Our Approach

By zooming immediately in on a single workflow, we can provide concrete improvements while also identifying systemic problems. But we need a workflow that touches multiple teams, involves real decisions, runs often enough that improvements compound, and produces something the organization already wants to improve. For this walkthrough, we chose a new product launch campaign at a global pharmaceutical company we supported. Eight months from the end of clinical testing through to a global launch. Seven teams involved. Three parallel workstreams. Five distinct technical systems.

We pull three to five recent instances through the organization's own systems. For this example: Clinical results packages. Strategy briefs. Kickoff agendas. Creative messaging documents. Design files. GTM briefs. Launch planning docs. We build a picture of how the work actually progressed through the organization. In this case, we found the clinical results took weeks to reach the marketing team. The creative strategy was approved but the direction given to external teams didn't carry the rationale. By the time the GTM brief landed, it had passed through four handoffs and lost its edge. The results from the previous launch? Nobody downstream ever saw them.

A full day with the cross-functional group in the room together. We put the workflow map on the wall with the artifacts we found. The group corrects it, adds context, fills in what the documents couldn't capture. In this case: clinical results were held up by Legal. Team resourcing was a constant bottleneck. CMO approval took five weeks. A key product stakeholder changed mid-process and the reasoning behind the strategy left with them. The approval process was unclear to everyone involved. Key team members got pulled off the project and the rework was significant. Midway through, the product team rewrote the strategy. Timelines shifted across the board. Regions pushed back on the strategy. SEO demanded prescriptive copy that overwrote the clinical messaging. And then: strong collaboration with the regional teams, once they were finally brought in.

The workshop captures the group's story. Observation captures what individuals actually do. The gap is where the most important findings live. In this case, we found seven patterns that the workshop alone couldn't surface:

In addition to task analysis (i.e., what humans vs AI should do), every observation gets coded against our taxonomy of intelligence loss: what kind of knowledge was lost, how severe, how often it happens. Then each finding gets connected through a chain of cause and effect. The CMO-Product conflict not being resolved meant the strategy kept shifting, which meant creative rework, which meant the campaign launched with weaker messaging than the clinical data supported. The global teams getting assets late with no context meant the secondary market launch underperformed. We also do the political work. For every proposed change, we think through who benefits, who feels threatened, and whether their concerns are legitimate.

This is the moment when leadership decides how to proceed and how to dose the necessary changes. In this case, five flights (a bounded round of changes) emerged from the findings:

We measure against the metrics we agreed upon at the start. Cycle time, rework rate, regulatory rejection rate, campaign performance, approval turnaround, etc. Some results show up in weeks. Others take quarters. We're straightforward about which is which. The deeper question is whether the team can now see intelligence losses on their own. Can they spot a broken handoff in a Monday meeting? Can they run a version of the workshop themselves? The goal is a permanent capability, not a consulting relationship.

Our initial engagements tend to evolve into two directions, often both at once. First: pick the next workflow. The product team's market research that was disconnected from the campaign workflow? It turned out to be disconnected from the medical education workflow and the sales enablement workflow too. Second: escalate. The CMO-Product conflict, the resourcing problem, the unclear approval authority: those weren't workflow problems. They're organizational problems that showed up in the workflow.