← The Digital Executive
Loading summary
Transcript18 lines
- [00:01]
A
Are all batteries the same? That's like asking if all soccer players are the same. Take Messi, the most decorated player ever. Is there any other player who has achieved that? No, just him. Now take Duracell. Is there any other battery with powerboost ingredients inside? No, just Duracell. Remember, goats only trust goats because they're built different. And Messi only trusts Duracell. I'm Archmanning. I'm Madison Skinner. I'm Eva Jovic. I'm Decoria Moore. Want to train like a Red Bull athlete? Tell us your fitness goals this summer to enter the Red Bull Athlete challenge. You'll get to try each of our workouts for a chance to win an ultimate Red Bull experience. Think you have what it takes?
- [00:54]
B
Welcome to Corazon Technologies, home of the Digital Executive Podcast. Do you work in emerging tech? Working on something innovative? Maybe an entrepreneur? Apply to be a guest at www.corazant.com brand. Welcome to the Digital Executive. Today's guest is Neil Marley. Neil Marley is the CEO of NeoLogic where he helps mid market companies turn AI into practical business transformation. Over a 25 year career spanning global software firms, Microsoft high growth startups and scaling consulting businesses, he has built teams that solve real customer problems with technology. His path into tech started early, from gaming on a ZX Spectrum to hours in a pioneering school computer lab and has since taken him from enterprise cloud adoption to leading fast growing businesses through scale and change. Well, good afternoon Neil. Welcome to the show.
- [01:46]
A
Thank you for having me. I'm very happy today. It's my 50th birthday today, so.
- [01:50]
B
Wow.
- [01:50]
A
This is for me a momentous day, I guess. Yeah, but I know I'm, I'm. I was looking forward to doing this. I thought it'd be a great time to talk about what we're doing and where we are on my 50th birthday. Yeah.
- [02:01]
B
That's amazing. Your 50th trip around the sun. And I feel honored that I get to do a podcast on your birthday. That's the first that this has happened. At least that I. I know of. I don't know if the guest has ever had a birthday during their podcast, but they never shared it either way. But I appreciate that. So thanks for sharing your birthday with me.
- [02:20]
A
The address below. Exactly.
- [02:22]
B
Awesome, awesome, awesome. Neil, let's jump into your first question. You started with ZX Spectrum and a school computer lab, spent 25 years across Microsoft, global software firms and high growth startups and eventually founded Neologic to bring practical AI transformation to mid market companies. What did you see happening in the market, particularly for companies below the enterprise tier, that convinced you that this was the right problem to build a company around.
- [02:49]
A
I think the first thing was previously to this business we were running a training business for mostly younger people to get onto the technology ladder in that business. One day we came across the first version of ChatGPT and the CTO and I looked at that and said this is going to change a lot of things inside businesses, particularly information worker roles. And as you look forward then and go what's actually going to happen and then how can I create a service that customers are going to love, support that? I think the mid market organizations, we've typically worked 100 million to a billion turnover dollars of turnover companies. Unlike large enterprises who have invested lots of people, time and energy in it, those organizations typically are more stressed on time, resource and expertise. And so I think the second thing is about those mid market companies taking opportunity afforded by the new features, services and so on in AI is that those companies have got the most to gain from disrupting the incumbents in their industry. So actually smaller organizations can move more quickly, be more agile, reinvent themselves more quickly than the larger companies. And I think the third thing that is happening is what's happening to software in general in the market. We think it's heading more and more to a sort of on demand state. And a lot of the organizations we meet are using the software vendors and tooling and we're showing them what you can build with AI in a fraction of the time. And so that, that like perfect storm of events is coming together for us to say we can go into the mid market companies, we can say here's a, an enterprise grade platform, you know, the Microsoft platform we work with, you can deploy hundreds and hundreds of workflow agents on demand to do exactly what you want and we'll help and support you on that journey. That's really resonating as well now with the market as well.
- [04:26]
B
Yeah, it's amazing how fast AI is advancing so quickly that as you talked about, you can actually build things that will take your company or your platform or idea to that next level. But I like your backstory. You were in that training business and you had that aha moment when you first saw generative AI and you're like gosh, there's certainly something here that we can leverage. And of course you did talk about those companies, especially the smaller companies who are more agile, have to be move around quickly, can be in a position to disrupt their industry. And I totally believe that so. Thank you, Neil. Neologic's platform is built around intelligent agents designed to automate govern business workflows with an emphasis on governance as much as automation. Why is that distinction so important? And what goes wrong when companies automate workflows without putting governance guardrails in place?
- [05:16]
A
On the basic level, when you're, if you're a prosumer or consumer user of say Claude or whatever, ChatGPT, you can afford for it to do certain things you don't expect. You can ask it to create a document or an output or an agenda or something and it can give a different answer every time and that's okay. A consumer level, you can deal with that. Enterprises, whether small or large companies, they need certainty of outcome. I think that's number one. So the thing that provides that is what information you feed it, how you prompt it, how you, how you control it, you never 100% control it, but how do you get to a point where you can control and govern the output? So they're reliable is really, really important. And I think the underlying LLM services don't provide that. They provide the capacity, if you like, they don't provide the bits. We call it scaffolding, business scaffolding that sits on top of it. And so we've learned, and it's fairly obvious, I think, that companies need reliable results. They need other supporting features like audit, they need to understand cost control, they need to understand evals and how quality of the answers is working. Are the answers drifting? What are all these new governance features now? Microsoft and Amazon are building some of those into their platform and they're developing that. But we see particularly in regulating industries that we work in. So mostly it's construction and financial services that we work in. There's still a big gap between the hyperscaler vendor platform and what our customer actually wants. And that's, that's true and very simple workflows. So create document with AI, you can put that into a neologic workflow and it'll do it in a very structured way. But when we get to autonomous agents calling each other, I think the problems are much harder. How do you, if you're a bank and you're running an autonomous series of workload processes and AI, at the end of one part of the process, the agent decides to go either A, B or C to the next agent. When you look at that map and that mesh, if you like, of agentic workflows, we think that's going to be a problem, which for us is an opportunity. How do you how do you then put a thread through those decisions and give auditability to that? That's what people actually want. So we think the governance layer is incredibly important and we think that as AI agents loose term develop in industries, we're going to see more and more problems which will require scaffolding and governance services to fix. So that's what we're building with Neologic. We've just deployed the platform into an organization in the us, actually a bank. They're not public yet with their sort of press release, but they've gone in and done a whole bunch of work on compliance around FFIEC regs and other parts of their industrial regulation and compliance. And we've deployed that agent to fix that specific problem for them about how they can manage certifications, regulations, gaps to audit and things like that. That is transformational for them. But they couldn't do it if we didn't have those guardrails and they couldn't stand behind the answers that came out of that. I think the other part just on governance is where is the data being held? This is really important for a lot of organizations. The hyperscale platforms, whether it's Azure or Bedrock, will call AI providers. A lot of them are outside the EU from our point of view, we're in London, some of them are available in all regions. But there's definitely a challenge here about where is the data going and who is processing it and for what purposes and how that's being managed. So our solution allows a customer to deploy the environment on their own subscription, their own tenant. So that's another thing for banks and other regulated industries, the data doesn't leave their environment. So I think these themes, while slightly less exciting than some of the funky things you can do with AI, are absolutely fundamental to production rollouts inside larger organizations.
- [08:55]
B
Thank you. Really appreciate that. There's a lot to unpack there. But you talked about that at the basic level, obviously consumers are. They've got the app ChatGPT on their phone and you can get 95, 99% of what they need. But you talked about businesses needing those guardrails. Quality, accuracy, compliance, governance. I liked your term scaffolding, by the way. I think that's a great term. And then you talk about data storage, especially in these regulated industries like banking. When you start using gentic AI, the workflows of processes get a little bit more complex. This is something that you're going to need more of an industrial solution. But I liked how you're offering this for the midsize companies. As well. So thank you Neil. Most mid market companies have run at least one AI pilot by now and very few have scaled what consistently separates the companies that move from a promising proof of concept to genuine business transformation. And what's the most common reasons AI initiatives stall before they deliver measurable roi.
- [09:53]
A
I recently did a presentation with the UK AI Federation over here in London about this topic. We've been working with companies now for about 24 months, 18 months in and more latterly in the last year. A lot of production rollouts and I think there's five things that really cause a problem. Number one, the biggest one is that AI is a business transformation, not a technology project. So one of the reasons that projects don't make it over from pilot to production is they don't do anything genuinely business impacting or useful. It's quite easy to get into that trap of proving a technology but actually not proving the business outcome. That's the first one. They don't add value. I think the second one is a lack of a clear comm strategy from the board or from the leadership about why people are doing AI. We see this quite a lot where someone just announces we're doing this and there's a lot of handbrakes and fear inside the organization to say whoa, why, why are we. I don't understand. I'm worried about my job, I'm worried about future, I'm worried about what's going on. So I think there's people are pulling back a little bit from going into production for that. I think people choose the wrong use case. There's lots of things you can do, but do they back to the business outcome? Do they drive the the right outcome? Are those use cases easy to do or are they really hard? And that's actually it's interesting. Some of the ones that look easier are hard and vice versa. Do those agents use cases have a safety or material critical risk to them which you shouldn't do for the first one? I think the other two bits are probably more cultural. Okay, well speed. So we see people create committees spending six months trying to fix a discovery to create a plan. That's a complete waste of time in my opinion. I think AI is moving so fast the idea of doing you start looking at the possibilities for your business and by doing a planning exercise for six months everything changes after that. So I meet a lot of companies are going through committee cycles before they realize actually you've just got to dive in. I think you've got to get into it. You've got to learn on the job. And that isn't really how we've deployed technology in the past. We've always done thoughtful discovery. I think AI is now a much different proposition. And the last bit is I think relying on tools to expect transformation. If you think you can do AI just by rolling out Copilot or, or if you can buy an app, a SaaS tool that does something with AI and think you are AI'd, I think that's also a false dawn. I think actually the opportunity for organizations to go into production with AI is actually much more about transformation of their business models and the services and customer experiences and other similar things. So it's not about tools, it's about how do I transform the business that I'm doing. So if you see an AI transformation vendor come into you and they say book a demo, you know you're buying a tool, not a transformation. So we think to make it stick, it needs to be part of our strategic effort. The board has got to communicate why it's got to be driven by business owners. Need help selecting the right use cases and go really fast at how do I deliver 1, 2, 3, 4 and tens of use cases into the business in six months, nine months? I think that's what makes the AI rollout stick in production in our experience. And we've done quite a lot of them now.
- [12:59]
B
Thank you, Appreciate that. Again, some great stuff there for the audience today. And you spoke about this, you said at the UK AI Federation recently, but kind of highlight those things you talked about. AI is a business transformation, not a pilot project. Lack of clear or common strategy for AI deployment is really important to focus on. And then obviously you talked about measurable outcomes, speed to delivery. It's a cultural item, obviously you got to get people on board. And of course transformation is not about the tool, it's about. I see this all the time. People talk about that they've piloted co pilot for two years and they think they're AI ready, but gotta be driven by the business for sure. It's gotta be part of that strategic plan. So thank you and Neil, last question of the day. Ernst and Young's research identified 2026 as a tipping point for AI to move from pilot to enterprise scale. But what does that inflection look like for mid market companies? Specifically, where do you see the practical AI transformation landscape, let's say three to five years. And what do you believe Neologic needs to get right today to be the platform that defines how that segment makes the leap?
- [14:04]
A
Yeah, I think it is A tipping point. I think we're seeing as in our engagements more and more. I think it's reached a certain critical mass now. I think a couple of things about the three to five year horizon. I think one is I don't think humans are going anywhere. You know, there's a lot of talk about AI replacing people and so on. I think, I think human first, AI enabled is the future. I think that means actually a whole bunch of people's jobs will change and I think new jobs will appear and some will probably not be there in the same way. But I think the opportunity, I'm very positive about the opportunity. I think the way that AI will particularly software and services vendors will intercept this. I think software vendors, the very interesting time to be a large AI SaaS vendor, as we talked about earlier, how much of that can be recreated on demand. What does that mean for SaaS vendors? I think the large ones will stay ERP, CRM and others, systems of record, high, high risk to move and so on will stay and be with us for years. I think the, the threat, if you like, is that the smaller software companies, we call them single shot SaaS, they do one thing, they're quite functional, they're quite departmental. I think a lot of that work will move from those vendors to AI driven solutions. And I think those solutions will change therefore from being customers buying 100 tools to customers buying a platform and building 100 slices of workflow. And I think it'll re aggregate the functionality in a different way. I don't think it's going to be lots and lots of vendors. I think it's a good thing for customers. I think customers will have many less vendors to manage. I think they will have the ability to full code customize exactly what they want. If you look at a SaaS tool today, whatever it is, you know, an RFP response tool or something like that doesn't matter, there's lots of examples. You probably don't use anywhere near 20% of the functionality and the vendor has to provide multi tenant solutions that are the same for everybody. That's part of their scale process. So customers have always been slightly separate from what they actually want and to what vendors can provide, there's always been a compromise there. I think that's changing now. So that on an AI platform where if you've got the right governance and scaffolding, you can instantly spin up hundreds of workflow tools, agents that do a thing that you need and it'll do exactly what you want, you can customize it fully And I think that's going to change the way people think about software. I think our job as neologic is to create a library of hundreds and hundreds of use cases of those things, those workflow agents, those slices. And we're seeing this now in finance and construction particularly let's build solutions for construction that talk about estimating from drawings or models, or let's talk about cash flow optimization from WIP and accrual revenue lines, or let's talk about very specific industry solutions and we want to have hundreds of those. And in finance, similarly for compliance for FFIEC or for FCA or for MITFID or whatever the compliance solution is, kyc, aml, there's all these areas of possibility and we're building immediately useful answers. I think the other interesting part of that is it typically takes us days to stand those up, not months. And I think that is a threat to the larger system integrators out there who have been used to spend 50 to $100,000 to do a proof of value or discovery exercise and then significant six figure work to do apps. That's been the sort of model, I think that's all being thrown up as well into the next three to five years as it becomes easier, not easy, but easier to create solutions on demand. So we're building things in days and deploying them in an enterprise grade environment and I think our customers are going wow, what can I do next? And I think that's going to displace a lot of those incumbents. So it'd be interesting. I think the number one thing therefore is we need use cases and a library of reusable solutions for our industries. We're starting with construction and finance as you said, because I think the regulations in those worlds are really interesting and a good target for what we do. But we'd like to have many more of those and I think that'll put us in the best possible place in three to five years for whatever comes next.
- [17:59]
B
Thank you so much, really appreciate that. And just to highlight a few things here, Neil, there is this big talk around, around humans being replaced and you, you said very frankly humans aren't going away. In fact it's human first AI enabled and I really like that. AI will certainly change the work, how work's being done for sure. We, we see some changes coming to how humans interact in their current industries or roles. For example, I noticed here you said the threat that you see is really to the smaller single solution SaaS providers that with AI you're able to build these apps again in days or hours, not months. And the upside at the end of the day is there'll be less vendors and platforms for the consumer that they'll have to manage. And I think that is certainly a positive for sure. So appreciate those insights. And Neal, it was certainly a pleasure having you on today, and I look forward to speaking with you real soon.
- [18:48]
A
Brilliant. Thanks, Brian. Thanks for having me. Appreciate that. I enjoyed that. Thank you.
- [18:52]
B
Bye for now.