What we do

Complex, risky programmes are where we run the full arc: legacy platforms, regulated environments, decisions too big to rest on estimates.

Discovery goes deeper than a traditional one because AI does the reading. Agentic analysis works through hundreds of thousands of lines of code and years of tickets, and Discovery Dialogues hears from every stakeholder in scope within the same few weeks. Planning then runs alongside a working prototype that users test, so the decision to invest is made on demonstrated capability and sizing built from your own codebase.

The build itself is where our way of working pays off. The methods that proved the prototype carry straight into the production build, which runs to a fixed milestone plan with real users steering the product from the first release. In regulated environments we lead a joint team that includes your own engineers, so accountability sits with us while the capability and compliance knowledge stays in your organisation. In less regulated settings we can build and deliver the whole system ourselves, usually inside four to six months.

Adoption programmes usually stall for the same reason: nobody knows where people are. Surveys get gamed and workshops hear from the loudest few.

Our programmes start from a measured baseline. Every person in scope has a voice conversation with Discovery Dialogues, and because the same AI interviewer holds and scores every conversation against one capability model, the results are comparable across a whole workforce; we have interviewed close to 200 people in a week, 138 of them in a single day. From there, experienced change professionals run the programme itself, finding champions, coaching teams, guiding tooling decisions and supporting leadership. Months later we interview everyone again and show you what moved.

At Cefalo, a Norwegian software consultancy, this cycle took 138 paired developers up an average of 0.9 levels on our five-level capability model in eight months, with 80 per cent of them improving. Cefalo drove the change; the baseline, the guidance and the like-for-like measurement came from this programme.

Plenty of organisations need experienced AI leadership now and cannot wait for a six-month search, or cannot yet justify the headcount. We provide it on demand: guidance sessions with leadership, a seat at your AI task force, advice on adoption, tool choices and engagement, and help shaping a transformation plan your teams can execute.

The advice comes from practitioners who build and ship AI systems every week, so it stays concrete: which tools, in what order, with what evidence you should expect at each step. We also brief boards, leadership groups and industry audiences; recent talks include "Agentic AI in development: the reality" and "AI in software development: what actually works".