How AI Is Reshaping the Way Enterprises Build Software - With Tim Sears of HTEC

**Daniel Faggella** (0:12)
Welcome, everyone, to the Emerge AI in Business Podcast. Today's guest is Tim Sears, Chief AI Officer at HTEC. HTEC is a global engineering firm focused on AI-centric software and hardware development, working across financial services, medtech, automotive, telecom, and enterprise software from more than 20 engineering centers. In this conversation, Tim examines how AI is reshaping the software development life cycle, and why the most important shift has nothing to do with individual productivity, and everything to do with how teams work together. He argues that software development is turning out to be the killer app for AI, and that understanding exactly why that is true is the clearest lens available for how AI will eventually transform every other business function. Today's episode is sponsored by HTEC. AI is moving fast. New tools, new research, new use cases every week. Emerge synthesizes what matters most, so senior leaders and practitioners can stay ahead without getting buried. Join the more than 85,000 subscribers and get the most useful AI business insights delivered to your inbox. Visit emerge.com/addone. That's emerj.com/adone.
Now the conversation with Tim Sears.
Tim, it's great to have you on the Emerge AI in Business Podcast today.

**Tim Sears** (1:54)
Hi, Yolande, thanks for having me.

**Daniel Faggella** (1:56)
I'm excited to pick your brain today. I want to get into the real stuff real quickly. You have described AI deployment problems in the past as more of a people problem than a tech problem. I like that. I want to dive into that first.
It's quite a striking place to start, especially when our topic is software development life cycles today. From where you sit today, what has actually changed about how enterprises build and ship software, and where are most of them still running the same playbook they've been running for the last five years?

**Tim Sears** (2:29)
Pretty big question. I mean, I would say a lot is changing. We talk about change, you know, old versus new. It's pretty clear that the tools, you know, AI in various forms, various modes, offer tremendous potential for productivity gains. We haven't figured everything out about how to put them to work the best possible way.
It's pretty clear that what looked like best practices 20 years ago are being outmoded, but we don't have, you know, we don't have a gold standard of how things should be done today.
That's in the process of being developed. In olden days, you know, say three years ago, we had, you know, requirements gathered, which is a pretty lengthy process, and it had to be done super carefully so that your precious software engineers could focus on doing exactly the right thing, iterating through things like Scrum and, you know, Agile. And then there's a whole variety of deployment and hardening tasks around deployment, and everything around that whole process is being affected. Today, we see a wide range of comfort and skill in terms of using AI tools and fitting in new ways of working, and it's having a tremendous impact on individuals. And I think the next phase is really about figuring out ways for teams to be highly effective using AI. Software development, business in general is a team sport. AI is helping certain players do performance enhancing things, but we need to, the teamwork aspect of that is really going to be, AI needs to become a catalyst for teamwork in the future. We're really in the process, the whole industry is in the process of figuring that out right now. So in a couple of years, I would expect the landscape to be quite a bit different. We could be maybe a lot more prescriptive.

**Daniel Faggella** (4:05)
I know that you've spoken publicly about that gap between AI experimentation and production deployment. Does that gap show up differently in engineering teams than it does in other parts of the business?

**Tim Sears** (4:17)
Well, I think looking at engineering teams is really going to give us all the evidence about how to figure out how to use AI more broadly. Maybe this will sound like a little bit of a departure, but I think it's quite relevant. Software development itself is kind of turning out to be the killer app for AI.
And all of the hurdles and benefits of that actually are illuminating. It's very instructive to understand exactly why that's happening. First thing you need to know, and everybody kind of does, is that if you've sat down with a chatbot, for example, and typed something in, you know, you get this pretty interesting, it's pretty good, but is it really what you need? Is it really correct? So the output of LLMs, rightly, is not totally trusted. Obviously, it's getting better, much better, and fast, but it's not completely trusted.