AI Workflow Workshop
Teams mapped real work first. Only then did they design AI into it.
Context & Constraints
KPMG teams in Legal, Corporate, SALT and Tax Transformation needed to understand where AI belonged in their own work cells.
A lecture would not change the workflow. The gap was practical: teams needed to see their own process, identify friction, and decide where AI could help without pretending every problem was a one-shot prompt.
- The workshop could not be a lecture. People needed to work on their own processes, otherwise the adoption would stay abstract.
- The room had two opposite problems: people afraid they knew nothing about AI and people convinced a single prompt solved the whole workflow.
- Some outputs became real projects, but the confidential solutions themselves could not be named publicly.
“Real AI adoption starts when people map the work, not when someone teaches prompts.”
Architecture
I designed and facilitated a 12-hour hands-on workshop. Each group picked a real process, mapped the current workflow, broke it into steps, and translated the best opportunities into AI agents, prompts, prototypes or solution specs. Some outputs became real projects.
My Role
I created the methodology and facilitated the sessions, with support from other consultants during group work.
Results & Validation
The workshop worked because it served two audiences at once: it gave confidence to people who thought AI was beyond them, and it slowed down people who thought one prompt solved the whole job.
Stack
Building AI that has to work every time?
I design, ship and operate systems like this one.
Let's talk