The future of supply chain planning

June 21, 2023

And the importance of setting up planning systems

Listen on:
Listen to the Tech Story Podcast on Apple PodcastsListen to the Tech Story Podcast on SpotifyListen to the Tech Story Podcast on Google PodcastsWatch the Tech Story Podcast on Youtube
The future of supply chain planning

Listen to the episode

In this episode, David Evans, co-founder of Oii.ai, shares his background and how his experience in supply chain planning led him to start a company that leverages science and AI to optimise supply chain planning. He discusses the critical importance of setting up planning systems correctly and the complexity of doing so. David also talks about the challenges his company faces in the market and how they are working to overcome them.

The key moments in this episode are:

00:01:55 Planning is critical.

00:07:47 Supply chain failures.

00:08:46 Overriding planning systems.

00:13:49 Supply chain maintenance.

00:18:30 Creative aspect of entrepreneurship.

00:20:05 Sales and recruitment mistakes.

00:24:45 Stochastic planning system.

00:28:02 Investing in legal early.

Watch the episode

Read the transcript

00:00 Welcome to the Tech Story podcast, a place where we interview founders and interesting people in tech, growing companies in London. I'm very excited to welcome David Evans on the show. David is the co-founder of OII.ai, a company which leverages science and AI to optimize supply chain planning. David, thank you for being on the show. Great to be here. Would you like to share a bit of background about yourself and the company?

00:28 Yeah, let me do that. So my background is I spent most of my career in the corporate world actually. I spent a long time with GSK, MSD in the pharma industry. And all of my career has been in supply chain. My educational background has got a business degree, but I've also got a master's in computer science and supply chain. And my big focus throughout that time has really been supply chain planning and in particular supply chain planning systems. So when I was at GSK, I led what was then one of the biggest deployments of Managistics. It's now called JDA. And I've done similar things at other companies. And I suppose the story of OII.ai, and it's quite difficult to say right, is that I kind of noticed that there was a real kind of issue with these planning systems. And the issue is that supply chain planning systems are completely manual in terms of the way you set them up. And setting up a supply chain planning system is actually a really complex task. So why is it critical that that is right? And why is it complex to get it right? Well, if you imagine over and above all the physical things that go on in the supply chain, I would say that the most critical success factor is planning, right? Planning determines what you make and when you make it, how you make it, what it costs you to make it. And all of the other peripheral activities which determine whether you get stuff to the customer are down to planning. And so planning is critical. Planning systems essentially determine exactly what happens and when it happens, yeah? And of course, if that process is manual, then the manual configuration of the planning system essentially determines how successful that planning system is, but it also determines how effective the supply chain is. So I saw that the manual configuration of those systems was a big issue because it drove service risk. It drove issues with inventory and waste. It drove issues with underlying costs. And so all this kind of inefficiency being really almost being the kind of the standard paradigm of supply chains. And I guess one of the interesting things of the pandemic is that everybody became aware of supply chains. They became aware of supply chains largely because of the Black Swan events of Covid and all things breaking down. But actually that underlying model of these very complex networks with lots of variability really behaving in an incorrect way for me is being driven heavily by the way those planning systems work. So I suppose the start point for OII was to say there's this big issue with the current paradigm. The planning systems are driving all this waste, all this cost, all this inefficiency. Is there a different way to do this? And then I was this was about four years ago. I had this conversation with this guy called Bob Rogers. Bob kind of reached out to me via LinkedIn. He was writing a book at the time on supply chain. He was writing a book on AI for the enterprise and he wanted to do a chapter on supply chain. And he reached out to me. I'd written a few papers on various aspects of supply chain. And we kind of hit it off. And through that I said to him, well, look, I've got this problem. Supply chains are kind of inefficient. They're manually configured. It's driving lots of inefficiencies and costs. Is there something we could do? And that was really the start point of OII to say, OK, could we use Bob's background in advanced modeling, AI, neural networks, digital twin technology to actually fix this particular problem? So that was the start point.

05:12 And here we are. That's great. And I guess where are you on this journey?

05:18 Are there any supply chains that are now configured with OII.AI? Yeah, no, exactly. So we are, I think we're still relatively at the start point of our journey. I think OII.AI is a disruptive technology in as much as it's a new approach. We don't replace the planning systems, but what we do is we plug in. So if a company's done a hundred million dollar deployment of Kanaxis or OMP or 09, then I want to replace that. But what our technology does, it plugs into that, takes the data from those systems, uses that data to kind of create this automated digital twin, which we then simulate how that supply chain is running, what risks it's driving, what costs it's generating, what waste, inventory, all of those things. And then we through the simulations find the correct set up. So through deploying OII.AI, you get this understanding of, OK, our current supply chain, we would expect to see it perform like this. And actually, we change it to this. You know, we will get these benefits. So we're relatively at the start of that journey because it's a different approach. And any new approach, any new company, you've got to get over that hump of people going, oh, you know, we're quite comfortable spending lots and lots of money with the established companies, but do we want to do something with a new company? That said, you know, we've done deployments with a number of companies. And I don't want to go into the details of exactly what we found, because, you know, obviously, some of that, you know, they probably wouldn't want to be showing in the public domain. But we've done deployments with three pharma companies now. And we've done deployments with three companies in the FMCG sector. And it's interesting what you see in the pharma world. We did we did a recent deployment with one of the very big pharmas, a pharma company that had actually done a major deployment of CanAxis, which I think is recognised as the, you know, the number one advanced planning system right now. And what we found was as follows. Surprisingly, we found that about 60 per cent of that company's items were kind of set up to fail from a service standpoint. And what that means is, is that, you know, if you took a hands-off approach with your supply chain and just let the planning systems work and the supply chain activities follow those planning systems, 60 per cent of the items would have had stock outs and would have not achieved the service goals of what you'd expect, which is which is really high. You wouldn't expect that in the pharma world. At the same time, what we found was was that the company had too much inventory in terms of their planning parameters and not wildly too much, but definitely too much and a cost imbalance of about 25 per cent, i.e. could you reduce the cost of running that supply chain by 25 per cent? The other thing we found was, was that and I think this comes back to the service thing, the company actually had twice as much inventory as what you would have expected to see based on the planning parameters. And what that told us was and this was, you know, consistent with what the organisation was seeing, was that because the systems were configured incorrectly, the people within the supply chain were actually overriding the system. So the manufacturing organisation was pushing stock into the channels because they didn't trust the planning system. You know, and then on top of that, you know, you're chasing stuff, etc. So they had twice as much stock as what they should add because basically they weren't really following the system. So that was very interesting for us. We saw a similar pattern in the other pharma companies, not so much on the service side, but certainly way more inventory than needed, very high discards, etc. So a real big opportunity to reduce costs and reduce waste and reduce underlying stock, you know, which, you know, when you projected the savings, you know, the savings within those pharma companies were at the level where you could actually move that cash and invest it in a whole drug discovery programme. So really big numbers. In the FMCG world, you know, we've done work, again, with a couple of the big multinationals. We just did a really big project with the food producer, their Indian division. And again, a similar story. The big headline there was, you know, we kind of expected in the FMCG world, you know, because their margins are smaller for the inventory to be tighter. But what we actually found was that every single item for this company had the same planning parameters. You know, so you know straight out the gate, well, they're all wrong, right? That company, 38% opportunity to reduce inventory, which was huge for them, especially when margins are tight and cash is tight and, you know, as money becomes more expensive, you know, that becomes a much bigger thing. The other thing for them was that as we were able to reduce their inventory, 14 days, you know, more shelf life on the shelf, you know, because things were going through the supply chain much quicker and therefore, you know, getting to the customer, you know, getting to the point where the customer could buy it with much more shelf life, which has a big impact on, you know, whether it gets sold and also, you know, good for the customer as well. So I think, you know, we've done projects with I think about seven companies right now. And really that's been over the last sort of 18 months since we got funding and really able to sort of start to scale up. Right at this moment, we're in conversations with about 10, you know, big names, right? So, you know, not all of those will necessarily go or, you know, some of them will take longer, you know, the sales cycle takes a bit of time. But, you know, we're really starting to see that pick up. And, you know, it's like a domino effect, you know, as you start to get, you know, people don't want to be the first one to use a not so well known company. But as they see other names coming on board, then they jump on board. And, you know, the big opportunity for us with this is that every planning system is manual, right? And the planning systems are doing very well, right? They don't need to kind of, you know, change the way they do stuff. They're selling out the gate, right? So all of the planning systems are manual. There's no real plans to kind of change that within those companies. So the opportunity for us to plug in and really start to, you know, fine tune those systems is huge. And there's one other thing that I kind of, you know, sort of need to touch on, you know, there's the issue of planning systems being manual. And that's a big issue. Because they're manual, even where, you know, companies spend a lot of time trying to get those parameters as good as they can. You know, sometimes they'll dedicate, you know, a very expert group to try and do that. It's not an exercise where you can do it very often. So what tends to happen is, you know, even where companies are actually fairly on top of it and the pharma company that we that I mentioned, you know, did a lot of work in this space. It very quickly goes out of sync, right? Supply chains are highly dynamic, you know, you know, demand patterns change, you know, variability in lead time. Performance changes, variability in factory reliability changes, you know, and as those things change, they very quickly get out of sync. What our software does, it doesn't just plug in and get it right. What it does, it plugs in and then maintains it. Yeah. So you get, you know, we called it kind of turn by turn. It's a bit like, you know, you know, when people went from maps to, you know, to GPS, you know, you, you know, as things change on the route, you know, the route, you know, changes accordingly. And it's the same with this. As things change in your supply chain, it constantly keeps it maintained. So, you know, the good thing for us is as companies come on board, you know, they're coming on board and then they want it, you know, as an integrated aspect of their supply chain, constantly making sure that their parameters are fine tuned. So, so we're fairly at the start point. You know, we've really, although, you know, we've been working on it for four years, it's really the last 18 months where we got funding. We were able to bring people on board, you know, to sort of scale the company. And we're really starting to see traction now. So it's exciting.

14:50 Yeah. Well, it sounds like the impact is huge. And even though from an entrepreneur's perspective, or from an outside perspective, it's it would seem like the obvious decision for any company to use OII to optimise the supply chain. But I it's the same with us when ultimately you're a newcomer and your clients or prospective clients have a solution in place, even if it's not the best solution or it's not optimal, it's still a solution. And so change, selling change is always a challenge. And I think it's you're right, it's just about getting that momentum. And then and then eventually everyone comes on board.

15:33 No, exactly. You know, and you and you want to be, you know, being innovative is great, but you don't want to be too far out of the curve, right? You know, you know, you want to be a kind of Led Zeppelin, right? Rather than a Velvet Underground, you know, you don't want to be that kind of niche thing, which, you know, everyone gets 20 years down the line. But I think it's really picking up now. And the other thing for us is we've started to really focus on the finance story. You know, supply chain people are often quite comfortable holding lots of stock and supply chain people, you know, and I'm a supply chain person, you know, they love kind of fixing things, right? You know, so the firefighting aspect of supply chain makes it quite exciting. You know, you're chasing stuff, you're putting it on low rates, you're the kind of hero that fixes stuff. Yeah. As opposed to, you know, having everything kind of like, you know, really, you know, proactive and smooth. The finance people, especially in the current climate, you know, when you say we can reduce your inventory by 30 percent and still maintain your service goals, they say, right, OK, we want you on board. And actually, we're speaking to a big FMZG actually on Tuesday next week at the CFO level, you know, because, you know, they've got a big inventory program, you know, they've been struggling to sort of figure out how to kind of do it, you know, so, you know, you're pushing on an open door. So, yeah, so all of those things, I think, are starting to kind of get traction and, you know, as things, you know, as you start to hopefully get on that hockey stick, that's the exciting part of the story. Well, best of luck and I'm sure you will. And so since starting the company, what's been your favourite moment so far? I think for me, you know, I have enjoyed the creative aspect of it. You know, I've always been in that world of kind of systems and business. This is great because, you know, in a small organisation like this, you know, I'm the person that's kind of designing those system screens. I'm not building them, you know, I'm working with the, you know, I'm working with the technical people who are doing the underlying sort of modelling stuff. I'm working with a really great creative team that sort of take your ideas and turn them into, you know, something, you know, which can actually be used. And that's the bit that I kind of really love. And, you know, that bit's always evolving, you know, you're always, you know, as you go into new clients, you start to see other use cases for your software. You start to see other kind of reports or screens that they want. And that bit's always developing. And, you know, that's the bit that I find most exciting, you know, sort of seeing it come to life, really.

18:30 Yeah, I think, you know, when you're a founder and you've got the idea and then it sort of comes to life, it's always special. And sometimes it doesn't come out as you anticipated or it becomes something beyond, you know, that can also be positive. It's not always negative. But

18:48 that's absolutely true. I mean, we're really lucky. I mean, my co-founder, Bob Rogers, he's a great guy. You know, Bob's background, he's a Harvard physicist, so a really kind of clever guy. His career is crazy, right? You know, I think he started off, you know, using digital twins to model black holes and all of this sort of stuff. But, you know, on the modelling side of things, you know, we've got a team which I think, you know, modelling and AI, I think we beat anybody. You know, one of our team, a lady called Madeleine Udell was Obama's AI expert during his election campaign. So it's that level of skills and capability. If I come up with an idea, you know, I know those guys are going to come up with a solution. Exactly. And at the front end, you know, our kind of UX team, you know, you come up with a pretty idea in terms of, you know, how it looks. And it's always better, you know, which is kind of nice. You can claim it's your idea, but then you go, it's better. Yeah. So that's exciting. Great. And so what do you wish you'd known before starting the company? Well, I think, you know, the tough thing with starting a company is you're kind of a hustler, right? You know, and the sales part of it, you know, the design bit, you know, the supply chain bit, you know, even the systems bit, you know, was always a strong card. The sales bit, I've never been a salesperson, right? I think it would have been great for us to get a salesperson on board sooner. But actually, the journey we've been on, which is always, you know, we always talk about, it's been very sort of synergistic. And, you know, the synchronicity has been great. We bought a sales guy on board about a year and a half ago, and it didn't quite work. We recruited them, you know, we didn't quite know, you know, what would make a good sales guy. And so we kind of forced it, and it didn't quite work. We've now, you know, actually a sales guy reached out to us, saw a podcast that one of our founder did in the US, got excited and said, I really want to be a part of this. And even came on board, just said, I'll do it for kind of shares. So I suppose the learning for me would be to kind of relax a little bit and not kind of force things. You know, we force things with the sales. We weren't sales guys, you know, we got some sales traction, you know, but we wanted to force it, you know, we bought a sales guy, but it didn't quite work. Sometimes sort of allowing things to happen, you know, knowing what you need to do, but not kind of forcing it would probably be the learning. It would be great if we did learn to be sales guys, but we would never be sales guys, right?

21:37 You know, we just, it's not an all we are. Interesting philosophy. I mean, I guess from my perspective, I had some sales experience before starting the company, but I also feel that if I didn't push put myself in that uncomfortable position of needing to sell, then the product wouldn't sell itself. So how do you sort of balance that?

22:01 I don't know, it's exactly that. And, you know, we weren't unsuccessful, actually. I mean, you know, we closed, you know, with that big pharma that I mentioned, you know, without any sales guy on board. So it's not impossible, right? You know, but I think you also have to kind of recognise, you know, your strengths and your weaknesses. And, you know, there's, there was definitely a need for us to kind of get somebody, a sales guy sooner, I think.

22:37 But yeah, it's- Used to be the right one and the like.

22:41 It's exactly that. You know, the recruitment bit is, recruitment is kind of crucial, I think, in a small organisation. In a big organisation, you can afford to get that wrong. In a small organisation where you're kind of burning, you know, investor money, you can't. And, you know, with hindsight, we'd have probably not, we'd have probably waited until we had the 100% right person. We knew it wasn't 100% right, we thought, okay, let's go with that.

23:07 Yeah, I think that's a great lesson. So, yeah, for any prospective founders listening, don't hire at all costs, unless it's the right fit. I think, I think it's exactly that. Great. And so you've, you're starting to see some traction and climb that hockey stick, what's sort of the vision for the company for the next three, five years?

23:27 I think the vision is this. I think we've, we need to keep selling the product. And I think that's really starting to happen now. I think we need to keep developing the product. You know, I think in terms of supply chain design and configuration, you know, automating that, we, you know, I don't think there's anybody out there that does what we do. And so that's exciting. I think the opportunity for us is, you know, how to kind of develop that part of it. Whilst at the same time, you know, keeping an eye on, you know, the planning part of the story, because one of the interesting things about our software is that under the hood, it is essentially a planning system. It mimics a planning system in order to configure a planning system. And it mimics a planning system in a very interesting way. You know, most planning systems are essentially a one stop hit, right? You know, you do one forecast, you know, you do one configuration of the network, you do one, you know, configuration of all the different planning parameters, and there's your plan. And if it's wrong, you know, you're out of kilter. Ours is essentially a stochastic planning system, because it runs thousands of simulations of how it should be. And therefore, it kind of gives you a range. Now, one of the opportunities for ours is to say, OK, do we go into that planning space? Now, we wouldn't want to go in there now, because, you know, we're too small. But as we start to grow, you know, our planning engine is, I believe, quite revolutionary. And lots of, you know, I sort of podcast yesterday, actually, from SAP, who were talking about stochastic planning and how they wanted to start looking at it. Well, we're already there with that. So that's an opportunity. So I think it's continued to sell, it's continued to develop the product that we've got, and it's continuing to keep an eye on, you know, where we can potentially, you know, branch out and go into

25:40 other areas. That's very exciting. And I guess it's one of those stories where once you do gather more data and more market share, then ultimately you can start to replace the incumbent.

25:54 Exactly. You know, and maybe that's partnerships. Maybe it's, you know, we step into the space and somebody comes and says, OK, we need to buy those people before they take our lunch away. Or, you know, maybe it's, you know, we do actually go in there and,

26:09 you know, disrupt that part of the market. So it's exciting. Yeah, well, very exciting and best of luck. And as a busy founder, I imagine you must come across quite a few contracts or legal documents. What are maybe the key ones and what can you share about them?

26:26 The key ones for us thus far have been, I suppose, two. You know, NDA is obviously, you know, we have to be, you know, as you start to have conversations with the organisations, we don't want, you know, them to kind of take our stuff. At the same time, you know, where we're taking data, etc. from organisations, then obviously they want to be able to trust the fact that, you know, they can trust us with their data. So NDA has been a big one. You know, we've got mutual NDAs and, you know, we've spent quite a bit with lawyers to make sure that those things were all, you know, tidied up. I think contracts, you know, with our customers, which again need to be written in a certain way, you know, contracts in terms of, you know, initial engagements, but then also, you know, ongoing discussions. And then I suppose finally, you know, contracts to cover, you know, shares, etc. You know, so we as an organisation, and I'm sure, you know, you see this with a lot of new companies, you know, we pay people both in terms of, you know, normal sort of salaries, but we've also paid a lot of people in terms of, you know, shares, you know, and that's a big endorsement for us that people are prepared to come on board and do stuff for shares, but then, you know, of course, you've got to make sure all of those things are covered up. So we've, you know, we've spent a bit of money with lawyers along the way, but I think, you know, that's a

27:55 good investment for us to make sure those things are done right and they're watertight. Yeah, a common pattern on the show is that we do find that it's important to invest in legal early on, because if you don't invest today, then you might lose IP, you might, you know, if employees do leave, then they might be able to create competitors. And it's ultimately a problem that can be solved later down the line, or might not be able to be, you know, you might not be able to solve it, but the cost of solving it will just be so much bigger than any legal investment that you do today. So start early with your legals is the key. Definitely. Great. So I'm conscious I've taken a little bit of time, so I'm going to ask you the closing question we ask all our guests. What's your favourite tech product? It can be software, hardware, it can even be your platform,

28:51 but yeah, what's your favourite tech product? Well, obviously my favourite tech product is OII.ai and our product Opti. But I suppose leaving that aside, you know, you know, I love all the Apple stuff and what have you, but the reality is our business could not have got to where it is without Zoom. You know, we're in San Francisco, we're in the UK, you know, we're in other parts of kind of Europe. You know, we could not have got to where we got to without Zoom. And in fact, 10 years ago, you know, this business couldn't have gotten off the ground. So I, you know, it frustrates me

29:27 from time to time. But I think, you know, that is has been a key enabler for us. That's a fantastic story. And we also have half our team in Spain. So without Zoom, equally, we wouldn't have been able to have the setup that we've got today. And yeah, great, great product. And yeah, it enables communication. So definitely. Well, thank you very much, David, for being on the show and best of luck growing OII.ai. Thanks very much. Really enjoyed it. Cheers.

Other episodes:

Clock icon
Listen on

Subscribe to our Newsletter

🚀 Early Access
👋 Community Polls
🤩 Shoutouts & Fan Participation