AI is the great equaliser. That does not mean everyone is equal.

There is a version of the AI story that goes like this: development is now democratised, anyone can build, the gap between a solo founder and a fifty-person engineering team has closed, and the cost of production has effectively hit zero. That version is not entirely wrong. It is just incomplete.

Access to capable AI coding tools is genuinely broad. GitHub Copilot, Claude, Cursor, and their peers are available to anyone with a credit card. A developer two years into their career can now produce output that would have taken a senior practitioner twice as long a few years ago. Scaffolding a Next.js component, wiring up a headless CMS integration, generating a typed GraphQL query, or standing up a deployment pipeline are all faster now in ways that are hard to overstate.

So the equaliser part is real. But what gets equalised is execution speed, not judgment.

The thing that AI cannot compress

There is a difference between writing code and making the right architectural decision. There is a difference between integrating a platform and knowing which platform should be in the stack at all. There is a difference between building a feature and understanding why a client's current approach will compound into a problem in eighteen months.

These are not small differences. In the digital experience space specifically, the consequences of getting these calls wrong are significant: costly re-platforms, brittle composable architectures that nobody can maintain, personalisation strategies that look right on paper but fail under real traffic and real content operations. AI does not protect against any of that. In some ways, faster execution makes the risk worse, because you can go further down the wrong path before the problem surfaces.

This is where solution architects, senior strategists, and experienced practice leads become more valuable, not less. When anyone can build quickly, the person who knows what to build, and what not to build, is the genuine differentiator.

What this means for how we work

At Gamma, we have leaned into AI-augmented delivery from the start. Our engineers use AI tooling across the full development lifecycle. We build on it, and we build with it. But the way we have structured our practice reflects something we observe consistently in client engagements: the leverage that AI creates is only as good as the judgment applied above it.

We do not measure the value of a senior solution architect by lines of code produced. We measure it by the decisions made before code gets written: the platform choice, the integration pattern, the data model, the delivery phasing, the risk profile. An AI assistant makes a good architect faster. It does not make a junior developer into one.

The same applies to strategy. A consultant who understands a client's operating model, their content governance constraints, their team capability, and their actual commercial problem will use AI to sharpen and accelerate their thinking. A consultant without that foundation will use AI to produce polished outputs that answer the wrong question faster.

The part the industry is getting wrong

The framing that worries me is the one where AI tooling is treated as a substitute for experience rather than an amplifier of it. Some agencies are restructuring around this assumption: fewer seniors, more juniors with better tools, lower cost base, same output quality. The cost base part is probably true. The output quality part is where it breaks down in practice, especially on complex programs.

Enterprise digital experience delivery is not primarily a production problem. It has never been. The hard parts are requirements that shift mid-engagement, integrations that behave differently in production than in documentation, stakeholder alignment across IT and marketing and procurement, and the accumulated judgment calls that determine whether a platform genuinely serves the business in three years or becomes a liability.

AI does not carry that weight. People do. 

Where this lands for clients

If you are evaluating an agency partner in this environment, the right question is not whether they use AI. Almost everyone does. The question is what sits above it. Who is making the architecture decisions? Who is responsible for the strategic framing of the engagement? Who will push back when the brief is pointing in the wrong direction?

The access cost of engineering has dropped. The value of the right engineering judgment has gone up. Those two things are not in tension. They are directly connected.