Bid Brief · Internal · For the Account Manager
DBT — AI Capability Programme
Zühlke proposal to the UK Department for Business & Trade · procured under DOS6
Status: Live bid — full response drafted
7 bid sections complete · 8 clarifications open
As of 19 Jun 2026
In one line: DBT wants to build real, lasting AI capability (skills + governance + culture) across ~3,200 staff — not buy tools. A near-identical £2.2M programme ("Meridian") failed here in 2024 and that ghost dominates everything. Our entire pitch is the antidote: "Proof in 90 days, not promises in 24 months" — we put ~£1.15M of our own fee at risk against change that actually sticks.
The deal
£7.5M
Envelope agreed in principle (24 months)
£7.35M
Our bid — £150K under, deliberately
£1.45M
Phase 1 · Foundation · 6 mo · fixed price
£5.90M
Phase 2 · Delivery · 18 mo · tapering
~£1.15M
Of our fee AT RISK against verified outcomes
Real month-6 go/no-go between phases — if Phase 1 proof isn't there, DBT stops having spent £1.45M, not £7.5M. A month-18 sustainability gate holds back fee until usage survives without us in the room. Outcome-linked fee doubles as the published KPIs the Procurement Act 2023 already requires at this contract size.
What they actually need
- Not a tech buy. Buying the skills, governance & culture to use AI well.
- The org: ~3,200 UK staff / 14 offices (+~220 overseas). Formed Feb 2023 by merging DIT + BEIS — cultures & tooling still not integrated.
- Uneven maturity: Advanced 12% / Intermediate 34% / Basic 38% / Low 16%. Real risk = a two-tier department.
- Hard deadlines: measurable proof by Dec 2026; demonstrable improvement in the April 2027 annual report (Perm Sec's public commitment).
Why we win / our angle
- Embed in real work — redesign actual workflows in real directorates, not standalone training.
- Govern as we go — policy grows from live use cases, so it's actually used.
- Design our own exit — named DBT owner per workstream; team tapers ~40% by month 24. We get cheaper as they get better.
- Willing-first: start with Investment + Data & Analysis (fast proof), approach Trade Policy & Legal last, with proof in hand.
- Fee tied to adoption that persists — verifiable by the client.
The ghost in the room — "Meridian" (must not repeat)
Meridian Strategy Group · 2024 · £2.2M · FAILED
- 180-module training catalogue + a governance framework
- 6 months later: 94 of 3,200 still "learning"
- Framework viewed 234 times, adopted by no one
- 100-champion network → 12 left; 71% used tools exactly as before
- Paid for activity delivered; value evaporated on exit
Our inversion — paid for change that lasts
- Capability built into real work, not alongside it
- Independent baseline first (Ekaterina) — before we have a stake
- Reuse Meridian's 2022 segmentation — refresh, don't re-run
- £1.15M fee released only on verified, persistent adoption
- Built against DBT's own 5 post-mortem recommendations
Who matters (and their stance)
Sir Michael Okafor— Perm Sec · ultimate sponsor (April-2027 promise)
Helen Marsh— Dir Digital Transf. · CHAMPION, staked her credibility
James Whitfield— Dir Investment · SUPPORTIVE · our first win
Richard Afolabi— DD Corp Services · SCEPTIC · owns procurement / VFM
Sarah Chen— Head of HR · NEUTRAL · win early
Dr Priya Singh— Dir Trade Policy · RESISTANT · accuracy-first
PCS / FDA unions— WATCHING · job-cut fears ("augment, not replace")
Afolabi is the gatekeeper — his test is "is this recycled?" The bid answers him on every page.
What could kill it
L4·I5Meridian 2.0 — activity delivered, behaviour unchanged. This is the default outcome here, not the edge case.
L4·I4Staff resistance — Trade Policy & Legal (2024 accuracy incident; AI banned for legal analysis).
L3·I5Union escalation if read as a job-cuts programme — can stop it politically.
L3·I5Sponsorship one-person-deep — enthusiasm = Helen Marsh only. Need contractual Director sponsors.
L2·I3Differentiator backfires — DOS6 mechanics may not allow outcome-linked fee → fixed-fee fallback ready (Q7).
Our team (SC-cleared core + associate bench)
Laurence Smith — Engagement Partner
Dan Bratton — Programme / Delivery Director
Manu Mukundan — Technical & Data Lead
Ekaterina Saparova — Discovery & Measurement (owns the baseline)
Gayatri Potawad — Capability & Engineering Lead
Henry Russell — Enablement & Internal Tooling (adoption dashboard)
What you (AM) should track
- 8 clarification questions are open and several are deal-hinges — esp. Q2 (will DBT contractually commit Director sponsors?), Q7 (can DOS6 accommodate outcome-linked fee?), Q1 (is the 2022 segmentation reusable — cost/time hinge?).
- Outcome-linked fee needs Corporate Services / commercial buy-in — it's unusual for this framework. Fallback exists but the differentiator weakens without it.
- Tone is the strategy: humble, practitioner, DBT-specific. Anything that reads generic loses with Afolabi.
- Watch the budget governance: £7.5M is "in principle," annual confirmation needed, >£500K needs Deputy Secretary sign-off — phasing is designed around those approval points.