On this episode of the Futurum Tech Webcast – Interview Series, I am joined by Pegasystems’ CTO, Don Schuerman, for a conversation on what the autonomous enterprise looks like and why it matters for today’s businesses.
Our discussion covers:
- The rise of enterprise AI, and the convergence of technology and industry
- How Pega has focused on the autonomous enterprise
- The benefits of a solution that can help to automate workflows and the opportunity for organizations using AI and automation to increasingly become self-optimizing
- How using AI can transform marketing approaches for personalized customer engagement
- What Pega is doing to drive growth for their customers, helping them to stay ahead with the skills and technology stack that enables transformation
Watch the video below, and be sure to subscribe to our YouTube channel, so you never miss an episode.
Listen to the audio here:
Disclaimer: The Six Five webcast is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Transcript:
Daniel Newman: Hey everyone. Welcome back to another episode of the Futurum Tech podcast. I’m Daniel Newman, CEO of the Futurum Group. Joined today by Don Schuerman, chief technology officer, Pega, a veteran, a regular here on the Futurum Tech podcast. Don, how’s it going?
Don Schuerman: It is going quite well. Quite exciting times.
Daniel Newman: It is, and it is great to have you back, I don’t know, veteran alumnus, regular correspondent contributor, I don’t know, but you’ve been with me many times before. Always great conversations. Been talking to you for goodness, probably the last five or six years and talking to you about something that suddenly has gained massive amounts of popularity, and that is AI. And I like to call us the AI OGs because we talked about AI before it was cool to talk about AI, but this year, or you’d call it now, last year, November 2022, with the advent of a publicly facing ChatGPT AI became all the rage and has now pretty much been the pacesetter of the public market. The pacesetter of every technology event in the world and pretty much the number one topic in every boardroom for every company in every industry. And it’s almost like a moment where Don, you can say, “I knew that.”
Don Schuerman: Yeah, it’s funny, I was just reflecting back in 2019, during one of our keynotes at PegaWorld, their annual event, put up a page of text and pointed out to people that text had been entirely generated by this new AI model called GPT-2, which was becoming available. This was back in 2019. I think if you’ve been in the AI world, you’ve seen uses for AI and you’ve seen some of these large language models and the power of them growing over time. I think the killer app of ChatGPT just suddenly brought it to everybody’s forefront, as I like to say.
Everybody and their CEO is now trying to figure out what to do with this, how to use it safely, how to use it effectively, how not to get left behind by some of the change we think it might bring. And I think as we go through that, it’s also really important to both understand what’s going on with large language models and generative AI, but also important to remember that AI is a family of technologies of which degenerative AI is a type in a specific use case. But there are lots of AI technologies that depending on the use case, organizations want to think about and have a plan about deploying across their businesses.
Daniel Newman: Yeah, you did call it. Actually, I remember that keynote, something you said, it stuck with me, but it was where you said, you were talking about how much better apps know us than the people in our lives. You did this diagram where your spouse, your mother, your friends, and basically that Facebook knows us better than we know ourselves. So if you ever wanted to know if you get credit for the invention of thoughtful stage fodder, that was one of the things that stuck with me. I remember tweeting about it that Facebook knows me better than I know myself. I think that’s become progressively true for the internet, that we’re all at the point where definitely the most intimate companion we have is the hard drive on our computer.
Don Schuerman: And I have to credit Dr. Rob Walker who’s our head of decision management with that insight. But it’s definitely true and I think part of the thing that’s happened is AI has been so immersive around us that we almost sometimes don’t notice it’s there. It’s in the algorithms of the social media tools we interact with. It’s in Netflix recommending what movies we should watch. It’s helping us do our Google searches, it’s recommending. Do we need to make changes to our 401(k) plan? It’s all over the place. I think intelligently and deployed in a way that helps people, I think it is incredibly effective both for businesses but also for all of us as customers and consumers.
Daniel Newman: Yeah, I like that you said that. I’ve been trying to have this AI’s been here conversation with the world for a little while and I feel like sometimes I’m banging my head against the wall until my hair starts to fall out. I hope you understand the humor in that, Don, but first of all, in many ways the generative AI tools that we’re all really excited about are just like search 2.0, meaning that it’s filters and recommenders, which by the way, we’ve been using very, very capably for a long time. You just mentioned a bunch of different ways we’ve been using those things, but the generative text from GPT is basically just a summation of the search responses with some algorithm that prioritizes which ones.
And the better generative tools use proper attribution and stuff because some of them don’t do that, and that’s a bit of a problem. And we can talk about that, but one of the things that the really exciting and important things that has come out of this as we’ve started to… I’m sure I’ll get a few people to yell at me when I say the commoditization of large language models, but to some extent we have seen that for the average person now. But what we have also seen is the rise of enterprise AI, the rise of this idea that… Look, the really important data isn’t the open internet data. Everybody’s had that for a long time and Google has actually organized it pretty well for a long time.
But what people haven’t had is a way to use enterprise search, for instance, to take the capabilities of any of these large language models. And pair it with vector search that can take things that are not just language, but image and everything else. And then couple it with all kinds of really high value ERP and HCM Data, and supply chain data and proprietary customer data. And put that all together now to start creating outputs that help you sell faster, customer service better, design more intuitively. Talk a little bit about how you’re seeing that pivot to enterprise AI.
Don Schuerman: Yeah, I think there are a couple of dimensions or pieces of this. I think for a long time we’ve had what I would call left brain or analytical AI. Organizations have been using it for a long time. We have a product called Customer Decision Hub. It is used to recommend Next-Best-Action. So what offers or conversations or services are you talking to your customers about? And we’ve got organizations who are using that hundreds of millions, billions of times a day to make these conversational recommendations. That’s analytical AI. It’s AI that makes decisions. It recommends, “Hey, here’s a Next-Best-Action or a conversation to have, or here’s a way to optimize a particular business process.”
And I don’t think that analytical AI goes away. In fact, I think a lot of the enterprise AI that we talk about is actually analytical AI. How do I take my data about my customers or my data about my processes and feed that through an analytical AI engine so I can make better decisions whether it’s to engage my customers more effectively, deliver better service, sell more product, run my processes more efficiently, et cetera. Generative AI is like right brain AI. It’s the creative side of AI. It can actually create things, it can create text, it can create images if you’ve played with DALL-E. And I think it’s important for all of us to remember it only knows how to create stuff because we fed a bunch of stuff into it.
It only creating by essentially at massive scale regurgitating stuff that it’s already found other places in the internet, but it does feel like it is creating text or summarizing or building us new images. And I think the really exciting opportunity is when you connect those two things together, when you connect your own enterprise, often proprietary analytical models. And then use things like public more commoditized generative models to do things like explain how decisions are being made or you mentioned the idea of search.
One of the big things that I think it’s also important for people to understand, I often hear people talk about, “Well, are we going to take gen AI models and train them on our business?” And my answer is, “Probably not.” Actually. What you’re going to do is you’re going to do really sophisticated prompt engineering for these generative AI models. You’re going to build, as you said, vector databases that have all of the documents say in your internal knowledge base.
And you’re going to use that content, combined with content maybe in your case management systems, your CRM systems to construct really specific prompts to a generative AI engine that can say, “Based on this content and this customer context, write me an email that will engage this customer.” And now I get something that is uniquely based on my content, uniquely aware of my customer context, but harnesses the creative regenerative power of some of these commoditized AI engines.
Daniel Newman: So I follow you there and I think we’re in this era of a cascading effect where we’re going to move from this large language to smaller and smaller, more sophisticated models that focus on fewer and more specific problem sets. Don, we’re seeing now, I’ve been briefed by a litany of companies that are trying to solve say one particular healthcare approach. And they’re trying to solve for allergies and diagnostics. And they’re saying we’re going to have a dataset, a foundational model, a generative tool.
And so we’re seeing something that’s interesting is this convergence of technology and industry. Now, we’ve seen the industries being veneered for a long time this way, but foundational models… Because you said they may not be trained to say, “Just learn about my business.” But what they are going to be done and trained to do is say, “Fix this specific problem.”
Don Schuerman: That’s right.
Daniel Newman: So as you’re seeing it, first I’d love you to weigh in on that idea more broadly, and second of all, with you really having been focused at Pega on the autonomous enterprise for a long time. How does this start to snowball into how you think about building product?
Don Schuerman: Yeah, so one, I think is we’re entering this phase where everyone begins to see, especially on the generative side, that there are revolutionary use cases for this. But most of the enterprises I talk to are still tiptoeing as to exactly where those use cases are going to be and how they’re going to implement them in a way that’s safe, and protected and explainable. Because things like explainability, being able to actually describe how a model makes a decision is really important. There are a lot of generative AI models, for example, as you pointed out, that can’t even provide references for how they answered a particular question.
If I don’t have that explainability and that traceability, it makes it really hard to deploy a model like this in a regulated industry like a bank or a healthcare provider, et cetera. So I think organizations near term are… I’m seeing a lot of look at human in the use cases, so how can I deploy generative AI in cooperation with something that a human’s doing where it’s assisting accelerating aiding what the human is doing, but that human exists as a check, as a backstop, as a validation of what the AI is doing.
So we’re about to release the latest version of our product portfolio called Pega Infinity ’23 and it’s got a lot of gen AI capabilities embedded in it that fit along that lens. So for example, a lot of our clients use Pega to build workflows that they want to automate. Well, we can use gen AI to say, “Take a name of a workflow.” Say, “I need to build a home loan application.” Well, I can now generate what are the most likely 20 steps of that workflow and what is the most likely data model that goes along with that workflow so that I can cut a lot of the initial design discussions out of building that.
And instead of my Locode developers starting with a blank page, which nobody likes to start with, they can start with at least an approximation of what the process might look like and then make their own changes, and edits, and overrides, and removes, and ads. But I’m accelerating the process or using gen AI to generate test data so that once I’ve built my workflow and I want to test it 100 times, I don’t have to send my developers off to fill out a spreadsheet with their best friends’ pets names. I can just say, “Generate me 100 records to test this workflow,” and we can do it.
Daniel Newman: By the way, you’re talking about synthetic data.
Don Schuerman: Yeah.
Daniel Newman: Okay. I’m just saying not everyone out there knows that, but that’s one of the coolest capabilities. That’s a lot of what we keep hearing about when we hear about this autonomous simulative environment is going to come down to the ability to create useful data to test hypotheses. We’ve been doing this for a long time in things like simulation and design, but it hasn’t been used in a lot of automation.
Don Schuerman: It hasn’t been used or it has required in some cases some pretty sophisticated data science skills to do it. And I think there’s the opportunity now to lower that barrier a little bit. Where we think all of this is going is this concept that we call the autonomous enterprise. And what I think that is the opportunity for organizations through use of AI, through use of automation to increasingly become self-optimizing, to have processes in the business that are automatically finding new ways to become more efficient.
To have customer engagement strategies that are automatically looking at customer behavior and recommending better ways, more effective ways to engage with customers, whether you’re trying to sell more product or improve your net promoter score or reduce your cost of service so that a businessperson can set goals for what they want their business to achieve. I want this process to have the lowest possible throughput, or I want this particular process to have the lowest possible cost ratio associated with it.
And then have the system through looking at data, through application of not just generative AI, but analytical AI models, data from process mining, pattern detection, there are lots of pieces here. But assemble that together to actually then dynamically recommend back out to a businessperson, “Hey, I found a bottleneck in the process. I think I could fix it by inserting an AI prediction to skip the step when it’s unnecessary. Do you want me to go ahead and make that change for you?” And when the person says, “Yes.” The system being able to like, “Great, I made that change. I’ve deployed it to your development environment. When you think it’s ready to go, let me know and I’ll put it through your DevOps pipeline to actually push it to production.” So that now all of a sudden, the process itself is pushing back out to the business owner where the opportunities are to become more efficient and become better.
Daniel Newman: So quick one, I’ve heard your boss at least on a few occasions talk about software that writes software.
Don Schuerman: Yep.
Daniel Newman: I think that’s the way, the line. Are we there? Is that there? I’ve watched some of the demos, so this is a bit of a leading question, but I’d like to get your take.
Don Schuerman: I think we’ve had software for a while. The Locode space is essentially software, the right software. I take a picture, I draw a picture of a business process, or I drag and drop a user screen, and under the covers, what we’re actually doing is generating the computer code that’s needed to actually run that thing. Where I think this is different is that still implies that a human being is driving 100% of all the decisions of what needs to be in this process. What are the steps? What’s the best way to lay out this screen, et cetera.
I think when you get to the autonomous enterprise, you’re going to have the system itself through data and AI models, generative and otherwise recommending how to build that process to drive the most effective outcome, how to lay out that screen to make sure that it’s easy to use and accessible to all users, and we’ll work across your channels. How to design a treatment.
One of the other use cases we’ve put in Infinity ’23 is the ability to use generative AI to take a marketing offer that you might be making and automatically generate treatments for different audiences. So automatically able to say like, “Great, I got an offer. What’s the copy text I should use for a millennial versus a new set of parents versus a retiring baby boomer and recommend to a marketer here are 10 different treatments for the same offer targeted at the different audiences that we see in your portfolio.” And then allow the marketer great, be like, “Great, I would tweak that, I would add this.” But now allowing the marketer to do far more than they could do if they were hand cranking all of those treatments themselves.
Daniel Newman: Yeah, I think there’s this big inflection debate impasse of companies. I think the ones that are going to win around generative and even enterprise AI more broadly are going to be those that are seeing this as an exponential growth opportunity. There is a little bit of that augment displaced talk and of course, I think just like history has always presented, people will need to continue to improve and invest, and learn, and grow, and the timelines from disruption to disruption will get shorter. And people that are slower to adapt and adapt are going to be the ones that have the hardest time.
But I actually think companies that can say, “Hey, we’re not going to try to do marketing at one 10th the cost, we’re going to try to do 10 times the marketing with the same spend. We’re not going to try to cut our sales overhead.” We’re going to do 10 x more sales, we’re going to produce and push revenue. By the way, we saw the same thing happen during COVID. There was a very good instance of this has nothing to do with AI. It’s just when you work more, you sell more, companies grew because people worked more hours. It was like a very… And this is like a machine version of that.
Don Schuerman: They weren’t commuting and all of a sudden you were getting more time spent doing stuff.
Daniel Newman: Yeah. You were meeting, you were selling, you were…
Don Schuerman: We saw this even before generative AI with the analytical AI that we’ve got in customer decision hub. The benefit there wasn’t to help marketers do traditional segment marketing faster. It was no, no, no. Segment based marketing traditional approaches that spam people with a bunch of stuff is broken. If you use AI, you can actually transform your whole marketing approach so that every customer conversation is personalized in the moment to that customer.
I can completely shift the way I think about each and every customer conversation that I’m having and that is what generates for many of our customers, hundreds of millions in increased revenue because every time they talk to the customer, they’re only talking to the things that are incredibly relevant to that customer in that moment. That’s the ability of using AI to completely rewrite the script, not just do the same thing you were doing before better.
Daniel Newman: So let’s wrap up here. I want to hear about what you’re doing at Pega. I want to hear about the evolution. You’ve always talked about the autonomous business, the autonomous enterprise. How does this inflection of AI adoption, the new-found yet slightly delayed fuse for the importance of putting this to work? Don, I’m glad we’re here, but how are you going to jettison forward, propel the growth, drive the next iteration? I’ll give you a bunch more cliches. How are you going to do that over at Pega?
Don Schuerman: Yeah, look, I think fundamentally the first thing that we start with is the stuff that at Pega we help our clients do day in and day out. And that is use AI powered decisioning and workflow automation to better engage their customers, better acquire and onboard their customers, serve them, run their operations, fix problems when they go wrong, really drive all of the decisions and workflows that surround how our clients engage with their customers. And I think by applying generative AI, we’re about to take a step change in how that happens.
So I imagine and we talked about PegaWorld, about this idea of being able to have an autopilot that can assist a user so that if a user wants to… Needs to assign an appraiser to a claim. Instead of the user going through and figuring out, “Okay, what appraiser is local in the area? Who knows that is skilled in the auto claim with water damage that we’re dealing with and who has availability this week? And by the way, what’s the schedule for the customer, et cetera.”
If they could have an autopilot that would just basically say like, “Yeah, I need to find an appraiser and get it scheduled for this claim.” And that autopilot could do the research, go through the library of appraisers, reach out to the customer via email, say, “What’s your availability?” And come back and say, “Hey, I found this appraiser. He’s available at this date. The customer is too. Would you like me to formalize that schedule?” So the claims’ adjuster can be like, “Yeah, I’m good.”
Think about the change in scale that thing would allow or think about the impact. If I had an autopilot for my business leader who could say, “Hey, I want you to monitor my claims process and I want you to continuously optimize it to reduce my cost while not impacting my net promoter score.” And then having process mining and analytical AI and everything working automatically under the covers so that business-users don’t need to know that they’ve deployed process mining or AI models or anything.
They just know that this system is hitting them with an email every week saying, “Hey, by the way, I found a bottleneck. I’m recommending this fix. Click here to approve it and I will go ahead and get it done.” That is to me, just transformative in terms of how businesses, now, as you say… I think disruption is going to come faster and faster. And the businesses that win are going to be the ones that are continuously changing and continuously optimizing, and continuously responding to those changes. So I think we’re pretty excited about being able to take the core of what Pega’s already had and use this next generation of generative AI to stitch that all together to drive really dynamic change for our clients.
Daniel Newman: Yeah, I think that’s a great place to leave it, Don. It’s a really exciting time. It’s a bit overwhelming for most companies, most leadership. I remember a chart where I was showing from the cotton gin to modern digital and basically that we’ve seen transformation eras shorten like half-lives. And that running a business now, the amount of time from disruption to disruption is getting shorter. And what we’ve actually entered is almost, we talk about things like agile. We’ve almost entered the agile era of digital because there is no real stop time where this new thing is going to change. It’s just changing all the time.
Don Schuerman: And I think that’s fundamentally, for years we’ve been talking about digital transformation and I think for a long time I would hear people talk about it like it was a project that you were going to start and get a team together and then you were going to have the end of your digital transformation project, and there’s be this big celebration. We did it, we digitally transformed. Whereas I think what people are realizing is that digital transformation is just about accepting the fact that you are going to have to constantly change. That you are going to have to constantly be disrupting your markets if you want to stay ahead and build an organization and the skills, and the technology stack that lets you do that.
Daniel Newman: And I think that’s a great way to end it. There is no end to digital transformation.
Don Schuerman: There is no end.
Daniel Newman: And for the leaders out there, I hope you hear that. I think you get it, but if you don’t, it’s going to get you. So I know there’s scaryism to end my show, but Don Schuerman, CTO at Pega, I want to thank you so much for joining me today here on the Futurum Tech Podcast. Always good to get your insights.
Don Schuerman: Thank you.
Daniel Newman: All right, everybody, hit that subscribe button, join us for all our Futurum Tech Podcast. We appreciate you being part of our conversation. Feel free to add a comment, share it with your friends, but if not, hopefully you’re learning, you’re spending time with us and that’s what we love here. See y’all later.
Author Information
Daniel is the CEO of The Futurum Group. Living his life at the intersection of people and technology, Daniel works with the world’s largest technology brands exploring Digital Transformation and how it is influencing the enterprise.
From the leading edge of AI to global technology policy, Daniel makes the connections between business, people and tech that are required for companies to benefit most from their technology investments. Daniel is a top 5 globally ranked industry analyst and his ideas are regularly cited or shared in television appearances by CNBC, Bloomberg, Wall Street Journal and hundreds of other sites around the world.
A 7x Best-Selling Author including his most recent book “Human/Machine.” Daniel is also a Forbes and MarketWatch (Dow Jones) contributor.
An MBA and Former Graduate Adjunct Faculty, Daniel is an Austin Texas transplant after 40 years in Chicago. His speaking takes him around the world each year as he shares his vision of the role technology will play in our future.
