Decision Intelligence in the Supply Chain with Fred Laluyaux of Aera Technology

Patrick Daly:

Hello, this is Patrick Daly and welcome to Interlinks. Interlinks is a program about connections, international business, supply chains, and globalization, and the effects these developments have on our life, our work and our travel in today’s world. Today on Interlinks, we will be talking to Fred Laluyaux, who is the CEO of Aera Technology, a company headquartered in Mountain View, California, the heart of Silicon Valley. It’s a company that works to transform the future of work through decision intelligence, and we’ll find out what decision intelligence is presently. So founded in 2017 Aera has a global team that serves customers throughout the world, including the UK and Europe, Asia Pacific and North America. And Aera works with many of the largest worldwide brands in consumer package goods, food, beverage, healthcare, and life sciences, manufacturing, and more. And Fred himself is a thought leader on the future of work and decision intelligence for the enterprise. And I’m delighted to have him here with us today. So you’re very welcome, Fred. Great to have you here.

Fred Laluyaux:

Thanks for having me Patrick.

Patrick Daly:

To kick things off, Fred. Maybe could you share with us kind of a brief overview of your background and what led you to start Aera technology?

Fred Laluyaux:

Sure, sure. So I started in technology right after college. So, that’s a few years back. Spent the last 25 years in the world of enterprise software, building applications to help large enterprises managed their performance. So career took me through different methodologies, approaches, techniques, technologies, and I realized about 10 years back when I was working at ACP that there was a massive tidal wave of transformation that was coming, as organizations we’re going to be forced to make more and more decisions, more accurately, closer to the point of impact, closer to real time that was led by digitization of the economy.

Fred Laluyaux:

And on the other side, the people that are making those decisions well, they cannot work 24/7, they’re not machines. And there was an opportunity to digitize a lot of those decisions with some new and emerging technologies. So that’s the Genesis basically of Aera. After SAP, I grew a company called Anaplan in the world of data modeling, and realized that we needed to go from people making decisions, using data tools, collaboration platforms, bespoke applications, to an era where those computer systems could actually make and execute decisions guided by people. So that was the Genesis of Aera back in 2017.

Patrick Daly:

And so Aera today is working, as we said, with companies all over the world in multiple sectors. And so tell me a little bit about those companies that you work with and what’s on their minds today? What’s front and center for them? What kind of challenges are they facing?

Fred Laluyaux:

Yeah, I touched on this when we launched the company. The realization that the acceleration of business cycle was a given, and could only go in one direction, which is faster. That the complexity of the decisions that have to be made is only increasing, companies have to think about a given problem across multiple dimensions. And all of that is front and central and creates a lot of model necks. And you compound that with the macro level, the disruptions that we’ve experienced recently with COVID and supply chain, and you have a world where the change is the only constant, the past then depicts to future, and the ability for these organizations to make the right calls, to make the right decisions, be it short term, do I ship this package that way or that way, or do I source my material here or there or, long term decisions around, how do I establish the right network of suppliers around the world?

Fred Laluyaux:

Do I bring my manufacturing capacity closer to my consumers? All those decisions are front and central. And I think this is now compounded again with the great resignation. And that’s something relatively new that we’ve seen becoming front and central to our big clients since COVID, which is the people that are making those decisions and doing the work are just not necessarily there anymore. So you have a giant issues around loss of tribal knowledge. People don’t stay in the job long enough, whether it’s a blue collar job or a white collar job, that knowhow is actually leaving the organization. So you’ve got those macro disruptions that are combined with the trend that I’ve just explained, that are just making the life of a manager in a large CPG life sciences, manufacturing organization, very complex right now.

Patrick Daly:

So, Gartner has recognized decision intelligence as a top priority or top industry trained this year. And you yourself have been described as a thought leader on decision intelligence. But what exactly is decision intelligence and how does it help with decision making? And I’m curious, what kind of decisions does it help with? Is it both structured and unstructured decisions? Because it’s very different say to choose between a number of valid options and maybe to make a decision where you might not even be sure what the proper question is yet. So what is decision intelligence and how has it applied to those different types of decisions?

Fred Laluyaux:

Well, yeah, Gartner just came it out with a definition of decision intelligence, which can be found online. I like to simplify it by saying it’s the digitization, the automation and augmentation of decision making. So you take a process that’s not digital, which is people collaborating, manually using tools to make a decision and you digitize that process fully. So my definition, our definition, and I think it totally overlaps with the Gartner one, is that again, the digitization automation and augmentation of decision making. What’s important is in our mind is that decision intelligence goes all the way to the execution of the decisions. So it’s just a set of practice and technology. And now to help you define and identify, I should say, what the best decision is to a given problem at any point in time, but also the ability to execute it. That’s why we talk about automation. That may be where we slightly differ with Gartner.

Fred Laluyaux:

I don’t know it’s still up in the air. And to answer your question on what types of decision, what we believe is that, and I can take the analogy of self-driving cars. As an organization you’re striving to get into full self-driving mode where the decision intelligence platform can actually run autonomously the decisions that you have to make. Meaning, identify the problem, whether it’s timed or whether it’s based on an event, or decision to be made, not necessarily a problem. Could be an optimization. But identify the decision to be made. Run through, as you said earlier, the different options, select the best option based on the set of rules that have been established by the organization, and then execute that recommendation via a decision. So that process is what we called human out of the loop, where basically the system can run fully autonomously and do the work.

Fred Laluyaux:

Now that’s imagine the self-driving car, you’re on the freeway. You get your hands kind of off the wheel, and the car is making the decision. There is decision augmentation, which is another sub-process of that, where the system will deliver a recommendation to a business operator and say, Hey, Patrick, I recommend that you make that decision. I’ve looked at all the options, but I need your eyes on it, right? Because maybe the impact of that decision is too high or touching a critical customer, whatever it is. So now you, as a business operator, you receive that message from the decision intelligence platform and you say, okay, I accept that recommendation. I reject it. Or maybe I’m modified.

Fred Laluyaux:

And there is also another mode, which is you’re leveraging the capabilities of the decision intelligence platform to make a manual decision. Maybe it’s something that has not been normalized yet. Not yet being defined. There is a crisis you need to react the numbers that are flowing into the system don’t make sense. And just think about it. You have to put your hands back on the wheel and make the decision yourself. And we believe that decision intelligence, while everybody tries to get toward full automation, for obvious reasons, we’ll have organization having to balance for a same decision set between manual decisions, the human in the loop, augmented decision, which is human on the loop, and then a fully automated, human out of the loop.

Patrick Daly:

And will the system or can system or the AI, I guess we’re talking about artificial intelligence here, will it learn from the input it’s getting from people who are benefiting from the system to take augmented decisions? Will the system then learn from them and be able to do better itself?

Fred Laluyaux:

So this is one of the most critical point. We’re talking about decision automation and augmentation effectively to accelerate the decision cycles, to make faster decision closer to the point of in time, closer to real time, 24/7. This is a huge benefit. It basically unclogs the decision making funnels. But the point you’re calling out is the most important, in my opinion, is the ability of the system to actually build a permanent memory of all the decisions that are being made on a given topic. So the example I was giving you earlier, the system comes to you and say, Patrick, I recommend that you accept my recommendation. You’ll say, yes, you’ll say no. Sometimes you’ll be right to approve. Sometimes you’ll be right to reject, but you’ll only know that over time. So that permanent memory is what allows the recommendation, the quality of the recommendations to become better over time.

Fred Laluyaux:

And it’s a story of Google. They have more data. They can make better recommendations when you do a search. It’s a story of Tesla, they have more data because of their feet of cars, so they can make better self-driving cars, autonomous driving system. And this is the story of Amazon that gets more data and leverage computer system, not just AI, by the way, there is heuristics. There is a bunch of different technologies to make better dynamic pricing decisions on time or ordering decision on time. So the vision that we’ve had from day one is to enable non-digital native organizations to actually perform like a digital native. And for that, we had to find a way to start building that data set that will allow them to get the efficiencies that the digital natives have, thanks to their immense data set.

Patrick Daly:

And those companies who have been successful in adopting decision intelligence, what kind of business benefits is that bringing to them? And is there going to be an opening up in the future of a gap between those companies that get on this and those companies that get left behind?

Fred Laluyaux:

There’s no doubt in my mind, I’ll answer the second part of your question first. There absolutely no doubt in my mind that the data is the new feel for intelligence. And unless you start building that decision data and you start recapturing digitally that tribal knowledge and augmenting it, you’re going to be left behind. I was talking with not yet a client, but a prospective client in Europe the other day. And we were like, look a few years back being able to deliver the grocery to our clients at their doorstep within two hours was unbelievably hard to achieve, yet we’ve done it. And now we’re moving to every 15 minutes. We have to be able to deliver within 15 minutes. I mean the exit or the level of customizations of the products that you’re buying and all of that, all those factors are not going away.

Fred Laluyaux:

There is a trend, that’s basically a pool for speed for accuracy. And you can only achieve that if you leverage fully those digital systems. So the benefits, to answer the first part of your question, the business benefits, the first thing that’s interesting to me is, well, it depends on the decisions that you’re digitizing. One of our largest clients now is digitized 22 decision areas within their organization on the global basis, on track to go with 100 decision areas this year. And the system will have delivered about a million recommendations, which is a lot, if you think about it. And the benefits depends on the decisions. So it can have an impact on reduction of canceled orders, because you cannot tell your customers exactly when the product’s going to be delivered. We’ve seen that as a benefit, we’ve seen increasing accuracy in forecasting. We’ve seen clients optimizing their procurement processes.

Fred Laluyaux:

We’ve seen a lot of benefits in terms of logistics and transportation, trying to optimize that process with benefits on cost, on service level, but also on sustainability. Because, you’re now asking organization to measure their carbon footprint, but it’s very difficult to go from measuring to optimizing, because that optimization is going to be an arbitrage, several hundreds of times a day, between shipping the goods that way or that way with different impact on carbon. It’s very difficult for humans to actually do that. Computers can do that very well. It’s basically think about decision intelligence as a intelligence layer that sits on top of your planning and your transactional tools. And that does the work, the decision making that humans are currently making. So anything that you would expect from a good decision, you can actually get it pretty much better and faster and more accurately with that memory work that I described earlier. And I think the benefits are very real.

Speaker 3:

93.9 Dublin South FM.

Patrick Daly:

This layer sits above my existing systems, whether ERP or WMS or LMS or TMS or whatever I have. And therefore it’s taking the data from those systems. So how does your decision platform work? And for people who might be listening to this perspective, future users, what does it take to actually implement this? So how are you enabling the adoption of the decision?

Fred Laluyaux:

Well, so going back to your very first question about my background and my friend Sheik Mansour, CTO’s background, if the goal that we had with the digitization of decisions, we quickly identified that we had to address four distinct areas and usually four areas of software that are not really embedded in the same technology platform. So the first thing, and you just touch on that, is if you got to digitize decisions, you got to bring 100% of the information that you need to make the decisions into a normalized data model. Humans are really good for making a decision with missing data. That’s where you experience and tribal knowledge. I don’t have all the data, but I know that this is the right call. Computers are not really good at that. They need all the information. So the first requisite to enable decision intelligence is to have a normalized model.

Fred Laluyaux:

We call it a cognitive data layer that pools data from your transactional system. You mentioned WMS CMSs, ERP, planning tools, billions of records have to be pulled in and augmented into that cognitive data layer. And that’s the first set of patents and technology that breakthrough that we’ve built because we can deploy that technology in the worst, most complex environments with dozens of different systems. And we basically crack that nut. I won’t go into the details of how we’ve done it, but it’s still similar than to how Google crawl the internet, crawling those systems. And we’re able to do it without impacting the performance of the transactional system. So now you think about regardless of the number of data sources, you plug this and within a matter of days, this normalized data model gets populated and it’s refreshed in real time. So populating the first thing, but then you have to refresh it in real time. Or as real time, as your transactional systems are being updated.

Fred Laluyaux:

That’s the first thing. So that’s the beta. Then you have to embed in that platform, all the intelligence capabilities that you need. You mentioned AI, but not only, it’s projections, prediction, optimization, allocation rules, all of that as to be available natively in the platform. So that’s data intelligence, then it’s automation. The decision is a complex flow. I identify the problem. I deploy the logic. I may need to validate, get some more data, go to somebody based on rules. All that orchestration has to be embedded in the platform all the way down to writing back into the transactional system. So pull the data, deploy the logic, interact with the users and write back into the transactional system. So data intelligence, automation, and the last part is engagement. You are delegating decision making power to a system that’s guided by humans, but that can actually work autonomously.

Fred Laluyaux:

So it’s fundamentally important that you build the trust, and that between the business and the system. And that comes through a series of capabilities from natural language processing to lineage, so that you always know what data and what logic is being applied to a recommendation or a decision to the ability to track the performance of those recommendation and decisions over time so you know exactly what value you’re generating. So we had to build a platform that brought all those different practices together. And of course it has to scale, because you’re now talking about a system that’s real time and always on. That thinking, that’s learning, that’s autonomous and you’re doing it at the scale of the largest companies in the world in CPG and oil and gas in manufacturing. So it’s quite complex. It took us a few years to build it, but I think it also took many, many years of experience in building enterprise software and technology and companies and connecting with the right people to be able to pull it together, not to mention the right amount of capital.

Patrick Daly:

What are some examples of how it’s being deployed, say in supply chain ecosystems?

Fred Laluyaux:

So the examples, let’s start with, if I think about demand forecasting, trying to get more accurate understanding of demand forecast. It’s dynamic, safety, stock, optimizing your safety stock in real time. Trying to minimize the amount of working capital that’s stuck in your safety stocks by looking at the supply chain end to end in real time and making the right arbitrage. In procurement, the ability to optimize when you place a PO to whom and do all the controls in real time. In logistics, it’s optimizing the routes. And I think what’s very interesting. I could go on and on. I mean, there’s pretty much every type of decisions that you’re making in supply chain gets impacted by this practice. But maybe what I call out is a couple of things. Because, those are things that we’ve just learned with our clients over the last few years. When you are thinking, just to your question, what kind of decisions are impacted?

Fred Laluyaux:

We naturally followed the path of what are the decisions that are currently being made, and I just named a few, and how can we deploy our technology to actually make those decisions better, faster, and more accurate fashion. What we’ve realized, since the platform became open and our clients took it and started digitizing all other stuff, is that this technology and the approach allows you to actually make decisions that are currently not being made. And that’s the most interesting part. And I think that’s where you’re creating value. For example, we partnered with WPP in the UK to connect media and promotion planning with supply chain, think about a large multinational CPG company. On one end, they’re working with a partner like WPP to run promotion campaigns and media campaigns. And on the other end, there’s a reality of a very dynamic supply chain. So the company might actually be promoting a product that is going to be running out of stock.

Fred Laluyaux:

Wouldn’t be available on the shelves and those things, Amazon is able to do it because it’s one integrated system, but connecting that dot to the supply chain dot is something that’s not usually done. Or connecting demand forecast to procurement. There is a signal happening that your forecast might need increase or decrease. Do I optimize immediately in real time, the type of orders that I need to place and where to, and by whom. So, I could go on and on. But it’s really interesting that think about your traditional supply chain decisions that decision intelligence can apply to it. But really it’s connecting now dots that would not otherwise be connected to help prevent the issues before they happen. A lot of supply chain work is around planning and controlling, but the execution, the intro, the execution where the quality of the plan disrupted by the reality of your operation, a truck doesn’t show up, blah, blah, blah. All of that is where the SNOE part, which is one of the areas that decision intelligence can really help with.

Patrick Daly:

And with companies deploying this type of capability over the coming years, what do you think the supply chains of the best or the best supply chains will look like in five or 10 years time? How will they be different from today?

Fred Laluyaux:

Well, there’s two different aspect, there’s a macro trend, as I mentioned earlier, around supply chain that has not nothing to do with decision intelligence that you do, how you build your network and so on and so forth. But I think the supply chain operations part, right? Writing the supply chain itself. I don’t know if my facts are a hundred percent right. But I was talking with a partner the other day, was telling me that… And this may not be hundred percent accurate. Just want to call it out. But directionally, I think it’s right. Amazon had hundred and plus people in the UK running dynamic pricing 10 years back, and now it’s down to a handful. And that handful of people is sitting behind systems and making the arbitrage decisions. But the rest is being digitalized or digitalization. I believe that supply chain operations are going to look like that, meaning that you have a lot less people doing the work themselves, but a lot more people sitting behind systems like Aera and driving how this amplification mechanism, because that’s ultimately what it is.

Fred Laluyaux:

It’s a decision amplification mechanism is actually being deployed. It doesn’t mean that everybody will have to be a data scientist. You don’t need to be a data scientist. You need to have some data science knowledge, but I think you’ll see a new role. That’s actually starting to emerge called a decision analyst, right? People who are going around the organization, looking at how decisions are being made, which by the way, is something very interesting. Again, learning as a pioneer organization and working with pioneer clients, very few companies actually truly understands how decisions are being made. They think at a high level, they understand it. So we start digitizing decisions. And then when it hits really the field, so to speak, the distribution center or the manufacturing plan, people say, well, that’s not exactly how we decide.

Fred Laluyaux:

Now, you’re smiling because yeah, we’ve kind of always known that, but now it’s becoming proven. So I think you’ll see a lot more effort in truly understanding how decisions are being made. The emergence of the decision analyst is I think, meant to happen. And as I said, you’ll see supply chain operations, moving from operators making every decisions, to operators guiding a system like Aera, to help scale, automate and augment the decision making process.

Patrick Daly:

And perhaps set the parameters for decision making and criteria and stuff like that. So it’s a very different type of role, different type of person. If you like, who will be employed in supply chain operations?

Fred Laluyaux:

I believe so.

Patrick Daly:

Yeah. So there’s a lot in this and I think we may need to get back on here for a second edition. As we move on to the last few minutes of this, we might just change tac slightly. And maybe I’ll just ask you a few questions about yourself. So when you’re not thinking about all of this fascinating stuff, what kind of things do you like to do with your spare time?

Fred Laluyaux:

What do I like to do with my spare time?

Patrick Daly:

Do you have any spare time?

Fred Laluyaux:

No, no, no. That’s not true. I mean, I spent time with my family first and foremost, and as much time as I can. I just went back from a vacation where I had an opportunity to do some scuba diving and with my son and my daughters, it was a lot of fun. So enjoying vacations whenever I can, I read a lot. I love music, trying to enjoy social life as much as I can. And keeping healthy is important. But I like to read. Yeah, you’ll find me on a Sunday afternoon reading a book or listening to music, that’s probably one of my happy places.

Patrick Daly:

Do you have any reading recommendations for our listeners? Are you reading anything inspirational at the moment?

Fred Laluyaux:

Today’s Monday, right? So literally yesterday I started reading a book that my kids got me for, I don’t know, father’s day or something on bees. And I didn’t know much about bees, but I spent an hour reading that, and the intelligence of a swam of bee as it compared to the human brain and how they make decisions. Now we’re looking back to our first topic. So I don’t know, I’m going to spend more time learning about bees now because I think there’s something really fascinating. I know at a high level, why they’re important, I just didn’t really know how much intelligence goes into how they’re operating. So I don’t know, that’s going to be my next topic for the next couple of days.

Patrick Daly:

So perhaps some beehives appearing around your house sometime in the future.

Fred Laluyaux:

Well, little secret, already have one. We have a little place in Northern California and this is why the kids got me to got the books on the bees, is because I started by getting the beehive. So yes, Fred the beekeeper is my new nickname now. I’m not sure I’m really reserving that title. We’ll see they stay in the hive for more than two weeks.

Patrick Daly:

Excellent. Excellent. And then maybe to wrap up then, how can people find out more about Aera? Where can they locate you and get in touch with you?

Fred Laluyaux:

Sure. Well, I mean, the easy Aera place is aeratechnology.com. That’s our website. Or reach out to me on LinkedIn and will be happy to engage. Of course.

Patrick Daly:

Excellent Fred. Many, many, thanks. It’s been fascinating and a pleasure to talk to you today. And I think we probably will need to get you back, because I think there’s a lot more in this conversation. So I wish you the very best for the future. Both personally and professionally.

Fred Laluyaux:

Thank you Patrick. It was a real pleasure talking with you this morning. Thank you.

Patrick Daly:

Thanks also to our listeners for tuning in again. So do stay well and keep safe, until next time.