AI Didn't Kill Services — It Made Them Worth More

There's an anxiety running under the surface of the services industry right now. Not quite panic -- more like a quiet unease that nobody wants to say out loud.

AI can write code. AI can draft proposals. AI can build infrastructure. It can do in days what used to take weeks.

So what's left for the people who do this for a living?

The Fear Has It Backwards

The assumption is that AI obliterates knowledge-based work, and therefore the people who sell knowledge-based work are finished.

But that gets it exactly backwards.

AI didn't destroy the value. It removed the busywork. And when the busywork disappears, what's left is what always mattered -- outcomes. Outcomes come from ownership, judgment, experience, context, and tooling. That's the hardest part to deliver and the hardest to find.

Here's what makes this interesting: those same consultants and engineers are empowered with the same AI tooling as everyone else. For the same reason a senior developer with AI outperforms a junior developer with AI -- experience compounds the leverage. So when it comes to delivering outcomes, who would you pick?

An AI isn't on the hook for the outcome. A human is. Someone still has to own the result -- understand what success looks like, navigate the ambiguity, make the hard calls when there's no clean answer. That's what services have always been about. We just couldn't spend enough time on it because we were buried in implementation.

And here's what's unexpected: validation has become the ultimate skill. Validating that something delivers on what it's supposed to deliver. Validating that the outcome is actually achieved, not just that the code compiles and the tests pass. These are the most important skills today -- and they will remain fundamentally human. Greater and greater techniques for automated validation will emerge, but judgment? Beauty is in the eye of the beholder. Someone needs to own that.

The Transcription Moment

From Note-Taking to Active Listening

Remember what meetings were like before AI transcription?

Half the call was spent scribbling notes. Trying to capture every detail. Worried about missing something important. Physically present but mentally somewhere else -- processing instead of listening.

Then transcription took that away. And something shifted.

We stopped writing and started hearing. We could practice active listening. Read the room. Follow the thread of what someone was really trying to say, not just the words coming out of their mouth.

We went from passive to present. From distracted to invested.

The notes didn't go away -- they got better, because a machine captured them perfectly. But we got better too, because we were finally free to do the thing that only a human can do.

This Is the Template for Everything

That same transformation is playing out across all of services work.

AI translates requirements into code. It compiles architectures. It handles the unadulterated, unglamorous work of building out the systems that have been built many times before -- the boilerplate Terraform, the standard pipelines, the patterns that used to eat entire weeks.

But someone still has to validate it. And what opens up is space for the work that actually determines whether an engagement succeeds:

That was always the most valuable part of the engagement. Now there's actually room to do it well.

Outcomes Over Hours

The Industry Is Evolving

The services industry has historically sold capacity. Hours, headcount, throughput. The deliverable was whatever could be produced in the time allotted, and the constraint was always how many people and how many hours.

That model is evolving -- and it's not just a hunch. Sequoia Capital recently argued that services are the new software -- that the biggest opportunity isn't selling AI tools (copilots), but selling AI-powered outcomes (autopilots). Their framing: for every dollar spent on software, six are spent on services. That market isn't shrinking. It's transforming.

When implementation gets fast, the conversation naturally moves from "how many hours will this take?" to "what outcome are we delivering?" A customer doesn't want 200 hours of Terraform work. They want a production-ready platform that meets their compliance requirements and ships in four weeks.

AI handles the boilerplate in a few hours. That means the next few weeks go toward translating requirements into reality -- ensuring the outcome actually matches the expectation, not just the spec.

Isn't that worth its weight in gold?

The Premium Is Judgment

From Capacity to Craft

When building is cheap, judgment is premium.

Sequoia draws the line between "intelligence work" (rule-based tasks AI handles well) and "judgment work" (experience-based decisions that remain human). The services industry is moving from selling intelligence to selling judgment. From capacity to craft. And that's what most practitioners got into this work to do in the first place.

The hard part was never writing the Terraform. It was helping customers ship their software reliably to their customers. That takes judgment. That takes presence. That takes ownership.

AI doesn't replace any of that. AI finally gives us room to do it.

The services businesses that thrive will be the ones that lean into this shift -- that use AI to handle the intelligence work so they can focus entirely on what an AI can never own: the outcome.

Want to talk about what this shift means for your team? We'd love to share what we've learned.