Sprint Zero.
The London AI Week.
An immersive week of Creativity powered with AI. Five squads, ten hand-picked builders, four intensive days in London — think tank, build, pitch, deliver. Each squad ships a working POC, a Product Plan, and a clear next step. The AI Lab's first build cycle — landing a month before the July Governance Session arms the decisions.
An immersive week of Creativity powered with AI.
Five squads work in parallel — think tank, build, pitch, deliver. Each one with clear time, clear direction, and the same bar: a working POC, a Product Plan, the next steps named, the work owned. Hands-on, in the same room, four intensive days. Creativity is the headline. AI is the amplifier.
This is the AI Lab's first Accelerator Sprint, run as the pilot that locks the bi-monthly cadence in. The Foundation set the bar explicitly — Appendix A says "it is not a social hackathon, not a brainstorm, not a show-and-tell event. It is not pizza and energy drinks." Sprint Zero is the proof that the format can clear that bar at scale.
The week lands a month before the Foundation's July Governance Session. That's by design — Sprint Zero is the build that arms the Governance Session. SLT walks into July with five validated POCs, five Product Plans, a Marketplace direction, and a film already in hand. The Sprint produces the material. Governance makes the calls.
One event, two names — each doing different work. Sprint Zero sets the bar. The London AI Week sets the energy.
Two names, one event. DS doesn't enter the BCG Creative AI space cautiously — it sets the standard, and Sprint Zero is the moment it does.
Five mini-squads. Five POCs. Five plans to ship.
Ten people, five mini-squads, one bar. Every squad walks out of London on Friday with the same five things:
-
01A running POCDemonstrable, on the screen, with an owner.
-
02A Product PlanWhat it does, who it's for, criteria, scale/iterate/set aside.
-
03Next steps · time neededWhat it takes to get to v1. Sized in weeks, not hand-waved.
-
04What's onThe commitment. Who picks this up after Friday. Named, not implied.
The Foundation calls this "the evaluation standard". If a squad can produce a credible POC, it ships. If it can't, the Product Plan documents why — and "set aside" is a successful output, not a failed track. The bar self-triages. That's how five tracks fit into four days without dilution.
This also resolves the Section 10 KPI gap in the Foundation — at the Sprint level. Outcome-based, not activity-based. Sprint Zero becomes the dry-run for the firm-wide measurement framework.
Seven tracks in. Five out.
Oli's original five, plus two additions I'd put on the table — Scoping Agent and Client Repo. We triage to five together at the Day 0 kickoff call one week before London, so the room arrives with tracks locked, squads pre-divided, and pre-reads already done. Grouped here by what they do for BCG, not by who proposed them.
Product evolution
Things case teams already use, made betterWorkflow utility
Things that make DS delivery cheaper and fasterIntelligence & pipeline
Things that connect DS to revenue and pipelineTriage and squad assignment happen at Day 0 — the virtual kickoff call one week before London. Me, Oli, Oscar on the call. Seven tracks reviewed against four criteria: POC achievability in four days, commercial leverage, cross-track learning, team appetite. Two get parked. They come back in Sprint 1. The five that survive get their pre-reads, their squad pairings, and their go-in brief before anyone books a flight.
Day 0 sets it up. Four days earn it.
Four intensive days in London — Tuesday through Friday, week of 23 June. Plus a Day 0 kickoff call one week earlier. By the time the room opens in London, tracks are locked, squads are paired, pre-reads are read, and Day 1 opens with Oli's intro and a teaser, not with orientation.
The full SLT. The CDs. Ten builders.
A deliberately stacked room. Sprint Zero builds in front of the firm's senior creative leadership, not in private. The week's outputs are reviewed live, decisions made in the same room, commitments earned in real time.
The leadership panel
3 named SLTThe Creative Directors
All CDs · region-spanningThe regional voice. Every CD in the room means every output is pressure-tested against the realities of every market the moment it's shown. This is how Sprint Zero outputs ship as firm-wide on Day 4, not as one-region experiments.
The Creative Technologists
10 · hand-picked across regionsSelected — not nominated. The criteria: AI-native builder, cross-disciplinary thinker, idea-to-prototype-to-pitch in a day. Hand-picked across regions so the Lab is born firm-wide, not single-office.
Already nominated by Regional CDs · mapped to Lab / Practice
31 named across regions · 10 to London · Lab membership confirmed Q3- Krishna Abhimanyu BEK → In London
- Yoojin Choi SEO → In London
- Yuna Moon ANZ → In London
- Hijing GC → In London
- Anshul Patria BEK → In London
- Edu Hernandez NXC · AMR → In London
- Bill Moore PIT · AMR → In London
- JP Lopez NXC · AMR → In London
- Ave Versteeg NXA · AMR → In London
- Ryan Davis PIT · AMR → In London Strong UX/UI AI craft — Cursor, Replit, Figma. A unique angle for the room.
- Jose Eguidazu MKA · EMESA → In London
- Paula Sanchez MKA · EMESA → In London
- Diwakar BEK · APACME Lead · primary
- Celestia GC · APACME Lead · primary
- Lee Gavin GC · APACME Lead · backup
- Nakul IND · APACME Lead
- Gema Arrieta NXC · AMR Lead · primary
- Mariana Dominguez AMR Lead · backup
- Nacho Rodriguez AMR Lead · backup
- Juan Biancardi ATL · AMR Atlanta rep
- Adam Bartczuk ECC · EMESA Lead · primary
- Jessica Manzano MKA · EMESA Lead · backup
- Oscar Armelles EMESA Visibility
- Shivali Chandra APACME → In London
- Kaivalya Joshi APACME → In London
- Suranjan Maiti APACME → In London
- Raymond Fung APACME → In London
- Shivam Soni APACME → In London
- Shiju B. APACME → In London
- Stephen Spencer AMR → In London
- Marcela Amador AMR → In London
The architecture: Foundation §3 defines Layer 1 — the Creative AI Lab as a global expert cohort, ~20 people, selected for visible craft and AI-native fluency. §4 defines Layer 2 — the Regional AI Creative Practice as open community where the field grows. Chips: → In London marks the 10 attending Sprint Zero · Lead marks Regional Practice Lead nominations · Visibility marks senior presence without layer designation. Final Lab membership confirmed in the weeks following Sprint Zero.
Plus me (running the week), Oscar Armelles (EMESA CD, regional anchor where present), and the capture crew working light-touch. ~20 people in the room across four days. Selected, not invited.
What Sprint Zero produces.
Five concrete outputs, all sized to feed the July Governance Session and the Lab's BAU rhythm from August onward.
Everything Sprint Zero produces needs somewhere to live.
Sprint Zero ships five POCs, five Product Plans, a teaser video, a film, a photo bank. Sprint 1 ships another five. Sprint 2, another five. Inside twelve months the Lab has thirty-plus pieces of validated AI work, plus all the GPTs, Claude projects, prompts and MD files that underpin them. That can't live in a folder.
A destination, not a folder.
A platform-agnostic, CMS-driven, database-backed web product — internal-first, externally-facing-to-BCG second. Fully responsive. Up-to-date by design. The single home for everything DS makes, finds, learns and ships in the Creative AI space.
What it hosts. Seven content types, one taxonomy, one search.
The trajectory. Internal first. BCG-wide second. External-facing (BCG audience) when the work earns it.
Not a SharePoint folder. Not a Notion page. Not a wiki. Not a Box drive. A real web product with a CMS, a database, a content model, an auth layer, search, taxonomy, and a design system that matches what DS already ships externally. Owned and operated by DS. Built for the firm.
Sprint Zero is not the build week for the Marketplace — that's a separate multi-quarter workstream. But Sprint Zero is the moment to lock the principle: every output we ship from this week lands in the Marketplace. The Marketplace gives Sprint Zero somewhere meaningful to point. Without it, the work disappears into the same folder graveyard everything else lives in.
Four flags worth naming.
Not problems — but the things that will go wrong if we don't address them upfront. Each has a proposed mitigation.
-
01 · Five squads for ten people might still be too many.If a squad feels stretched, POC quality suffers and the bar slips. The Day 0 triage call is the first defence — but mid-Day-3 we may need to consolidate if a track is clearly going to "set aside."Mitigation: the Day 3 mid-day demo session has explicit permission-to-park built in. Killing a track on Day 3 is a successful Product Plan, not a failure. Resources flow to the squads that need them most.
-
02 · Building under the gaze of senior leadership changes the room.SLT in the room means every demo is also a moment of evaluation. The risk: squads play it safe, build the demo rather than the real thing, or stop being honest about what's broken. Building for approval is different from building toward truth.Mitigation: Day 2 is closed — no demos, no presentations, no leadership drop-ins. The day belongs to the squads. SLT visible Tuesday (open) and Friday (evaluation); deliberately absent from the build days. Combined with explicit Day 1 framing that "set aside" is a successful output.
-
03 · The room can't honestly kill its own work.If the cohort feels they're being evaluated on building something that scales, no track will produce a "set aside" recommendation — even when that's the right call. The Foundation supports kill decisions; the room has to feel that's true.Mitigation: name it explicitly on Day 1. Celebrate a credible "set aside" in the Day 3 evaluation. The first track to honestly call its own death sets the culture.
-
04 · Capture feels intrusive and dampens the room.Film and photo crew working a small in-person event can change the energy if they're heavy-handed. Especially in moments where people are struggling — and those are the moments worth capturing.Mitigation: light-touch crew — one shooter + one producer, brief them on Foundation language and the "no pizza and energy drinks" ethic. Lav mics over boom. Capture build moments, not staged ones.
What I need from you.
01 · Naming — Sprint Zero · aka The London AI Week. Sign-off on the two-name approach. This becomes the calendar invite, the doc header, the slack channel.
02 · Dates — Tuesday–Friday, 23–26 June. Day 0 kickoff call Wednesday 17 June. Confirm the calendar holds — and that the July Governance Session timing is fixed.
03 · The room — SLT confirmed (you, Richa, Justin), all CDs in attendance, 10 Creative Technologists hand-picked across regions. Sign-off on the selection list before invites go out.
04 · Track triage — the Day 0 call with you and Oscar. Seven tracks in, five out. Scoping Agent and Client Repo earn their place.
05 · Teaser video — commission the Creative AI Lab teaser this week. You reveal it during your Day 1 opening. ~60–90s. I'll brief production.
06 · Marketplace direction — in-principle approval to scope DS Marketplace as a Sprint 1+ workstream. Sprint Zero outputs land here. Owned and operated by DS — built for the firm.
07 · Budget envelope — separate doc. I'd like averages from you so the costed plan is credible without being to-the-pound.
This is the moment DS takes the BCG Creative AI space. Let's lock the week in.