Today’s interview is with Don Schuerman, the CTO at Pegasystems, an industry-leading low-code platform for AI-powered decisioning and workflow automation. Don joins me today to talk about the upcoming PegaWorld 2025 (PegaWorld is taking place at the MGM Grand in Las Vegas from June 1st to the 3rd), Agentic AI (obvs.), how some organizations are implementing agentic solutions and it is causing chaos, a better way to do it, Pega Blueprint, legacy transformation and how to get it right.

This podcast is sponsored by Pega.

This interview follows on from my recent interview – Starting a resolution revolution – Interview with Tom Eggemeier, Adrian McDermott and Matthias Göhler of Zendesk – and is number 538 in the series of interviews with authors and business leaders who are doing great things, providing valuable insights, helping businesses innovate and delivering great service and experience to both their customers and their employees.

Here are the highlights of my chat with Don:

• These days, CTO really just means Chief Translation Officer.
• You can’t run a tech conference these days without talking a little bit about AI.
• I think we are all being forced in one way or another to consider agents.
• What we’re going to really focus on:
  • What are the real steps an enterprise can take to get value from and deploy this technology at scale? Once you step past the hype, what do you really need to do?
• There are two big themes that come out of that for us.
• First, is the accelerating need for legacy transformation. If we want to use all this cool AI and agent stuff, it’s not going to work if we’re still got our data trapped in a bunch of old systems running on Lotus Notes and mainframes and those kinds of places.
• We’ve been having a lot of success with clients using a tool we call Pega Blueprint that can generate and design workflows for you on the fly.
• It uses AI and a library of best practices to design workflows.
• It’s also really good at taking the documents and screenshots and code snippets from your existing legacy apps and using that to construct what a modern version of that app would look like in a couple of minutes.
• We’re seeing the real acceleration of clients using blueprint to drive legacy transformation.
• We’re going to be leaning into how you do that, what new capabilities we’re adding into blueprint and the overall architecture to support that.
• Secondly, we’ll be looking at how we believe enterprises can get real value in a safe way from AI agents.
• The problem we know with prompts is, one, it actually takes some fine tuning to get them right. And even if you get them right, they’re not necessarily predictable. Now, how do you test that prompt when OpenAI et al are releasing new versions of their models every couple of weeks?
• How do you do that when you don’t have just one or two of these problems? But you have thousands of them running around the enterprise. I think that becomes a pretty unscalable approach and introduces a level of risk and non-determinism that most enterprises for a lot of their core business processes aren’t going to aren’t going to accept.
• We actually think there’s a better way to do that, which is. Instead of trying to run your agents off prompts, run them off workflows.
• But the great thing is, with Blueprint, I can design a workflow faster than anybody can design a good prompt. So we can use Blueprint and the reasoning power of agents embedded inside of Blueprint to design the workflows, and then use the workflows to power the agents at run times. You get the power of an agent, but you get the predictability of a workflow based into it. And we’re going to be showcasing a lot of that technology at PegaWorld.
• If you don’t take into account your present anchors then you won’t get it right.
• A lot of people are talking about this technology and it’s almost like leaping forward without actually being cognizant of where you’re coming from.
• The thing that sets me apart, the actual definition of my IP, is the processes I use to fulfill that service.
• We should start thinking about legacy systems as not legacy but heritage systems. Because in many ways, if you’re an organization that’s been in business for decades, those legacy systems do contain some of the heritage, either data, transactions, best practices, processes etc.
• So, our thinking is, how can you take that heritage and not just replatform it, but learn from it, improve it and transform it so we can carry that heritage forward to a modern world, not just technically, but also in terms of the processes and the experiences I want to deliver to customers.
• One of the reasons you tend to know when you walk into a contact center that there is an opportunity for system improvement is you see a bunch of Post-It notes stuck around someone’s screen.
• You could take a photo of all those Post-It notes and begin to distill down what’s inside of that so that you could then pull that into a more modern system that encompasses all of that learned IP so that it’s embedded in the enterprise, but hasn’t made it into my system yet.
• It is really important that we acknowledge that there is embedded expertise and IP in the human supercomputers that exist in your organization and that have managed these systems for the longest time.
• A lot of the AI that actually gets deployed at the enterprise is statistical AI based on machine learning algorithms that have been around for decades. And that stuff still generates billions of dollars in increased
• revenue and value and efficacy.
• The power is not, in my mind, that Gen AI eats everything and everything just becomes an agent running a large language model.
• Rather, it’s how does the large language model enhance and make the statistical AI models better? How does the large language model enhance our workflows that we’re orchestrating?
• How do I go from either science fiction or science experiment to actually something that I can use in my business. That’s the conversation we want to have because that’s the value moment.
• Anybody can go to Pega.com and try Blueprint out. It’s a great way to really experience what’s possible here.
• Don’s best advice: How are you using AI to rethink some of your core workflows? Because at the end of the day, the things that deliver on your employee experience, the things that deliver and get work done for your customers, or the workflows, the processes in your business. I think we have an opportunity to very rapidly rethink some of those processes and make them much more customer-centric, much more outcome oriented. And doing that sets you up to then be able to plug in agents and take full advantage of some of the optimizations we’re going to get with AI.
• Don’s Punk CX brand: Pickup Music

About Don

Don Schuerman is CTO at Pegasystems, responsible for Pega’s industry-leading low-code platform for AI-powered decisioning and workflow automation. He has over 25 years of experience delivering enterprise software solutions for Fortune 500 organizations, with a focus on digital transformation, mobility, analytics, business process management, cloud and CRM. Don has led enterprise software implementations and provided technology and architecture consulting to senior business and technology executives from Fortune 500 organizations, including American Express, Citibank, JP Morgan Chase, and BP. Don holds a BS in Physics and Philosophy from Boston College.

As a longtime improv comedian, Don believes that “Build for Change®” is more than a tagline – it’s a mindset, and that great teams bring adaptability, diverse voices, and a sense of fun to all they do.

Check out Pega, say Hi to Pega and Don on X (Twitter) @pega and @donpega and check out Don’s LinkedIn profile.

Credit: Photo by Soheb Zaidi on Unsplash