Compelion
An AI consultancy asked us to think about their own AI, not their clients'. Here's what changed.
saved per proposal once the generator went live
from kickoff to first internal tool in production
AI workflows shipped against weekly-hours-saved, not novelty
on-brand outputs after voice guardrail rollout
What was breaking before us.
Compelion's job is helping organizations turn ideas into delivered AI projects. The harder question was the one most consultancies avoid: how should we use AI on ourselves, day-to-day, before we tell clients to? They needed a sparring partner, not a vendor.
What we did.
The work itself, in order. No padding, no buzzwords.
- 01Mapped where AI would actually move the needle inside their own delivery, proposal scoping, knowledge capture across engagements, project handovers
- 02Recommended a small, opinionated stack: a knowledge assistant trained on past project artefacts, a structured proposal generator, and a brand-voice guardrail for client-facing comms
- 03Sequenced what to ship first based on weekly hours saved, not novelty
- 04Documented the recommendation as a working playbook the team could execute internally
Other engagements.
The Bootroom Collection
Two stores, one Shopify front, one Odoo back office, and stock that finally agreed with itself.
Read
Liverpool FC Store
Official licensed retail running on the same operating system, POS, online, stock, and finance all on one screen.
Read
Mo's Joint
Running a multi-location burger brand on one screen, POS, kitchen, stock, and supplier orders that finally talk to each other.
ReadWant a sparring partner for your AI roadmap?
Tell us about where you are. We'll come back with how we'd build it, fixed scope and fixed timeline.