#300 Fred Laluyaux: How Decision Intelligence & AI Agents Are Redefining Enterprise Operations

How are Decision Intelligence and AI agents reshaping enterprise operations today? In this episode of Eye on AI, host Craig Smith sits down with Fred Laluyaux, CEO of Aera Technology, to unpack how organizations move from dashboards and ad hoc workflows to a system that senses, decides, and acts. AI is not just about chatbots. At the heart of this transformation is decision intelligence: connecting data, analytics, AI, and automation to optimize decisions across the enterprise. Fred explains why this is becoming the operating backbone of the modern enterprise and how it accelerates the shift toward autonomous, self-driving businesses. We look at how to build a decision intelligence stack end to end, how AI agents collaborate with people, and how to stand up a control room that monitors decisions across supply chain, finance, and customer operations. Learn how leading companies model decisions, govern them safely, and measure impact with clear metrics that matter, including service level, cost to serve, cash flow, inventory turns, and time to resolution. Stay Updated: Craig Smith on X: https://x.com/craigss Eye on A. I. on X: https://x.com/EyeOn_AI

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  To remain competitive in this growing digital world, companies would have to think about their decisions and the impact of their decisions across the value chain, breaking the silos, not just data, but knowledge and organizational silos. When you look at the volume of decisions coming in and you look at the constraints that are basically driven by this human operated model, our vision was from day one to move from people making and executing decisions supported by machines, by data, by software, to a world where machines would be able to make and execute decisions guided by people.

• 0:35Craig S. Smith?

  I usually start by having you introduce yourself to listeners and explain what Era Technologies is. And then I wanted to talk about decision intelligence broadly and then how it relates to Era. But why don't you start by introducing yourself?

• 0:56Speaker 00

  Okay, so first of all, thanks for having me today, Greg. My name is Fred Laduio. I'm the co-founder and CEO of Aera Technology. Been working on this adventure for eight years now. Prior to that, you know, I spent my entire career in the world of helping enterprise manage complexities with software analytics. So my career took me from Europe to the US, starting with data modeling, business intelligence, ERP, AI. Prior to building Era, I built another company, the space called Anaplan. Prior to that, I was with SAP. So basically, spent the last close to 30 years really

• 1:41Speaker 00

  working on performance management with data and analytics.

• 1:47Craig S. Smith?

  Okay, and this is all built around this idea of decision intelligence, is that right?

• 1:55Speaker 00

  It is correct, yeah.

• 1:56Craig S. Smith?

  And you've been pioneering decision intelligence for years.

• 2:01Craig S. Smith?

  Uh, long before most people were talking about it. Can you explain what decision intelligence is? And why it's becoming important?

• 2:10Speaker 00

  Well, I was pioneer decision intelligence before it was called decision intelligence. When we launched Era in 2017, June 19, 2017, I'll never forget that day. We said, welcome to the self-driving enterprise. Why did we do that? It goes back to 2010 when I was at ACP at the time and I wrote a paper around uh, my vision, which was quite simple, related to the fact that or driven by the fact that uh, we believe that the digitization of our economy was going to drive three fundamental things. The first one is an acceleration, a continuous acceleration of business cycles. And if you go back, you know, a decade, two decades ago, and you looked at the planning and the execution cycles,

• 2:53Speaker 00

  you were counting in month, you were counting in quarters. You're now looking at real time. So that continuous acceleration has been proven, uh, you know, a valid assumption to start the company. The second one was back to the digitization, driving now a consumerization of our economy. Uh, and, and the result of that was decisions that would have to be made at a much finer grain. We used to plan and think with segments, with aggregates, with categories. And now we're thinking and executing down to the store, the SKU level, the consumer level. Okay. And the third element was vision that to remain competitive in this growing digital world, companies would have to think about their decisions and the impact of these decisions across the value chain,

• 3:38Speaker 00

  breaking the silos, not just data, but knowledge and organizational silos. So for me, again, that goes back 15 years ago. The combination of acceleration of business cycle, consumerization and complexification was going to drive an explosion in the volume of decisions that companies would have to make to remain competitive. And then you look at the way companies make decisions and the way they're structured. There, it's a large pyramid of 10, 11, 12 levels of hierarchy, right? Sitting on Top of a bedrock of transactional system. And people within this big pyramid are scrambling with.

• 4:18Speaker 00

  Uh, you know, tools, bespoke software, data analytics, workflow systems. To actually make the decisions. So when you look at the volume of decisions coming in. And you look at the constraint that are basically led by or driven by this this human operated model. Our vision was from day one to move from people making and executing decisions supported. By machines, by data, by software, to a world where machines would be able to make and execute decisions. Guided by people. So that's that's really foundation, the the foundation of of era. This is why we started. We said, okay, we got to build a technology stack from the ground up to enable that process.

• 5:02Speaker 00

  People called us nuts and crazy and a bunch of funky names back then. And you fast forward eight years and and decision intelligence, which is now defined as the digitization. The automation and augmentation of decision making. Is a category that's being, you know, it's an AI category. And the the validation of the the growth in the market and and and the interest in that category is the fact that. You know, in December or maybe January twenty twenty six, December twenty five, January twenty six. Gartner is going to release the first magic quadrant for decision intelligence platforms, which is squarely what what we've. Helped define as a category for the last eight years.

• 5:47Craig S. Smith?

  But how does decision intelligence differ from uh? The agentic AI that's spreading so quickly?

• 5:57Speaker 00

  Um, if you, well. The station intelligence leverages agentic AI and other AI and analytics capabilities. With the the goal of digitizing specific decisions for the enterprise. So. The agentic is a tool, but the category that supports the full end to end process of. Digitizing decision is decision intelligence. So within decision intelligence, we're leveraging L L Ms, we're leveraging. All different kinds of capability, but it's not relying exclusively on that. If you want to make a decision intelligence, a a reality in your enterprise, you need to resolve four fundamental problems. The first one is the data.

• 6:41Speaker 00

  If you are going to make decisions digitally, you need one hundred percent of the information that is required for these decisions to be made to be available in a normalized data model. And that's a lot to unpack and unpack, but that's the first block. This can leverage LLM and agentic, especially when it comes to pulling in unstructured data, which is new. Back in the early days of DI, there was no real way to leverage unstructured data. Agentic technology is allowing us to do so. The second is the intelligence. And again, there's a lot to unpack in the intelligence. First is data, second is intelligence.

• 7:24Speaker 00

  Intelligence is the ability to reason. LLMs can be fantastic for that, but not always needed. You can have deterministic processes that allow some reasoning, but it's also the ability to do all the math that is required when you make a decision. You need to project, you need to predict, you need to allocate, you need to select, you need to calculate a bunch of stuff, which is not what LLMs do. So the intelligence is a combination of the reasoning and the calculation. And those are two different technologies, right? The third is the automation. When you think about decision intelligence, you don't have just a system that tells you what to do. It would not solve the problem I've explained earlier of the volume of decision. You need a system that executes the decisions as well.

• 8:09Speaker 00

  So I'll give you an example. You have a, the system will predict an inventory stockout, meaning you will be running out of stock and some customers are not going to get what they ordered. Well, then you have several options, right? Conceptually, you can make more products, you can transfer products, you can expedite, you can do a series of things. The system will calculate the different options, will select the one that's the most relevant for that specific combination of product, of customers, of in time, based on expected business outcome. I want my customer service level to go up. I want to make sure I get the revenue. So you select all of that. But then, once you make that decision, the system needs to execute your decision.

• 8:50Speaker 00

  It will go and create a stock transfer order, for example, in SAP. It will go and update your planning parameters. So it's the data, it's the reasoning, it's the execution, which is not an LLM process at all. And the last part, LLM play a big role of that as well, but not only, in national language understanding and generation, is the engagement. How do I engage with the users? How do I learn from the users? Last year, our technology enabled our customers across different industries to execute about twenty five million decisions, right? So a lot of that is automated. The system does the process I've just described on its own. But in many cases, you have exceptions. You need to go in front of a human.

• 9:36Speaker 00

  Then you will have your inbox, your era inbox, and you'd say, okay, a message for Greg from era saying, I've done the work, I've done the analysis, I identified a problem, I should say. I've done the work, done the analysis, I make that recommendation, but my confidence score for that recommendation is maybe below eighty percent. I need your input. So then you open the message and the action you take is, you of course need to understand the logic, the data, the source, you build the trust, but then you say, accept, reject, modify. So that action needs to happen with business operators who are working anywhere and everywhere. Could be in a warehouse, could be loading trucks. So that's where you have to have a level of engagement that's super smart.

• 10:18Speaker 00

  And the software will then capture your decision and execute it. So data, intelligence, automation, which is very important, and an engagement. And across those different tiers, you will leverage multiple flavors of AI, including Agentic, which is changing the game for us right now.

• 10:37Craig S. Smith?

  Yeah, yeah. So decision intelligence, who are era's technology stack. The Agentic part is is kind of at the bottom of the stack. There is a lot more going on in in context analysis and and in that sort of thing. Are there other systems that are doing this kind of thing?

• 11:03Speaker 00

  You mean systems inside our technology or other vendors out there?

• 11:07Craig S. Smith?

  Other out there in the world because I haven't heard it described this way.

• 11:12Speaker 00

  Um, the, as you have in the market right now, you have two, like always when a, when a market is forming. So first of all, I would say that in the last year, what we've seen is the word decision became ubiquitous. Everybody is talking about decision. I can guarantee you that even three years ago, nobody was talking about decision. They were talking about AI on one end, they were talking about data systems and data lakes and analytics and. But now everybody's converging toward decision. We talk about process intelligence, talk all this good stuff. Everybody's converging toward decision intelligence, decision automation. Because that's the future, that's what's uh obvious right now for large enterprises.

• 11:52Speaker 00

  So if I break down the market for you, I would say a lot of the systems of differentiation, which is if you think about the enterprise stack, if I trivialize it, you've got your systems, system of records, your ERPs, your CRM, you get your systems of differentiation. This is the work, this is what tech, this is the technology that allows people to manually process data, collaborate to make those decisions. Could be a plan, could be a, a decision to re-point sales order, to allocate inventory uh to uh this. I'm talking a lot about operations, but it goes, it goes beyond that. And then you get your system of analytics, really basically fundamentally allowing the information to be distributed across the organization.

• 12:34Speaker 00

  Everybody who's in the category of system of differentiation is trying to now bring that intelligence uh to their system. So a lot of the vertical vendors, where a planning vendor, a procurement vendor, are trying to bring that level of, of service, basically of, of automation uh to their software. It, then there is another category, which is the pure play, the decision intelligence platform. And I can talk about why I think this is so important to differentiate the two. And in that category, you'll find Palantir, you'll find Era, you'll find very few others that have been able to build a platform end to end. And then you have the do it yourself, which is a big, a big part of the market today,

• 13:17Speaker 00

  especially in the early stages where large companies who have access to a stack are trying to couple things together to make DI work for them. And, and that's pretty much how the market is structured. Now, who's coming into this market? I talked about the system of differentiation, the analytics vendor are trying to get into this market, the RPA vendors are trying to go this market, the ERP vendors, everybody's trying to converge toward that new El Dorado, which is decision intelligence.

• 13:43Craig S. Smith?

  Yeah, that's fascinating. The, and what are the core components of, of decision intelligence? How, and how do they work together?

• 13:53Speaker 00

  Well, I mentioned that a bit earlier. I think for us, we break down four key components, the data, I mentioned it, right? You have to be able to crawl. So the data is bidirectional, the data is structured and unstructured, the data is a big, that's where we come from initially, right? We, we said, we got to resolve that one problem first to enable the intelligence and the automation and the engagement later. So if you just break down the data process, our technology connects to the transactional system, it actually crawls them intelligently. It can be a data lake if you have that, but we, we recommend to go straight to the source, which creates a lot of dialogue with our customers because there's a lot of questions about it. How do I crawl those systems without impacting their performance?

• 14:36Speaker 00

  Then the data, then again, that's for the structured data, a big part of what we do is work on top of, of SAP, of the, the traditional, you know, systems of record. But you also bring in external data, unstructured data, it can be crawling websites, it can be looking at PDF for claims management or emails or images. So that's data, it's all combined into an indexed and structured into our, what we call the decision data model, which is a normalized data model. Why do we call this decision data model? Because it's a model that has all the metrics that allow you to understand your business. Calculated, refreshed in real time in the in the system.

• 15:16Speaker 00

  But it also going to contain the memory of all the decisions that are made with the decision intelligence. So the connectivity between the transaction, the augmented metrics and the uh the the decision data is is quite a big piece of software. Then, as I mentioned, you have all your engines, so the capabilities you need, I need to model my data. So um it it's a multi-dimensional structure that allows me to use formulas to to cascade stuff, to project stuff. So you need to have that capability. We have a cortex, which is basically our our um machine learning, uh you need to have machine learning, yet you'll be able to do statistical forecasting, all the different flavors of uh of AI.

• 16:02Speaker 00

  You need to have an agentic uh capability with agents, agent teams, agent functions that can orchestrate a lot of that logic that can also be leveraged uh to understand and structure data. And so that's a another big component of that. So you've got your analytics, you've got your data modeling, you've got your machine learning and other AI capabilities. You have your um uh agentic capabilities and all of that gets orchestrated. Uh then you have the ability to write back and orchestrate the decisions back into the decision memory and into the transactional system. And as I mentioned before, you have the you need the ability to engage with the users.

• 16:45Speaker 00

  And the way we looked at it, and it's very important, because when people think about DI today, as the market is maturing, they think about a system that's going to generate recommendations, going to provide some intelligence that allows people to individually make decisions. Well, that's great, but it creates a lot of noise. And you know, you're just pounding the users with, hey, I'm I'm a smart system here, and I'm going to tell you all the things that you should be doing. People don't have the time for that. You need to automate the work and you need to engage with them at the right level at the right time. And one of the foundation of Era after twenty five years building software.

• 17:22Speaker 00

  I'm going to square a little bit here, but I wanted any user experience that didn't suck. And, and, you know, I've looked at my previous companies and, you know, it's all those systems are required. They introduce a level of complexity that forces people to learn how the software works and get familiar with it. And how do I, but, but we said, we're going to move from, and it sounds marketing, but it's not. It's really part of core to our vision, going from a people that people have, people have to learn how to, I'm sorry. Moving from people that use a software that they have to learn how to work with to a software that learns how you work, a software that learn how, learns how you decide.

• 18:05Speaker 00

  And that's what we've built for that. We needed to create that experience, just like Siri or Alexa. This is what Era is called Era, four letters, A E R A. It took a long time to come up with the name that you could call in any language and you talk to Era. You engage with Era in natural language. Now, when we started, we were already doing demos of that technology eight years ago. That was literally talking to Era, but it was very scripted. It worked. But I had to engage with that. Greg, I would say, Era, what is the evolution of my forecast this month? And if I didn't spell it properly or spook it properly, the system would not understand.

• 18:44Speaker 00

  The addition to L L Ms in the last two, three years has been fantastic because now I can have a conversation with my decision intelligence agent that is completely natural. I can say, uh, what is the volume of blocked inventory between Stuttgart and, uh, and Frankfurt uh, over the last month. And I'll say, break it down by product category. And then show me the ones that are impacting this customer. And I use those, this language. The L L Ms and the agents that we built in the system now allow me to have a natural language conversation, not just a using words to interact with the software and, and keeping the context and having composite questions.

• 19:27Speaker 00

  All this is super important. So that was part one, engaging as much as possible in natural language. And the second part was for us very critical is. Uh, I, I, I, what, what do you do every morning? You check your inbox. You go into your email and that's what we do. So you use Era, you can ask questions, get answered, you can run simulation. But you fundamentally start your day by going to your inbox and seeing how many messages in English or in whatever language you want, properly documented as a system prepared for you. And I don't need to train people on how to use an inbox. But I have to train people on how to use a complex software with a bunch of buttons and windows and, and, and processes. So taking care of that, taking care of the workflow dynamically is all the stuff that we had to build.

• 20:11Speaker 00

  So the components to wrap up data, intelligence, I broke it down between the engines and the agentic, the ability to orchestrate sometimes in a deterministic fashion, a simple process, the automation and the engagement. And, and this is the stack that makes DI um, a reality for some of the largest companies in the world today.

• 20:32Craig S. Smith?

  Yeah. Now you talked about uh, memory and, and a learning loop. Uh, is the memory of uh, offline, is it in uh, you know, a, a graph database or, or something that gets updated? And the learning loop, um, how do you avoid overriding previous uh, previous memories or previous information? I mean, that's catastrophic forgetting that everybody is working on.

• 21:10Speaker 00

  So that's, that's, it's a good, it's a, it's a super interesting topic. When we talk about, I described the stack and described the components, which is the technical part. But when we talk about what does Era do for you? Now, we have the question, you go and meet many and many companies all the time. Last week I was doing a great conference with Hershey's and they're deploying Era right now. What does, and for the end users, what is this decision intelligence agent called Era that you have access to? What does it do for you? First thing is it answers any question that you have in natural language. The second is it allows you to do simulation, what if planning, uh, uh, you know, collaborate. Third, is it? generates recommendation for you for your job. You do your job every day. You have to, your job is to reduce, manage inventory or reduce waste in components.

• 21:55Speaker 00

  The system provides all of that. It executes the decisions. It learns. So Era understand, Era recommends, Era executes, right? And then it learns from the decisions that you made. Couple of things. It just doesn't learn from the decision that I make. If you're a company like Unilever or She's or Dell or whoever, they have for a specific function, you'll have dozens or hundreds, sometimes thousands of operators at scale, right? So the key is that the system needs to learn not just from my decision, but from the decisions made by the entire network of people. So how do we do that?

• 22:33Speaker 00

  It's quite tricky to develop and code, but now it's deployed very easily. There is, there is a couple of components. The first one is, you got to build the data for the learning. The data for the learning is basically what is the context of each individual decision that is being made? So I could have backpedaled a little bit and if you allow me, I'll do that. One of the biggest question that I had when we, we, we, we launched Era was, it's very simple.

• 23:02Speaker 00

  I've been working for twenty five years or twenty some odd years in analytics and, and, and the message for all the tools that we built and all the companies that we created around that was, we're going to help you make better decisions. However, the dimension of decision is not captured anywhere. If you look at a system of information today of any large company, you say, where do you capture your decisions? Where are they memorized? Nowhere. Now, it is in some specific sectors. If you look at automatic trading, they do. They know that. They've done it. They're ahead. But if you look at the world of operations and, and sales and finance in an enterprise, you don't capture that. And I'm not just talking about. Strategic decisions, which is do I go into this market? Do I?

• 23:47Speaker 00

  No, talking about the decisions every day. I have that many sales order. I've got that much inventory. I've got to match this. Those are micro decisions that are very important. I need to ship my goods a certain way. I need all this stuff that is thousands of decisions that you make every day. They are not memorized. You rely on people to drive that process and the decision stays in their head. Now they follow guidelines, they follow rules, they follow everything. But the first thing we said is like, how do I digitize a decision? And our approach was to go A B testing. Like if I come to you, Craig, with a recommendation, fully documented, that you can understand. When you press accept the recommendation, now I'm digitizing your decision. Right?

• 24:32Speaker 00

  So what is the decision memory? It's the context of the recommendation. So the business context, what are the metrics that have been driving the recommendations to be generated? It could be on schedule, it could be on demand, it could be, it depends, on event. Then I capture your decision. I capture when the recommendation was made, when you made the decision to accept, maybe you rejected it. If you rejected it, the system should prompt you and say, why? Select from the list or enter a little text. Well, Fred, this is the reason why in our, in East Cleveland, at that time of the year, we have something that wasn't captured in the software. Great. Or maybe you ignore it. Why you ignore it? Or maybe you modify the recommendation.

• 25:16Speaker 00

  That needs to be captured. It's memorized, all right? And then what the system does, so now I've got the context of the decision, the expected outcome. I expect my, to generate X amount of dollars in saving. I expect to generate that much improvement in service level. I expect to retain, we have a fantastic customer called Western Governor University. It's one of the largest or the largest digital university in the United States and they'll be speaking at our conference in a, in, in, in a month. They're leveraging decision intelligence to help with student retention. So I'm not talking about logistics, I'm not talking about operations, I'm talking about leveraging data, make generic recommendations to help retain students.

• 25:58Speaker 00

  And, and the same logic, I need to have the context of the decision, the recommendation, the decision itself, and then what is the expected outcome? And as the system runs and refreshes.

• 26:12Speaker 00

  Every day, multiple times a day, we have a client who refreshes the model every fifteen minutes. The system will see and identify whether the expected outcome was reached or not, and if not, to what degree. And that's when you bring a feature that we built called auto decision learning that allows you to calculate over time, when there is enough data, when the data is rich enough, it will calculate the confidence score for future recommendation. So now you look in your inbox and you have a recommendations to rebalance inventory from a DC to another DC. It will explain the root cause, the logic, the expected outcome, and it will say, Era is confident at ninety-two percent that if you say yes, the outcome will be reached.

• 26:55Speaker 00

  And the system continually learn. It's not a numbered, maybe if you had run the same scenario three months later, maybe the confidence score would have been ninety-five percent. I don't know. So that's one part of the learning. Now, there is another dimension, which is super interesting. With Agentic, if you're leveraging agents and you're prompting the decision logic, which is we're doing now with some of our most advanced customers, this is brand new technology. It's the ability to really use Agentic to run the decision logic. Now, there is a closer loop, faster loop for the learning, which we're experimenting with,

• 27:37Speaker 00

  that will allow me to understand from, if you said no to that recommendation and you give me a verbal explanation five times in a row, I can now start detecting faster than from the data whether there is an adjustment that could be made to the logic. But that needs to be also very much supervised. The first model that I described, which is auto decision learning, doesn't require human supervision. It calculates. It's math, right? This one, well, you'll see the prompts coming up saying, "Well, Craig believes that this recommendation needs to be improved, the logic needs to be improved, or the threshold needs to be changed." Based on his input, I can identify a pattern

• 28:18Speaker 00

  and then I can go manually and adjust or not automatically adjust the decision logic, the prompts. That's new. That's starting. We got one customer fully live on that now. And that's going to create a decision learning process that will be faster, but will require more adult supervision, so to speak.

• 28:40Craig S. Smith?

  Yeah. I'm interested in the mechanism of this memory and and learning loop. I mean, is it, are we talking about updating weights in a neural network? Are we talking about updating a database that that then is referred to

• 29:04Craig S. Smith?

  in future transactions or decisions?

• 29:10Craig S. Smith?

  And yeah, how does how does that work practically in the system?

• 29:17Speaker 00

  So, as I mentioned, there is the there is the current way and there is the the new way. In the current way, you basically built a data set that is getting enriched over time. And we're using machine learning to derive the likelihood that, so it learns from all the decision made. It does it does know the context and then it it flags over time whether the decisions that were. So it doesn't learn right away. Let's me start with that. The the feature is out of box, but it doesn't learn right away. And it's always a question last year with last week with with Hershey at this conference. They were saying, we need about 10,000 recommendations to start learning.

• 30:02Speaker 00

  And I said, you know, it might be 10,000, it might be more, it might be less. We don't know. You have to build a data set. So what the system does, to keep it simple, it it flags whether a decision was has delivered the expected outcome or not. So that's the first process. You you record all of them and over time, could be a minute, could be a week, could be a month. It will say, check, this was good. This this recommendation in this context delivered the expected outcome. And as you start building and then those others will say, no, it's not. And then he will look at the next recommendation that's coming. And with machine learning, it will actually identify, it will calculate the likelihood

• 30:46Speaker 00

  based on historical data that the next one, it's a predictive model, is actually going to generate the expected outcome. And that that is getting refined over time. That confidence scoring is attached to every recommendation, but it's, as I said, is something that evolves over time. So that's what it does. The other way, go ahead.

• 31:06Craig S. Smith?

  It sounds.

• 31:09Craig S. Smith?

  So it's minimizing the error prediction, is that right?

• 31:13Speaker 00

  Yes.

• 31:14Craig S. Smith?

  And is that through Bayesian inference where you're looking at

• 31:23Craig S. Smith?

  probability distributions? Or is this looking as in traditional neural networks?

• 31:34Craig S. Smith?

  The error between a prediction and and some baseline and then correcting the weights?

• 31:44Speaker 00

  It's the latter.

• 31:46Craig S. Smith?

  Okay, using backpropagation or?

• 31:48Speaker 00

  Yeah.

• 31:49Craig S. Smith?

  So then the question is, in that context, there's this problem of catastrophic forgetting that as you update weights, you lose some of the earlier intelligence. How do you guys manage that?

• 32:06Speaker 00

  Well, the process is supervised, right? So you have in the control room, the system runs on its own, but you have in the control room, the ability to test, the ability to monitor all the different parameters that are driving the error learning capability. So when we talk about

• 32:28Speaker 00

  guided by people, this is it, right? So you have decision architects, you have decision analysts who are monitoring this, who are testing the assumptions, who are cleaning, who are updating the models over time. And this is a fundamental part of decision intelligence, which is it's autonomous, but it's supervised. That's why we keep talking about guided by people. You have a control room, you have people looking at the performance of the different algorithms, you have looking at the sorry, the word escapes me right now, not the sanity, but the quality of the data set and you have work and we're building the tools to help over time improve that process.

• 33:11Craig S. Smith?

  And that those people are sitting at the enterprise using not at at era on the

• 33:20Speaker 00

  Ah, very good point. There, there, there are. Every company that deploys decision intelligence at scale needs to have a center of excellence. They have literally physical rooms sometimes where you are monitoring this. So our platform is a SaaS self-service. We provide all the tools, all the capabilities, the the analytics, the, the, the features that allow you to train sometimes your own algorithms or to bring your own. A lot of the work we do is with very mature companies that have already built some kind of capabilities. They have their own a prediction engines. They have all this

• 33:58Speaker 00

  machine learning models already built, but they need to bring it into the platform. They know how it works. They need to bring it into the platform to operationalize it. You can also use some bespoke capabilities that we have already or services that we already built in the platform. When it comes to Agentic, for example, we're, we're building a library of agents that will be uh, that are available in the platform for people to do either what we call business agents that can help, I don't know, allocate stuff or optimize an inventory. And you have also the functional agents that will be reading emails or understanding pictures. So

• 34:38Speaker 00

  all of that library is available, but the assemblage and the running of those capability, the configuration of that is done by the customers on their own, which is absolutely critical. Because we're talking here, Craig, about digitizing the work that is running your business. And you, the first thing you need is an open platform and, and you need to own that IP. So the transformation that we're, we're discussing with our, all our clients all the time is yes, this is an efficiency play, but it doesn't mean it turns off. You have to optimize and there is more to it, right? So. When you start digitizing decisions,

• 35:19Speaker 00

  you enable new types of capabilities in how you think about your enterprise. Today, you have a planning function, you have a logistic function, you have a procurement function, you have all those functions. Why is it structured like this? Because you need a high level of specialization of the individuals that are beyond those functions. They need to know how to make the decisions, they need to how to use the software. So when you digitize that process, you can now connect dots that are not connectable today. I can look at building a skill. A skill is a competency of Era. It's something that Era Learn has how to do that connects my real-time digital media activity or planning

• 36:04Speaker 00

  with my supply chain levels, which was unconceivable just a few years back. So that work and the question that comes next is like, but how do I prioritize my decisions? Now, the way a company works, you build the plan and then you have hundreds of people, thousands of people that are doing micro decisions. They're not aligned. We try to align them through processes and tools, but they're not aligned there. There's a lot of micro conflicts between those decisions all the time. But it's okay. At the end of the day, you, at the end of the month, you'll check, did we hit the plan? Did we hit our targets? And what could we have done better? Eventually, you do this work.

• 36:43Speaker 00

  Now imagine that I'm able to orchestrate my decision out of fine grain, right? I can say, well, if I do this, the impact on that is different. There's a negative impact on this other decision. And that is the area that we're exploring right now. And this is why you need a decision intelligence platform, because decision intelligence is not just generating recommendation. As I mentioned, it's executing, it's building a memory, and it's the orchestration of the decisions across your value chain. That's incredibly complicated. And on a scale of one to 10 in terms of maturity, where we are as a vendor, but where our customers are, we're on two. The next five years, we're going to move to seven or eight.

• 37:28Speaker 00

  This is really fundamental. Think about the, think about the analogy of the self-driving car, which I use quite a bit. When we launched the company, we said, welcome to the self-driving enterprise. And it says, this is a self-driving car, and this is going to be a self-driving enterprise. At the time, we had a picture of a self-driving car because they were not self-driving yet. Then it evolved. Now in San Francisco, where I'm talking to you from today, we have Waymos everywhere. Fantastic. Yeah. So that, what's Waymo? It's a regular car with self-driving capabilities. But you still have a steering wheel, you can still drive the car if you need to. Now we have the Zooks for the last month or so, which are, you removed the

• 38:10Speaker 00

  technology that allow a human to actually drive the car. This is gone. But it doesn't mean that the driving is not happening, it's happening digitally. It also means that the, the, the decisions around driving, so there's the operational driving, but the decisions about where do I go at, are happening in a network. There are people behind some screens in a control room at Zooks, I guess, who are looking at the entire network and optimizing each individual car. Same thing is happening for decisions. We're digitizing every single decisions, but we're creating the memory and the intelligence that allows you to optimize your decisions in a network. And as, that's the part where I'm telling, we are on a two out of 10.

• 38:53Speaker 00

  We're, we're, we're dabbling into the, we're, you have to get the first fundamentals in place. But this is why I like the name Decision Intelligence. It's the intelligence that you can deploy now on top of the decisions that you make every day. And this is incredibly rich. And this is going to change the world because it's going to allow you to remove the human-based inefficiencies from, from our supply chain, from our production, from our manufacturing. You're going to be able to make the right amount. Deliver it in the right place to the right customer in real time. You cut waste. Cutting waste is the number one driver for our customers to deploy the AI today.

• 39:35Speaker 00

  And it's raw material that's getting destroyed. It's components that are being sold into at a loss because because today it's it's so inefficient. Someone says, we're going to sell 10 million of this, uh, units of that. And then this big machine starts, you know, uh, running. And every step of the way, there are inefficiencies. Because it's physically impossible for humans to optimize across that network. But it is possible for machines to do it.

• 40:06Craig S. Smith?

  Yeah, wow. Uh, well, how is adoption? Uh, I mean, enterprises have been slow to adopt.

• 40:14Speaker 00

  Yes.

• 40:15Craig S. Smith?

  Those basic agents. I I can imagine that that there's some challenge in uh in getting them to implement decision intelligence. Um, and actually, let me add a question on that.

• 40:27Speaker 00

  Sure.

• 40:27Craig S. Smith?

  There's that question, but also this control room that you described.

• 40:32Speaker 00

  Uh huh.

• 40:32Craig S. Smith?

  Uh, where people are monitoring decisions and optimizing.

• 40:37Craig S. Smith?

  That sounds like a fairly high level. I mean, it's in in the self-driving enterprise that you envision, i- is that the C-suite ultimately?

• 40:50Speaker 00

  Uh, help me understand what you mean by the C-suite.

• 40:52Craig S. Smith?

  Is it the top, uh?

• 40:54Speaker 00

  Oh.

• 40:55Craig S. Smith?

  You know, the CEO, CTO, CIO.

• 40:58Speaker 00

  Um, okay, super interesting. So, um, maybe I uh I'll address the the adoption and then I'll address that. Okay, so adoption. Um, well, tell me about it, Slow. I mean, I've been at it for eight years. Uh, and I wish it moved faster, but it's the good news is it's happening now. As I mentioned to you earlier, the fact that Gartner is positioning now a magic quadrant for decision intelligence platform. And the conversation in that magic quadrant is around all the questions we we've we've discussed today. How does it work? What algorithms, what machine learning processes? How do you make you you model the decisions? How do you run them? How do you execute them? How do you learn from them? So it's very technical. How do you drive adoption?

• 41:43Speaker 00

  So adoption, uh, there's two level. The first one is the a customer we decided to roll out decision intelligence. How do they get the users to adopt it?

• 41:54Speaker 00

  The reality of most companies is the operators of the business are completely overwhelmed by the volume, the complexity of decisions. So when you provide this agent, that that it's like, you know, it's it's it's your personal decision intelligence agent. Again, I go back to Siri Alexa. The system works with you. It removes a majority of the dull repetitive work that you have to do. Imagine that you have a question, you get an answer, and it's not copilot thing. It's not just for your personal productivity. It's a question around the enterprise data, and you get the answer in real time, and you can validate the truth, and you can. So it it it provides you oxygen. We have a great client in Canada who said,

• 42:39Speaker 00

  a wrap provides oxygens to the users. So but there is always going to be a concern around, am I training a tool that's going to replace me? And and that's a fair point. The reality is, if you don't embrace this approach, you're going to be replaced. I do believe that this technology is creating a fair amount of work for a fair amount of people, because you can unlock, if it's just looking at individual, yes, it's going to replace a lot of the work you're doing individually, but it's creating so many opportunities that we discussed around the network, that you have to think about things completely differently. And there's a massive amount of of of of of work that can be done to deploy this kind of technology.

• 43:20Speaker 00

  Companies are moving toward DI today. Look, I've got a personal view on this, but it's moving very fast. I was wrong in when would the inflection point be, but I'm always wrong on that. My my my job as as a founder and CEO is to figure out, well, the puck's going to be, and make sure that by where the puck is, my my technology is ready at that point in time. It took twice as long as I thought. Thought it would take four years. Well, COVID happened, but I thought it would take four years. When we launched, it take it took eight years. That moment is now. Why? Because of the pressure that I've discussed earlier, which is the volume, the complexity and the speed, that's not going away. But there is another element, I think, where agentic AI and ChatGPT and

• 44:04Speaker 00

  I have demonstrated that this intelligence is real. Yeah. Now, we can debate whether it's real or not, but the fact is, I use it every day at home. I do my homework with it. I find my whatever. I use it. And I don't have to be proven that that kind of intelligence is a reality. The next question is, how do I deploy that kind of intelligence in my enterprise? But but CEOs and everybody says, we have to use this somewhere in the company. So that helped us a lot. I think without the explosion of of generative AI, we would not have been able to feel the traction we're seeing in the market.

• 44:47Speaker 00

  So the demand for it, the demand for a decision intelligence, the fundamentals are screaming louder than ever. Acceleration, complexification, a granularization of decision. But what's enabling companies to say, let's go, is that push from the top, from the CEO, from the board, like, how can I do this at home and not in the office? And that's where DI comes in and said, this is our job and our number one discussion with CEOs today is, DI is a very real way or opportunity for you to operate your business, leveraging this level of artificial intelligence.

• 45:28Speaker 00

  And the problem was operating. I need to operate. I need to let go and let the system operate. Now, operating and energetic is fantastic to read documents to, you know, if I'm doing a call center or service desk or I'm in the legal department and I need this LM to read all my documents and extract on letters. It worked beautifully. On numbers, it did not. And what DI does is it build the bridges between your enterprise data. A lot of it is numbers and rules and processes. And and and you know, the ability to make those decisions. So so on the number is no question, but on the ability to operate, as I said before, your business with that technology.

• 46:13Speaker 00

  And that's why DI comes in, and that's why it's moving up. And very briefly, on on your point about executive and the control room. No, the control room is not is not the place where you find a CEO or the C-suite. Not yet. I think it will be, but not yet. The control room is where you find the operators of the decision intelligence, the decision architect, the data scientist, the decision analyst, the business leaders. I say, again, I'm repeating what our friend Douglas Guirmeis said talked about last week at a big conference we did together. He said, I can see now in real time what decisions are being made in supply chain in the scope that's been deployed. I could not see it before.

• 46:57Speaker 00

  I can see in real time what decisions are being made for what impact. By the way, fantastic story where he said that the entire project was paid back in 90 days, which is remarkable. That's his words, not mine. But but why? Because you can measure the business impact of every single decision. The executive is the sponsor, is the is the push. And they have access to, you know, they will use ERA with, you know, asking questions and getting answers. But in a control room is really where you find the experts, the operators of that technology.

• 47:31Craig S. Smith?

  Yeah, okay. The I mean, trust is a huge issue. How do you benchmark this? I mean, to get people to adopt it, to get people to to to allow it to make decisions?

• 47:46Speaker 00

  It's I I try to answer quickly, but it's a multi-step process. So your trust is around the data. You have to certify data. So we do a bunch of work in the data layer with and the decision data model with involving Agentic to basically certify the source of the data that will generate the recommendation. Then you need to bring full transparency on the logic that's being deployed on on the scale. And this is why. About fifty percent today of our new business comes from companies who have built some level of capabilities of the AI on their own. But they're doing it by coupling things together. And therefore, they lose the lineage, they lose the logic.

• 48:29Speaker 00

  I pushed the data into that optimizer, then I brought it back and I pushed it into this thing. And every time you do that, you lose the trust. So fundamental to me from the get go was to build all the capabilities that are enabling the intelligence to run to be context aware and in a single environment. So wherever I am in Era, there is a very clear thing, which is you're one click away, one click away from understanding the logic, the source of the data, who's done what, right? So the data has to be certified, the logic has to be accessible. And the explanation of, does it work? Comes naturally because each recommendation, as I explained before, is tagged with expected outcome.

• 49:15Speaker 00

  So you'll see if it's working or not. If it's not working, you go and fix the logic or you go and fix the data. But the system doesn't lie. And by the way, all good. Sometimes you start a new skill and your adoption rate is low because the user is saying, it's not doing what I expected it to do. Great, fix it.

• 49:35Speaker 00

  So, so the math, you know, the math is the math, but is the logic alright? And, and you start sometimes with a forty percent acceptance. And within a few weeks, you end up at eighty. Most of, most of the skills that have been built, they have, you know, seventy-five, eighty percent acceptance. But they also move toward ninety percent plus automation. Automation is when you have enough confidence, enough trust that you can let the system run. But you have to monitor it in real time. You gotta, it's back to the control room. You gotta look at every day as a manager, what is the impact of the work that Era is doing with my people? What is the impact on cash, cost, service level, carbon, water, student retention?

• 50:19Speaker 00

  You name it, you look at those and that's how trust comes, it comes with results.

• 50:24Craig S. Smith?

  Yeah, I mean, do people pick up a business function and and run ERA? And without adopting the decisions, but just running it virtually to see how different its outcomes are from the human decision making?

• 50:46Speaker 00

  No, no, we, I understand the the point, it's it's logical, what you're saying, very logical, but it's not the way it works because you need the human in the loop or on the loop or out of the loop, you you need the human to work with the system. It's you could, I mean, you could run the two things in parallel, but then you don't have the value. So I don't, I can't recall any customer who's done that. What they do is they will take a limited scope.

• 51:14Speaker 00

  And I always recommend customers to go with high volume of decisions, medium complexity, not trivial, not super complex optimization that only three PhD in a company understand. You go with medium complexity, several choices, high volume, and then you run it. So ERA, you can push in super complex, you can push in super, super heavy in terms of volume, kind of find the the the the middle ground in complexity and a high volume. And you'll see people going like, oh, I'll give you an example. I mean, at a very high level, if you think about the language that you use in a company and how you work, you look at classification of everything.

• 51:57Speaker 00

  You classify your customer segment, you classify your products, A, B, C, D, X, Y, Z. Those are, if you work for a CPG company, this is, everybody knows what it means. Why? Because it's a way to allocate the minimum, the the the limited amount of of decision-making power that you have on the decisions that have the most, the highest impact. I'm going to put eighty percent of my people on the A products, twenty percent on B, C and D, the long tail of products, we'll we'll look at it once a year. Digital brings unlimited compute power. I don't care if the impact of that recommendation is a cent or a million dollars. I can equally treat every decision. So you want to, you want to think about it that way.

• 52:43Speaker 00

  It's not a competition between people and system. It's people enriching the system and system helping people. So it's not just math. It's really that interaction that defines a decision intelligence platform. Is that what defines the success of a project? So we've always opposed. Now, sometimes the, the, the fight, so to speak, the friendly fight is like, hey, I'm doing a demand forecast. I'm going to look at era's forecast versus our forecast. Well, fine. The algorithms work the same way. Our algorithm is not better than theirs. It's the same math. It's the same logic. You know, a, a random forest is working out. There is some subtleties in, in subtleties in all of that. But, but fundamentally, it's never been a compete.

• 53:28Speaker 00

  My number is better than yours. It's how do we bring that together to bring the humans and the machines and build that memory? And again, from people making decisions supported by machines to machines making and executing decisions guided by people.

• 53:43Craig S. Smith?

  How are companies measuring success to justify and continue investment? And what does a truly decision first autonomous enterprise look like? And when do you think will, will arrive at that?

• 53:59Speaker 00

  So, so success is measured at two level, right? You look at this, the, the, the impact of every individual decision. So the first time we look at a project, decision by decision, I can look at. So when you start a project, you know, very often you get the feedback from the customer. Oh my God, last week we had three decisions that enabled us to save $500,000 that we would not otherwise see. At a conference in London, AstraZeneca mentioned the fact that, hey, as they're deploying era for clinical trials. And our friend Sam Mulligan said, it saved lives. We were able to, to save lives. It's, it's on the record. It's, it's very interesting to watch. So you measure that decision that era was able to make in real time enabled.

• 54:45Speaker 00

  The medicine to be in front of the patient at the right time. Otherwise, it would have been missed. So you you look at this at the micro level if you want. But the next thing is you say, hey, we're deploying ARI to reduce the amount of waste that we that we generate. So very simple. You look at the before and after and and and that's how it works. And it doesn't work every time. There are situations where the system is not that helpful. Fine, move on. There's another element, which is the you you can have a big focus today on network optimization, reducing logistic cost. And you build the skill to do that. But then your priorities evolve. Now it's about something else. You have to continuously read, adjust the way the system works, the same way you would

• 55:30Speaker 00

  expect your people to adjust their priorities. So it's not like I deploy it, I measure, and it runs forever. Guided by people comes back in the game. And you asked a question about what is a truly, you know, DI first company? We're starting to see that. We're starting to see that. Again, the market is evolving. Well, it's it's a company that puts the decision intelligence agenda at the center of their value chain, not looking at it from a specific line of business or a specific a use case, so to speak, but start thinking about the optimization. Some of our clients, to enable to get agility and reduce cost are a lot of companies out

• 56:12Speaker 00

  there are just outsourcing full functions of their organization, planning, claims to to low cost operators, which are which is the way they could do it. Digitize that process instead. You're creating that memory of data. You don't need that low cost approach because you can fully digitize it. So that's what I mean by DI first is thinking about end to end across the value chain. And yeah, it's it's really starting to happen. And I think that agentic technologies and L L M driven agentic AI is going to really accelerate the adoption because it resolve one problem of DI, which is it's quite hard to model the decision logic.

• 56:57Speaker 00

  The rest is easy, operating easy, but the logic. So the ability to leverage now L L M based agents to to prompt that that decision logic is going to drive massive adoption and unlock in the market.