AI Workforce Transformation. A function rebuilt in 45 days.
A function rebuilt around AI in 45 days — the playbook we ran on our own teams before we ever sold it to a client.
TRANSFORMATION · PEOPLE + AGENTSWhat this engagement is
Tools don't transform functions; operating models do. This engagement takes one function — support, finance ops, content, QA — and rebuilds how the work flows: what agents do, what humans gate, what gets measured, and how the team's week actually changes.
We ran this playbook on ourselves first. Our own delivery organisation runs on the platform we sell — specs gated, code agent-assisted, releases signed — which is why the 45-day claim is a schedule, not a slogan: two weeks of mapping verbs and policies, four weeks of running the new model in shadow and then in production.
The deliverable is a function that runs differently on day 46 — with its people doing judgment work, its agents doing volume work, and its metrics reported from the graph.
The numbers behind it
What ships
Verb mapping
The function's real decisions, captured from live work — not from org charts.
Role redesign
Human attention repointed at exceptions, quality and improvement.
Agent rollout
Volume work moved to gated agents with confidence floors.
Shadow → production
The new model runs beside the old until the metrics say switch.
Enablement
Your team trained to author skills and tune policies themselves.
Outcome dashboard
The function's new metrics, live from the brain.
How the engagement runs
Five phases, forty-five days, one function rebuilt around AI.
Proof from production
We ran the playbook on ourselves first
“Before offering it to anyone, we rebuilt our own delivery, QA and support functions around AI in 45 days — eight production agents now run our operations, with every exception owned by a named person.”
Questions teams ask
Is this a layoff program in disguise?
No. The playbook redistributes work: agents take the volume, people move up the checkpoint chain to judgment and exceptions. Your team designs the target model with us — that's why it holds after we leave.
Why 45 days?
Long enough to rewire one function properly, short enough to hold the organisation's attention. Scope discipline is the feature: one function, measured against its own baseline.
What if our function is too messy for AI?
Messiness is what the baseline phase is for — and the messiest functions gain the most. The undocumented workarounds we surface in week one usually become the first workflows worth automating.
What happens after day 45?
The function is yours, with its measurement cadence and its workflows feeding your intelligence layer. Some teams keep us alongside for the next function; nothing about the model requires it.
We stopped buying projects and started owning capabilities.
Pick your function. Own the intelligence behind it.
Discover one opportunity, engineer one capability, deliver one measurable outcome — then scale.