Podcast Ep. 95: Exploring Causality and AI-Driven Digital Transformation with Dr. Jerry Smith

Share on facebook
Share on twitter
Share on linkedin
Share on email

Episode Description: 

In today’s episode of the Agile Coaches’ Corner podcast, Dan Neumann is joined by Dr. Jerry Smith who you may remember from a bonus episode of the Agile Coaches’ Corner a few weeks back. 

In this episode, Dan and Dr. Jerry explore the topic of digital transformations. Dr. Jerry takes listeners through the process of a six-step AI-driven digital transformation, the challenges of the process, as well as the key benefits. 

Listen on Google Play Music

Key Takeaways

  • What is digital transformation?
    • Changing the behavior of the organization in relation to their customers 
    • Changing their journey map so that they can achieve the business outcomes they want
    • Looks at changing the behaviors to create new opportunities
    • It is not “transforming digitally” (i.e. moving to the Cloud, etc.) – the order of words is important to note
  • AgileThought’s AI-driven digital transformation
    • It is a six-step process that gets a business to actually bend their business curve
    • It is implemented in a set of capabilities; there are over 67 capabilities that transition an enterprise’s data and transform it into insights and actions, and is a systematic process
    • This process puts a customer into the position of changing their business
  • The six-step AI-driven digital transformation process:
    • (1) “Data is the debris of human activity. We collect it all, but all is not important.”
      • The first thing that is done is data collection
      • The most important question to ask when you begin is: “What is data?”
      • Data is because of us; not in spite of us
    • (2) “We determine what data is causal to the business problem. This allows us to only focus on those areas we can control.”
      • You need to ask: “Of all this data we collect, what is causal to my business problem? What should I be focusing on?”
    • (3) “Using causal data, we build digital twins – surrogates – of the problem. We create an artificial model of the real world.”
      • They build high-quality, predictive algorithms (from step two’s causal data/input)
    • (4) “Within the artificial world, we organically grow perspective solutions designed to optimize the business outcome.”
      • Now that you have the model, it is important to optimize the digital surrogate
    • (5) “We implement the prescriptive solutions, wait for change, and collect new data.”
      • In this step, you are running through optimization, changing those inputs, and looking for a combination that results in that output achieving the business goal
    • (6) “The cycle repeats, bending the business curve.”
      • When you have the behavior of the people that marketing, sales, and product development will have to change, you can then wash, rinse, and repeat
    • When you go through the six-step process in cycles, you need to give enough time to see the ripples go through to see the changes and continue to iterate and refine
    • “This is why this six-step process is important for customers; because for the first time we’ve actually connected business and IT together.” – Dr. Jerry Smith

About Dr. Jerry Smith: Dr. Jerry Smith is AgileThought’s managing director of analytics and data science. As a practicing AI & data scientist, thought leader, innovator, speaker, author, and philanthropist, Dr. Jerry Smith is dedicated to advancing and transforming businesses through evolutionary computing, enterprise AI and data science, machine learning, and causality. 

Mentioned in this Episode

Transcription [This transcription is auto-generated and may not be completely accurate in its depiction of the English language or rules of grammar.]

Intro: [00:03] Welcome to the Agile Coaches’ Corner by AgileThought. The podcast for practitioners and leaders seeking advice to refine the way they work and pave the path to better outcomes. Now here’s your host, coach, and agile expert, Dan Neumann.

Dan Neumann: [00:16]
Welcome to this episode of the Agile Coaches’ Corner. I’m your host, Dan Neumann, and happy to be joined today by Dr. Jerry Smith, managing director of analytics and data science here at AgileThought. Hey, thanks for joining again, Dr. Jerry.

Dr. Jerry Smith: [00:29] Oh, it’s great to be back. Um, and it’s always fun to have these kind of open conversations where, you know, folks like you and I can chat about interesting things and hear back from, you know, the listening audience as well.

Dan Neumann: [00:40] Definitely just a little bit ago, you did a bonus episode on artificial intelligence. And so we’re going to probably reference maybe some of the content in that as we go forward today and explore digital transformation.

Dr. Jerry Smith: [00:55] Yeah, yeah. That one, that one was a lot of fun because we actually went really deep, which is unusual for a first one. We went deep into causality. And so today we’ll reference that in the process. Uh, so, uh, for those folks who, um, who are interested in causality in a different way than you often think about in a business way, um, you ought to reference that, uh, that, that, uh, podcast, uh, because we’re going to talk about it today and put it in context.

Dan Neumann: [01:23] Wonderful. Well, let’s, uh, let’s start building some of that context. I think the term digital transformation gets used kind of loosely gets thrown around sometimes in a meaningful way. And other times it’s just a buzzword bingo that, that people are playing. So what, when you use the term digital transformation, how would you describe that to people? What does that mean to you?

Dr. Jerry Smith: [01:45] Yeah, that’s a great place to start because you know, marketing is powerful, right? You know, people marketing is, defined the term digital transformation. They defined artificial intelligence, but after all that definition, we often don’t recognize it. And we don’t know how to use it. A couple words, digital transformation, the position of those words actually have meaning. There’s a difference between transforming an organization digitally that is moving them to the cloud, getting them into Amazon, Microsoft, Google, IBM, right? Getting, getting their products in a more digital format versus digitally transforming them, which means what? Which means changing the behavior of them with relationship to their customers, right. Changing that journey map so that you can achieve the business outcomes you want through the behavior of people. So it’s a fundamental change. All digital transformations have to have a transforming digital component to it, but not all transforming digitally, people moving to your cloud, actually deal with the digital transformation.

Dan Neumann: [02:51] Yeah. Just that, that movement of a machine from your computer to somebody else’s computer in the cloud, whether it’s Amazon or Microsoft, isn’t going to change the behaviors. It’s not going to create the new opportunities that, that people are hoping it will just by moving where the compute power happens to be happening.

Dr. Jerry Smith: [03:09] Yeah, you’re absolutely right. And, and there in lies a disconnect between business and sort of the IT digital transformer, a business, come in and say, Hey, we, you know, we need to change some new, uh, get some new results out of our business, our revenues, lower margins aren’t what they have our customer satisfaction is that there, right. If you think about some verticals where people are still sick, you know, people committing suicide at a higher rate, our products are spoiling on the, uh, our products are spoiling in our retail stores, our supply chains aren’t working as well as they want, we need to do a digital transformation. And the way it hears that is, Oh, I need to transform your digitally. So here’s what I’m going to do, I’m going to take all these disparate systems and move you to the cloud. I’m going to take all this work and move you into a warehouse. And that’s what they do. Uh, one of the beliefs that I have is that this field, AI specific, which we’re going to get into is too important to be left up to IT. Right. It’s just too important. And we do need to move it back and there in lies. The first thing I think we want to talk about today, which is AI driven, digital transformation.

Dan Neumann: [04:06] I do want to, I do want to talk about AI driven digital transformation. And as you were saying, some of those things, Hey, we want new results. That’s the same thing we see I’m in the Transform practice. That’s the same thing we see when people are like, Oh, we want to go get us some of that agile stuff. We want to go bring in scrum. Or we want to bring in a scaled framework, et cetera. It’s, it’s treating maybe some of the symptoms much like moving the computer from on prem to the cloud, as opposed to really looking at new behaviors, they’re going to help get you those new results. So, um, that, that was kind of the, the epiphany for me. Yeah. Yeah. And if you think about our practice of Transform, Build and Run, here’s a key, here’s the conversation I had with one of the largest retailers on planet earth. Um, they started off with a vendor, another competitor of ours who said, listen, we need to, we need to transform our practice. Our, our, our stores, you know, the thousands of stores out there often are out of products. And the person said, yeah, let’s transform you. Let’s build some wonderful, uh, uh, inventory, user interface, some displays, we’re going to create a great display for you. And so they had that, you know, Jerry and Dan would dial into their respective stores in the morning and they’d say, Hey, this is what’s out of stock. This is, what’s not out of stock. This is the timeframe it’s going to take to get it, et cetera, et cetera. And that was the, that was the digital transformation, right? They got them out of paper and they moved them into digital. And it was a success. You, you, you checked off the statement of work, the owners of that products that we were very successful. It runs there’s very little error, but then you go back and look at the company and you say, are you meeting your business goals? Has your revenue increase has your customers been more happy? Have you sold more of these products? And the answer is no. And it was a simple set of questions that we ask. That’s part of this process we’ll get into, which is, what do you do with that? What do you do that presentation? And this person, I remember saying, here’s what we do with it. We actually look at it and I say, you know what? I’m out of this product. I wonder if I go to this other store that I know is near me. If I go to this other store, I wonder if they have that product, Oh, look at that product. They have, I’m going to ask for inner store transfer. And I said, well, would it help you? This is question one part of the transformation. Would it help you if I predicted for you, which of these products you have that will be out of service or out of stock tomorrow, or will not be out of stock tomorrow, sort of a red, yellow, green, red you’re out of stock yellow, you’re, you’re kind of maybe out of stock in green, you’re going to be, and he goes, that would be awesome. Can you do that? So that’s one of the first steps. That’s not even, that’s not even transforming you digitally. That’s that’s, I mean, that’s not trans digitally transformed. That’s transforming us beginning to think through the power of prediction. I said, now, what would you do with that? And he said, you know what I do now that I know tomorrow, I’m going to be out of stock. I’d definitely go to take a look at that other store and see if they’re in stock tomorrow. And then I put the transfer. I said, well, what if we did that for you? What if I then took and did that for all the stores? And then I looked to see who in your area was in stock. And then I created that SAP entry for you. He goes, can you do that? And I said, certainly we can. So now where are we? We’re at a spot where we’re predicting you’re going to be out of service or out of stock for something. We then look for stores, which are in the area that have a low cost of transfer. We put the order in to move those things over through overnight thing. And the next day you have those products. I said, now, what would you do? He said, I would focus on those things that just make my customer happy now. That’s 80% of my day is just trying to figure out where my stock is and what I need to do. I can really sit down and actually focus on customers. In that simple journey, we transform them digitally. And didn’t do anything with the cloud. Now, of course it’s implemented in the cloud. But the point is, is that he went from just looking at stocks and blocks from paper, into creating reports, into actually understanding the world can be, uh, predicted to the prescription being movement. And now when he’s done through that, he’s working with his customers. I think every business person wants to be in that same situation.

Dan Neumann: [08:08] Yeah. The business engaging with its customers is where, where the real value gets created. I, it seems that getting, making sure that you have stock for people to buy as the blocking and tackling of any of any, any business really.

Dr. Jerry Smith: [08:24] And therein lies the challenge that a lot of these companies have, right? Um, business people are super smart, right? But they often over rely upon our IT shops, which are also super smart, right? Because a business person doesn’t know what artificial intelligence is by the technical definition or what clustering and classification or time series analysis. They outsource the implementation to a group of people they believe. And so, you know, AgileThought we’ve actually broken this thing down. We, we have an AI driven transformation process, a six step process that gets a business to actually bend their business group. That is then implemented in a set of capabilities, right? We have over 67 capabilities that transition your data, your enterprise data, whether it’s your supply data, your customer data, your ERP data, your IT data, how they’re interacting with your system or open source data, right? Third party data, whether it’s economic, social, retail, data, et cetera, taking that data and transforming it through data, information, knowledge, and wisdom, into insights and actions. It’s a systematic process. And what these two things together allow us to do is a put a customer in the position of changing their business through these six steps which we’ll get into next. And then the second thing is, as practitioners as Dr. Jerry, I’m able to say, have I missed anything right? Have I missed that? You know, this is, this is supply chain. Have I done an event analysis? Have I have I take a look at a temporal analysis or a spatial analysis? I mean, these are all fun words that are interesting to guys like me, but not fun words to business people they don’t care about.

Dan Neumann: [10:02] Nor should they have to, they’re focused on their business results. And that’s why they partner with, with technology folks. I can only remember about three to five things. And so I want people to know that as we go through this, we’ll put an artifact on the show notes where they can download a picture of the six steps in a process so they can have that to refer to them. They don’t have to kind of try and create a mental model independently of, of that.

Dr. Jerry Smith: [11:11] Yeah, we’ll do that. That’s fun. And, uh, and again, I would encourage anybody that’s out there looking at this stuff, uh, to, to leave notes and comments and, um, spend some time talking through it. The piece I’d like to focus on first cause the capabilities model is really a visual model and you and I still have to figure out a way to have a show on that where it’s highly visual, visible, or visual, I should say. But the first part is the transformation process. It’s a simple six step process and there’s some unique things in there and there’s some things that aren’t unique. So let me just walk through the process real quick and then we’ll dig into very first thing we do is we collect data and you’ll hear me say this over and over. So I’m going to pause for a second. And this is for the people that are running businesses and the most important thing that will transform your life is to, to answer those questions. What is data now, you’re going to in the back of your brain. And you’re going to sit there and say, yeah, I know Dr. Jerry is trying to trick me on this one. You know, data, some sort of ones and zero aggregated into words and constructed into characters. No, no, no. It’s really simple. Data is the debris of human activity, right? It’s because of us, not in spite of us, right? The data that is in your systems, the stuff that you analyze in those reports is because of a human being, right. Interacting with what, an application that you paid to build, which then does what? Persists, whatever it has into a database, which then we do what we create reports around. Right? And it’s in those reports that we don’t like the results. I don’t like what my revenue is doing. I don’t like what my supply chain is out of date. I don’t like my customers being sick. So we generate these reports and we often focus on the data. So the very first step is to recognize that data is the debris of activity. Why is that important? Because of step two, we talked about this the last time in detail, I’ll call it out here and then I’ll move on, which is the very first step. Every business person on planet earth needs to ask is of all that data you collect Jerry and Dan. What’s causal to my business problem? What should I be focusing on? Right? That is a complex step. That’s a statistic process otherwise known as mutual information theory is a whole bunch of practices in there. It’s part of the influencing set of data science techniques that we have. But the most important thing is you take a hundred things that we store and we narrow it down to those 10 that are causal. Why is that important? Go back to step one. That data is the debris of our activity. So if I know what causal data, I go back to the application, I then go back to the person and herein lies, the Rosetta stone for business, what behaviors now do I have to change in the person to get that data to change that way? So there are two key roles that are missing in every transformative program out there for the most part. And that is once you know how that data changes. Step two, what’s causal. And you go back through the application, go back to the human. You need to sit down with a psychologist. Somebody who actually can say here’s how human behavior changes and a sociologist, somebody who says here’s how people change in general to ask the question, how do we get these changes? And then that information goes to marketing and goes to sales and goes to product development, et cetera, et cetera, et cetera. That’s the thing that drives those areas that is transformative for coming. So now that you have step two, you go into the easy stuff, right? We, now that we know the causal data, step three, we build digital surrogates, right? We build high quality predictive algorithms. The inputs are what step two, the causal data. We know if we change this data, we know it’s going to change what the business outcome, which is what we want. We want, uh, you know, less spoilage in our, in our products and service. If we want customers to be happier, we want less sick people out there. So changing the inputs, we model it. Deep neural networks. If that floats somebody’s boats, the decision trees, there’s a lot of different tools that you can develop these surrogate models and step three. Therein lies where most people stop the process. Hey, I got a prediction. You’re telling me that my supplies are going to be out of stock, right? But we need to go one step further because now that we have that model, what can you do? Optimize, optimize the heck out of that digital surrogate, because what is it telling us? Hey boss, if you tell me some inputs, I’m going to tell you what I’m going to do. I’ll tell you if I’m sick or not. If I spoiled or not, et cetera, et cetera, my supply chain. So what do we do? We run through an optimization. We change those inputs and look for the combination that results in the output, achieving our goal. And when I have that, what do I have? I have the behavior of the people that marketing, sales and product development will have to go out and change. They go do that work wash, rinse, and repeat. You wait a day, a week, a month, you collect new data. You figure out what’s causal. You, create the new models. You then look for the prescriptions. And over time, this is the only process because it’s closed loop. We’re connecting your data to real programs that are changing human beings. You actually bend your business curve. It’s an amazing process.

Dan Neumann: [16:18] And as you described that process, you mentioned it, the wash rinse and repeat the speed at which you can get through those loops, those cycles. Obviously you need enough time that whatever you tweak, do you want to see that effect ripple through so that you’re not just thrashing making changes, making changes without them coming through with actual output behavior changes. But you do want to iterate through that and continue to refine it. So you, like you said, you bend to that curve. You keep getting a better and better results coming out the other side from your customers.

Dr. Jerry Smith: [16:51] Absolutely. And there’s actually a technical term for that process. So everything in this process has, has science behind it, which is the kind of cool thing. And again, that’s where IT, we listen, I love IT. Right. Um, but sometimes we outsource things to a group that isn’t in the position of best representing the business. So when it comes to things like how long do you wait? It’s the Nico’s theory, right in Nico’s theory says it’s basically, Oh, it was established when we were trying to think through, at what rate do we sample the frequency of, of, uh, audible information to reproduce the fidelity of our ears, right? And then the theory is that you, the practice is you, you sample at twice the rate of the highest frequency you want to hear. We do the same thing in this space here. We ask the question, how often do people make these decisions? And we let the period of time go twice as long for that. So it gives time. So if you’re in, if you’re trying to sell cars, right, the average person out there is looking buy cars. It usually comes across a desk every once or twice a month. That’s the average frequency two or three times a year. Somebody in the back of their head is going to say, you know what, maybe I’ll look at a new car. It doesn’t make sense to go do this and then change your models every day, because you’re not capturing the change in human behavior. So what do we do? We do it on a quarterly basis, right? It just makes sense. We, we make the prescriptive changes. We’ve run it for a quarter, go out there, spoilage in a store. That’s a daily thing, right? So we actually want to update our behaviors two or three times a day to catch those behaviors. So there’s a, there’s a science just behind. That’s step six of waiting. How long do you wait in order to capture, uh, the signals that human beings are delivering to you? So this is, this is why this six step process is important for customers because for the first time we’ve actually connected business and IT together. Business and IT are finally connected together, joined at the hip in terms of achieving business, real business. People say, these are the things that are important to me. And I now know the data that it’s causal. And IT says, this is the data we’ve collected. And we waited long enough for us to then re-predict that stuff and re-prescribe, that stuff, the prescriptions are a common, um, point of view in which we, uh, we help the company change out of that. And again, we all, we, we use the capabilities model to do all this background, fancy work in there, right? It’s just like architects back in the day, right? You could be a TOGAF architecture. You can be architect. If you’re an architecture, you use those models in data science, AI, machine learning, AI, we use this capability model to achieve our same result.

Dan Neumann: [19:26] Yeah. And I think we’ll have to put a link to those because even though I had a little AI back in school, I was just like, yeah, I don’t know what those models are, but that may be yeah. Either for another day or a deep dive or something like that. But we’ll leave. We’ll put something in the show notes for, for people to go with that.

Dr. Jerry Smith: [19:40] Well, I mean, a data science is pretty simple, right? We got three categories of stuff. We got data science, then we have AI. And then we have cognitive computing, right. Artificial intelligence. And in the data science area, there’s only three practices. Um, there’s the first practice, which is the frequency stuff, which is the clustering and classification, time series analysis. You hear a lot about that. The second one. And by the way, this is for my audience. This is the only joke I have. It’s on the statistical area. It’s in the inferencing area. Now wait for it. This is the area where five out of four people have problems with statistics. Scott let that sink in for LA five out of four people in the last one’s in network theory network during this is the most powerful and most underused, right? It’s, it’s the, you know, six degrees of freedom, Kevin Bacon stuff, right? How are things connected together? And by combining all three of those use establish a very, very powerful set up data science, which is the study of data, so that then you can model it and predict and then prescribe from it. Yeah. Pretty interesting stuff. It’s easier than people think it really is. I mean, I, that’s why I started off with saying marketing has done a really good job of, of creating an understanding, simplifying it to the point where it’s so simple. It can’t be used. Our job is to simplify it, but make it usable, simplify, but not simple. That’s our job.

Dan Neumann: [20:58] Yeah, definitely not simple. And then it also strikes me that agile frameworks and agile values collaborating with the business and the IT collaborating together and that continuous improvement process and inspecting and inspection and adaptation, all those things fit really well into an AI driven digital transformation.

Dr. Jerry Smith: [21:18] And therein lies, lots of examples, right? Um, this is, you know, the, the, the process of, of there’s five steps, there’s going from data, right? Your data going to data engineering. So we have to take all that data together. And then we create what? A machine learning record. And there in lies the bridge, the, the contract between the data world and the insights world, right on the insights, you start with data science. The second word is important. It’s called science. And that doesn’t mean that we’re going to actually achieve the results you want. If somebody says, Hey, you know, can you help me identify the causal data in this? I can go through the process, but there may not be causal data here, right there. You may not have collected the data that captured the causal characteristics of the human being. By the way that’s a good thing. If we find out you don’t have it, what does that mean to you? You have to digitally transform yourself. On the other hand, if you have that data, you may just have to transform yourself digitally in that particular case. So it’s a good thing. And then you move into, after the data science piece, you move into the machine learning, which is fundamentally about what? Predicting tomorrow, next day, the day after that, predicting how those, that data will change, Oh, spatially left and right coast. And then finally, now that, you know, predictions now that you know what your inventory is going to be based upon the signs of that data, you can prescribe actions, right? So it’s all tied together. Now you’re going to get a PhD in AI like myself, out there arguing those definitions. But these definitions allow business people to have meaningful conversations with their IT partners in delivering services. So they don’t get confused. That’s the whole idea. Tell me what my data is doing, what I should, what’s important to me, then help me understand how that’s going to change as a function of my business concerns. And then lastly, put me in a spot now that I know that, that I can take meaningful actions through artificial intelligence.

Dan Neumann: [23:13] Wonderful. Well, um, I’m inspecting our time box and I think we’re going to have to adapt here. So, um, super cool that you were able to join and share on AI driven digital transformation. And we’ll look for some feedback from folks on maybe where, where they want to go next, as far as exploring this topic. So thank you very much for joining.

Dr. Jerry Smith: [23:34] Definitely been a lot of fun, and I enjoy these kinds of conversations where we just sort of riff off of, uh, off of a topic. And if there’s topics that people want to hear out there, business problems, they want to see solved with this approach or, or just some technical stuff. I can’t imagine anybody wants to get into the science, but if they do right, um, uh, let’s put it in the, put it in the show, put it into the notes and we’ll respond to it.

Dan Neumann: [23:57] Wonderful. Yep. They can email or they can tweet it with the #AgileThoughtPodcast or, you know, put a comment on the, uh, the website. So yeah, lots of different ways we’ll look forward to it.

Dr. Jerry: [24:09] Cheers, Dan.

Outro: [24:12] This has been the Agile Coaches’ Corner podcast brought to you by AgileThought. The views, opinions and information expressed in this podcast are solely those of the hosts and the guests, and do not necessarily represent those of AgileThought. Get the show notes and other helpful tips for this episode and other episodes at

Stay Up-To-Date