Web3Audience (Laurent Chemla)

Laurent operates through several vehicles — web3audience.io (his email domain), a personal holding company (name garbled in transcript), and Six Degrees (AI). This page tracks the Ghost Team partnership thread with him; he is a channel/deal partner, not a buying customer.

Current truth

  • Partnership structure proposed by Laurent (2026-07-01): SPV per client opportunity, not a rev-share or distribution agreement. His holding contributes client relationships, industry knowledge, and IP; Ghost Team contributes MCP/discovery technology and execution; royalty split agreed upfront per SPV. His stated benefit: the end client can later acquire that SPV/IP block (Laurent says enterprise clients increasingly want to own or license IP, and that legal/ICT compliance blocks perpetual-licensing relationships) without Ghost Team exiting the whole company. Elliot probed the benefit vs a plain per-client rev-share; unresolved.
  • [idea] Laurent’s read: Stellantis used-car division → Peugeot is the shortest path to a deal — the global used-car head is AI-savvy, Paris-based, has budget already unlocked for GEO/perception, and was shown agentdiscoverability by Laurent in late June; Peugeot is in a brand-perception crisis and GEO is already a live internal subject across several Stellantis brands.
  • Laurent’s stated position on Stellantis broadly: he won’t engage further without CEO/chairman-level commitment; he is pushing them to spin off an independent AI-native entity with a €100–200M injection, because pilots without executive mandate stall after ~3 months (his account).
  • [idea] Six Degrees (AI) — another Laurent company — is about to launch a large Peugeot/Fiat pilot across Europe including AI share-of-voice / brand-perception reporting (his account). Floated option: Ghost Team white-labels its live-prompt architecture (API or white-label) as the AEO/GEO measurement layer under that work. Our side flagged it needs internal discussion: capacity, focus fit, commercial terms.
  • We owe Laurent (from 2026-07-01): a two-column pricing summary (MCP app build/maintenance; discovery platform retainer) and a customized version of Elliot’s event presentation (add BMW MCP example + used-car funnel narrative, tailored to the used-car contact).
  • Laurent owes us: a date for a call with the right Stellantis-side stakeholders.
  • Beyond Stellantis, Laurent floated HubConnect and other groups as further client sources for SPVs.
  • Separate thread, not Ghost Team business: Vincent and Laurent also run a separate joint venture under web3audience (Vincent has a web3audience.io address). Deliberately not tracked in this shared brain.

Open questions

  • Laurent’s surname and holding-company name are inferred/garbled — “Chemla” from lchemla@web3audience.io; holding name unintelligible in transcript. TODO: verify (no web lookup during ingest).
  • Is the SPV-per-client structure actually advantageous for Ghost Team vs a simple rev-share agreement?
  • Do we want the Six Degrees white-label AEO work at all (capacity + focus + commercial terms)?

Timeline

  • [2026-07-01] (call — Vincent + Elliot + Laurent; recorded by both founders, merged) Reconnect after ~4 months. Laurent debriefed his keynote to the Stellantis global marketing event in Rome (14 brands; ~70–140 people; “AI is an asteroid, not a wave”; recommended spinning off an AI-native entity with €100–200M): says he mentioned Ghost Team in the context of the agent-driven purchase funnel and showed agentdiscoverability to the global used-car head (AI-savvy, building an internal used-car AI brain with RAG, blocked by compliance). Anecdotes from his account: legal blocked a fully AI-produced Jeep commercial at Cannes over AI-character likeness/IP; Stellantis fears Chinese-orchestrated negative sentiment against Fiat in Italy. We walked our positioning and pricing: full-service offering targeting marketing/product-marketing teams; two segments (no AI-distribution presence yet vs has an MCP app but can’t optimize); retainer $5–10k/mo, managed service on top → ~$10–15k/mo total, one-off template-based MCP app build $10–30k; low-cost self-serve data SaaS tier planned; platform pro tier $149/mo for 50 prompts, charged on prompt volume; roadmap up-funnel to AEO/citations and down-funnel to ad bidding by intent. Named monday.com as signed to optimize their Claude connector and cited Kiwi at 500k users/month via their Claude connector. Competitive picture as we presented it: one direct discovery competitor (copied us), 4–5 VC-backed build/infra players ($3–8M raises). Elliot explained ADO mechanics (Claude-only organic today; ranking = training-data prominence + app health + registry keyword matching; optimization cycle in weeks — closer to ASO than SEO). Laurent proposed the SPV partnership structure (see Current truth) and flagged Peugeot/used-car as the priority use case. Mentioned to Laurent: rebrand upcoming, angel round upcoming. Next steps: we send the pricing table + customize the presentation; Laurent gathers stakeholders and finds a date; explore the Six Degrees white-label AEO option.