Our Response, Packaged as a Working Dossier.
The Magnum response room — four working tools grounded in the real RFP, plus Ask Mara over the approved material below. TMICC produces 1,400–1,800 adapted assets a year across an 80+ market footprint; our answer is MRCL MAKE — one connected Human x AI craft system with the Magnum heart at the centre. MRCL MAKE handles production speed, Occasionist shapes cultural relevance, Product Builder maps local flavour culture into dual-flavour routes, and the Food Checker governs market compliance, all one pipeline. Open any tool from the working-tools rail, or the response pack under Downloads.
This is your Magnum dossier.
MRCL MAKE, Occasionist, Product Builder and Food Checker are one brief, one corpus — the RFP timeline, the win themes, the core team. Ask Mara alongside brings approved facts, citations and tool links into one place.
Ask Mara about the response.
Mara answers from the client-approved documents in this dossier only — ask about the timeline, the team, the win themes, MRCL MAKE, Occasionist, Product Builder or Food Checker.
Voice coming soon
Your core team
Growth Engine · NBOs · MRCL MAKE
Adaptation & Localisation Partner Pitch
Invited to respond to The Magnum Ice Cream Company's Adaptation & Localisation Partner RFP — a nine-brand portfolio (Magnum, Cornetto, Twister, Ben & Jerry's, Carte d'Or, Breyers, Klondike, Talenti, Yasso) producing 1,400–1,800 adapted assets a year across an 80+ market footprint. The incumbent is a Publicis-competitor holdco. Our answer is MRCL MAKE — a connected Human x AI craft system with the Magnum heart at its centre, built for production speed, cultural relevance and market compliance in one pipeline.
growth-engine → NBOs → Magnum → MRCL MAKE
Grounded in the real TMICC & Magnum brand world
Sources: TMICC corporate identity + RFP pack · Magnum 2024 Brand Book Lite — no stock
Crack — DBA
Bite & Pour — DBA
Where Art Meets Innovation
Brand Evolution
M-Stamp — DBA
Needs Content Grid
Twister — TMICC
Cornetto — TMICC
Portfolio — TMICC
Life Tastes BetterRFP timeline (real — from growth-engine NBO record)
Deepti Velury & Jay Humphreys presented the HNK intro deck; most time on the Heineken case study.
Adaptation & Localisation Partner RFP issued; WIP Q&A doc started.
Kick-off call — R&Rs, game plan to submission.
09:00 CET — internal vs external split, Media team asked first.
30-min Q&A with TMICC; pricing matrix template + test briefs shared.
Updated deadline (was 3 Jul; client confirming).
Win themes & core team
est. end-to-end turnaround
scored across 9 brands × 23 products × 80 markets
9 real rule regimes, EMEA / Americas / APAC
Move /creativetech into growth-engine's real RFP location
MRCL MAKE — the connected Human x AI craft system for the response, with the Magnum heart embedded in the identity. It covers master adaptation and variation adaptation, fed by the Opportunity & Compliance Radar: Occasionist finds the moments, Product Builder creates culturally credible flavour routes, Food Checker clears the markets, and all three route briefs and fixes straight back into production.
MRCL MAKE
The Human x AI production workspace — brief, deconstruct, variant, sound, orchestrate, review — the real build from engines/wev/maker (Magnum's actual adaptation demo, La Pistache → La Pêche).
Open workspace ▸Occasionist
The real Occasionist recommender ported for Magnum — same composite formula, score bands and cross-market pattern detection as the source engine. Grounded in the Magnum Brand Book's own Art/Fashion/Festival passion points, now spanning 65 moments from Cannes down to local city carnivals across the full 9-brand TMICC portfolio (23 products, 80 region-grouped markets), each moment hand-annotated and most tagged to a home market so scoring reflects real local credibility. Brief, Playbook, Horizon, Products, About.
Open tool ▸Product Builder
A researched flavour-intelligence layer for Magnum: global, regional and country-level dual-flavour routes ranked by cultural fit, Magnum right-to-play, novelty, occasion fit and compliance risk, then routed into MRCL MAKE as a product brief.
Build a flavour route ▸Food Checker
Every adapted asset checked against real food-advertising and AI-disclosure rules, market by market, before it ships — HFSS, EU claims & AI Act, kids' rules, halal, green claims — with fixes routed straight back into MRCL MAKE.
Run a check ▸Internal console · RFP Checker
How much of the RFP have we answered?
Score the latest written-submission deck against the real TMICC Adaptation & Localisation Partner RFP — §08 “How to Respond” plus the §04 evaluation KPIs. Load the copied deck assessment for an instant read, then use the live Mara interrogation to re-check the corpus.
Deck intake
The latest deck has been copied into the working corpus. Upload remains available when the live Mara corpus needs refreshing.
After upload — file the canonical copy
- 1 Deck ingested into Mara’s corpus — chunk count confirmed in the log beside this card.
- 2 File a copy of the same deck to the git-approved SharePoint folder:
Manual step — SharePoint has no API bridge from this static site.
…/PP/NewBusiness/Magnum/Corpus/OurDocs/ - 3 Confirm the corpus copy and the SharePoint copy match before running the coverage check.
Latest deck assessment
Use the local extraction of NEW_Magnum RFP_Written Submission.pptx as the current scorecard.
88% coverage · strong with commercial gaps
146 extracted slides were checked against the same 16 RFP requirements used below. The result is strong on operating model, AI production, hubs, governance, process and reporting; it still needs a contract-ready rate-card appendix, references appendix, and explicit pass-through-cost policy.
- Commercial appendix is partial: rate-card logic is present, but the completed TMICC table is not visible.
- Pass-through costs are not explicitly stated.
- Reference proof is mostly case-study narrative, not named contactable references.
Coverage interrogation
Sixteen real interrogation calls to Mara — one per RFP requirement — against the latest deck in the corpus. Run them all, or refresh any single requirement.
Verdicts (ANSWERED / PARTIAL / GAP) come from Mara’s own STATUS field — each interrogation asks her to answer in a structured STATUS / EVIDENCE / GAPS / NEXT format, which is parsed directly. When she doesn’t follow the format, an honest heuristic fallback (recommendation + confidence + gap-language) takes over and is labelled as such on the card. Triage aid, not a scoring authority. Coverage % = (answered + ½·partial) ÷ assessed requirements.
Super admin · access
Invite users.
Admin surface for the Magnum response portal. The live token authority remains PP Engines Access; this panel keeps the admin workflow visible inside the one portal rather than hidden in a separate app.
Send a response-portal access request
Enter a stakeholder email and role. Until the invite endpoint is connected, this prepares the exact PP Access request so a super-admin can issue it from the approved access service.
Current admin session
Reading your token entitlement...
Super admin · Mara
Mara inbox.
A single inbox for questions asked in this browser session. Client questions are logged with client-safe mode; internal questions are logged with frank internal mode. A persistent backend audit stream still needs wiring before this becomes the full operational inbox.
Questions and answers
ElevenLabs handoff needed
Mara voice now has a clear product rule: client voice must stay response-focused and never surface weaknesses; internal voice may discuss gaps, risks and missing material frankly.
Mara Internal · RFP interrogator
Interrogate the corpus
Corpus 9 docs
Showing the known seed-pack source list.
Add a document
RFP timeline
Same record as the RFP snapshot above — growth-engine NBO timeline for Magnum.