of years back it was september last year so this time last year uh

i think it was in dallas yeah

perfect we'll check that out because it's really clear you can actually see dwight and see it in action and yeah it kind of blows your mind when you see actually working in real time so yeah thanks so much for coming on today thanks for having me cheers good man