The AI narrative is constantly framed as a way to revolutionise and improve how we do work. But it’s not as simple as this. We use AI tools every day and we are big believers in them but without the right strategy, expertise or shared processes in place, AI can also silo teams and cause distractions.
When everyone builds their own thing
It’s very easy to pick up an AI tool, get locked in and build something useful, all on your own. That energy is a good thing to see. The risk is that everyone ends building their own version of things in the background, and sometimes solve the exact same problem in parallel without ever realising it. Suddenly, the tool that was meant to connect teams ends up doing the opposite.
AI needs an expert at the wheel
None of this is an argument against AI. The gains we have seen are real. Our developers, for example, have taken to these tools quickly, and they are producing work to a higher standard and at a faster pace than before.
What has not changed is the need for people who know what they are doing. The skills look different now, but the same critical thinking is required: prompting the tools well, and sense checking what comes back rather than taking it at face value. You still need experts to steer AI tools in the best direction. AI makes good people quicker and sharper. It does not replace their judgement.
There is a foundational point too. These tools are only ever as good as what you feed them. If the underlying data is not solid, you are building on an incomplete foundation, and the tool will happily fill the gaps by making things up. Getting the basics right first is a very necessary foundation to building anything with AI.
Focus on coordination, not caution
The answer is not to slow people down nor hold the tools at arm's length. It is to be deliberate. For us, that has meant putting together a prioritised plan for AI rather than letting a hundred experiments run wild. Each initiative has a name against it, someone who owns it and sees it through. We are choosing a small number of initiatives to work on properly rather than chasing everything at once. And we are making sure people reach for the same tools and share what they build, so that two people are never unintentionally solving the same problem in parallel.
Collaboration is not automatic
The promise of AI making teams more collaborative and productive is real, but it is not automatic. It takes a little more planning than people think to set strong foundations and introduce shared practices. That’s worth considering, whether you are rolling AI tools out across your own team or choosing a partner whose team has already worked out how to use them well.
Emma heads up delivery and operations here at Unified. She sits on the leadership team and has the responsibility to keep everything ticking along as clients expect.
