For Anaplan partners, consultancies and PE value-creation teams
Anaplan + AI agents, as one person.
Anaplan's agent era has started: CoModeler, CoPlanner, Agent Studio. Every practice is about to be asked the same two questions by clients. Which numbers may an agent touch, and how do we prove its answers are right. Those engagements need two halves: the model, and the agents and controls around it. I do both, as one person, as an associate on your engagements.
01 · Who I am
I'm a Certified Master Anaplanner who also builds production AI systems for finance. CIMA-qualified, fifteen years in finance operations before I built systems, production AI since 2024. Anaplan work at Goldman Sachs, Astellas and Exscientia.
02 · Three proofs
Agents around a live estate
At Fox I built the AI agents that draft board commentary and variance analysis around a live Anaplan estate, in production, on one engagement.
A platform built on verification
Grounded runs single-tenant for finance teams: every answer cited to the exact cell, evals and guardrails on everything, an audit trail throughout, and designed refusal when the documents can't support an answer.
A public benchmark
The Board Pack Test: 34 documents, 25 questions, a deterministic grader anyone can re-run. Raw frontier models scored 66 and 71 of 81 machine-checkable points. The verified pipeline scored 81. The whole exam is published, so you can check the claim rather than take it.
03 · How I work
As an associate on your engagements. You own the client relationship; I bring the piece most practices are being asked for and don't yet have in-house: building the AI agents around the estate and proving their answers are right. I work to your delivery standards, document everything, and hand over cleanly.
One honest line: I haven't built on CoModeler or Agent Studio yet. The two halves those deployments need, the model and the agents and controls around it, I've done in production.
Available now. UK-based in Reading; London, remote or hybrid on-site.
If you want your bench upskilled, a one-day workshop for model builders on agent design, evals and governance.