Most engineering teams are still sized for a constraint that AI is removing. Headcount has always been the main lever for output, which is why squads, ceremonies, and roles are built around the number of people available. Take away that constraint and the questions worth asking change, starting with how big a team actually needs to be.
The reasoning behind the traditional squad size was straightforward. Teams needed enough people to cover the required range of skills while sustaining a predictable delivery rhythm, and the ceremonies that grew up around that model were all designed to coordinate human effort and keep work flowing through the pipeline.
What AI agents are now changing is that underlying calculation. The output available to a skilled engineer working with the right tooling is growing considerably, and the team structure designed around the old output ceiling no longer reflects what a well-equipped team can actually achieve.
The sharpest version of that question is whether a team is better served by two teams of six or six teams of two. It is deliberately provocative, but it surfaces something real. With agents doing the execution, what a team can achieve depends far more on the quality of thinking at the top than on how many people are in it.
The Three Amigos model from Agile, which brings together business, development, and quality perspectives to define and validate work before it is built, is a useful lens here. With AI agents available to those three people, a small senior team can now orchestrate a scope of work that would previously have required a much larger group. It is one of the more concrete pictures of how a smaller, AI-enabled team might actually operate.
One pattern we are starting to see across ClearPoint's engineering teams and client work is a move toward greater generalism. Engineers who previously specialised in one domain are finding that AI agents handle enough of the domain-specific complexity that they can work more confidently across the full stack. It is one way the role could evolve, and it is already shaping the kind of engineers we look for, those who can engage with a problem from every angle rather than just their corner of it. This does not mean specialism disappears. As the execution work generalises, it is the judgment and oversight work, the parts that decide what good looks like, that becomes the place specialist depth matters most.
For as long as software teams have existed, there has been a gap between what the business envisions and what engineers build. Product and engineering have always operated at some distance from each other, and AI is now closing it.
Product owners who can now prototype ideas using AI tools arrive at the engineering conversation with something concrete rather than a requirements document. Engineers who carry strong commercial awareness can engage with those prototypes critically and creatively. When those two things happen together, the cycle from idea to working software compresses considerably, and the work that comes out of it is typically better for having had both perspectives involved from the start.
A common concern about AI's impact on engineering teams is that specialist roles will become less important. The pattern ClearPoint is seeing across its client work points in the other direction. Quality engineering and design disciplines are becoming more critical as AI agents take on more of the execution work, because organisations need experienced humans to oversee what those agents produce and ensure it meets the required standard.
This puts quality engineers, designers, and platform specialists among the most valuable people in a modern engineering organisation. As AI takes on more of the execution, the roles that direct, validate, and shape that work only grow in influence.
After decades of laying technical foundations, AI is now giving engineering the space to operate as a creative discipline, not just a technical one. As AI takes on more of the production work, what is left for people is the part that was always the most human, deciding what is worth building, and shaping how it serves the people who use it.
For engineering leaders, that shift creates a real opportunity to build teams around capability and creativity rather than headcount and coverage. The question worth considering is not how many people are in the team, but what that team is actually capable of, and how well the structure around it is set up to support that potential.