Rebuilding Around a Hybrid Workforce

Every firm reading this has access to the same AI models we do. That’s the first thing I’d tell anyone trying to figure out how to compete in this environment: the technology itself was never going to be the differentiator. It’s already table stakes, and it will only become more so.

What will actually separate firms is a willingness to rebuild the business, not bolt AI onto the edges of it.

I see a lot of firms treating AI as an efficiency layer. They apply it to a handful of use cases, they shave time off a few processes, and they call it transformation. I understand the appeal. Those wins are real and they’re immediate. But they’re also superficial, in the sense that they don’t touch the actual structure of the business. And structure is where the real advantage lives.

The Hybrid Workforce is Already Here

Here’s the concept I keep coming back to: the hybrid workforce.

For as long as professional services has existed, the workforce has been people. You bill people out at a rate, or you staff a project at a fixed fee, and your entire cost structure and pricing model is built around human hours. That era is ending. The workforce of the future is people, software, agents, and tokens, all of which carry their own costs, all of which need to be planned, staffed, and priced as part of the same system.

I have not met a single customer yet that has actually built their business around this idea. Not one. Everyone is still running two separate mental models: how do we manage our people, and, over here, how do we manage our AI spend. Those need to become one model. If you’re forecasting a project, you should be forecasting the token cost and the agent cost right alongside the human cost, because that’s genuinely what it’s going to take to deliver the work.

This Has to Touch Every Process, Not Just One

This is harder than it sounds, and I want to be honest about that. It’s not a matter of adding a line item to a budget template. It means bringing this thinking all the way back to the center of the business, and then touching every process that connects to it. Your pricing model changes. Your resourcing model changes. Your margin analysis changes. Your client conversations about scope and deliverables change. If you try to do this at the edges, tweaking one process at a time, you’ll spend years struggling with it instead of actually transforming.

The Race to Zero

I think about this in terms of levels, too. Level one is the chat interface everybody got introduced to AI through. Ask a question, get an answer, move on. It’s useful, and it’s also already commoditized.

Level two and level three look completely different. At these levels, the technology isn’t just answering your questions, it’s taking action, assessing a situation, and actually eliminating work. Not doing the same work faster. Eliminating it entirely, so your team’s time gets reallocated to something that actually requires a human.

If you’re only operating at level one, you’re not building an advantage. You’re running the same race as everyone else in your industry, and that race has an end point I call the race to zero. Once everyone has access to the same basic efficiency gains, there’s nothing left to squeeze. You end up competing purely on price, racing to be the cheapest version of a service that looks identical to your competitor’s.

Where the Real Advantage Lives

The firms that will actually separate themselves are the ones using these technologies to enrich what they deliver, not just shrink what it costs to deliver it. Smarter analysis. Better judgment applied faster. A higher quality outcome for the client, not just a faster invoice. That’s a value proposition your competitor can’t just copy by buying the same tool you did.

So if you’re a firm leader trying to figure out where to start, my honest advice is to stop asking which use case to automate next and start asking a bigger question: what would our business look like if we assumed, from the ground up, that our workforce includes agents and tokens as first class citizens alongside our people? Once you can answer that, you’ll know which processes actually need to change, and you’ll be building something your competitors can’t replicate just by subscribing to the same AI models you’re using.