A mid-sized reinsurer had invested heavily in a major technology programme. New dashboards. New workflow systems. A data lake that would change everything.
The dashboards sat untouched. The workflow system had more workarounds than actual workflows. Treaty data still lived primarily in spreadsheets, despite the expensive infrastructure that was supposed to replace them.
The project had delivered exactly what was specified. That was the problem. It had been designed around system capabilities, not around how work actually happened. Nobody had mapped how data and decisions flowed through the organisation before choosing platforms. The result was a digital façade atop manual processes - new user interfaces that slowed down experienced actuaries, and dashboards filled with data nobody trusted.
When we were brought in, the brief wasn’t “fix the technology.” It was “help us understand why smart people keep building things nobody uses.”
What we did
We started with something the previous project had skipped entirely: mapping how information and decisions actually move through the business - including all the workarounds, informal processes, and “I just email Sarah” moments that don’t appear in any documentation.
We call this Dataflow Design. Before making any technology decisions, we mapped every data flow, every decision point, every manual intervention. We identified which processes were actually complex and which were complex only because nobody had ever simplified them.
Then we worked backwards from decisions. For every capability being considered, we asked: what specific decision will this improve? Who makes that decision? What information do they need, and when do they need it?
Only after these flows were understood did we design solutions - and those solutions looked nothing like the previous programme’s output. They were simpler, more targeted, and built around how people actually work rather than how a vendor’s platform was structured.
Why it matters
In reinsurance, technology transformations fail more often than they succeed. The incentive structures are misaligned: IT teams are rewarded for delivery, consultancies for complexity, and vendors for licences. Nobody is rewarded for adoption or business impact.