A private equity firm

The firm's HR team came to us with a practical question: where in their work can AI actually lift productivity, and what are the first moves worth making?

The engagement

Six weeks of analysis, sequenced across several months and designed to take no more than a day of each stakeholder's time. The unit of analysis was the role; the unit of action was the task.

What we did

• Harmonised 21 job descriptions into a single task model
• Decomposed each task into discrete actions and scored them against four AI capabilities
• Validated the model with the firm's HR team and folded their feedback back through the analysis
• Built three tracks from one evidence base: use cases for employees, agent designs to commission, and a vendor shortlist for procurement

What they took away

• A prioritised evidence base for where to deploy AI first
• Role-level automation profiles and a task-level heatmap
• 63 use cases — three per role, anchored to each role's top three tasks. Each is a workflow loop with step-by-step guidance and ready prompts an employee can run today
• 23 agent designs — structured agent definitions sized against the function's work, ready to take into a build or commissioning conversation
• A scored vendor shortlist — 150 AI products evaluated against the function's work, top 20 recommended for targeted procurement discovery
• A single skills taxonomy across the cohort, usable across both AI prioritisation and mobility planning

The use cases are already going to employees across the function — a near-term path to AI adoption running in parallel with the longer agent build and vendor procurement tracks.

In numbers

Talk to us