Getting AI Right: Architecture, Investments, and Tough Choices

How should leaders think about their tech stack and architecture in the agentic era? 

There's a lot of stress inside companies right now around architecture. The technology is advancing at such a rate, and the milestones are so unpredictable, it's really hard to have a definitive and solidified plan around architecture. So the name of the game for companies is going to be flexibility. We need to build architectures that can adapt as the technology evolves, and companies need to get comfortable that they're going to make some bets and make some mistakes and then have to backtrack a little bit. 

I think we're in a new era of architecture where things are going to be a little less solidified, a little less predictable, but I think that's what's required to move with speed in the generative AI era. 

When it comes to generative and agentic AI applications, buy or build? 

I think everyone is rightly biased to buy AI, not build it. And that makes a lot of sense. You're drafting off an ecosystem that is investing to put capabilities in software layers and all of the technology stack. But if you're just buying what everybody else has, you're not creating competitive advantage. So I think, increasingly, companies are going to have to find places where they have bespoke, unique workflows that make them different and build something unique around those. I think that is how competitive advantage is going to be created, even as you bias towards buying, not building. 

What trade-offs do companies need to make to successfully apply AI? 

Well, there's a big one in technology organizations. Legacy infrastructure and stacks have to be modernized. At the same time, CIOs are under enormous pressure to find cost savings to reinvest in generative AI. So we're seeing a pretty profound shift in IT departments from run the business technology spend to change the business technology spend. And that requires sort of a relentless pursuit of optimization and rapid application of this tool set. 

I think modernization has always been on the agenda of many of our clients. I think it's now an imperative if you're going to find the funding to invest in generative AI.

These captions were automatically generated.

Part one of Chuck’s talk

What to Expect from AI in 2026

Boardrooms are abuzz with fear of an AI bubble. Bain’s Chuck Whitten breaks down the reality.

Part three of Chuck’s talk

Preparing for AI’s Next Wave: Agentic, Quantum, and Managing Change Throughout

Goodbye baton pass, hello “bilingual” teams. Bain’s Chuck Whitten outlines the new approach to change management.

Part four of Chuck’s talk

AI’s Evolution: What Good Looks Like Now

Bain’s Chuck Whitten explains what top leaders get right—and how others can follow.