Ghost Team Brain

Ghost Team Brain

9 min read

Index — generated by tools/rebuild-index.sh, do not edit

  • Fundraising Playbook — Fundraising playbook for Ghostteam’s current stage — angel-round mechanics, valuation testing, pitch narrative, timing,
  • Investment Fundraising — Investment/fundraising relationship tracker — warm investor/advisor contacts, raise timing posture, and proof points to
  • Partnerships — Active partnership relationships — hosting platforms, integration partners, referral loops, and where partner/customer r
  • BD Pipeline — One line per active thread: company | stage | next action | owner | last touch. Updated by ingest-ca
  • Podcast Appearances — Talks and podcast appearances — public narrative moments where Ghostteam explains ADO, agent discoverability, and the ag
  • Design System — Ghost Team design tokens and component rules extracted from the live marketing frontend for reuse across slides, website
  • Naming and positioning — Naming and positioning for the unified Ghostteam / Agent Discoverability brand — name criteria, headline candidates, aud
  • Alpic — Partnership stage: cautious integration design. The substantive partnership call happened 2026-0
  • Awaze — Awaze — PE-backed European holiday-rentals group (Novasol, Cottages.com); warm discovery prospect via Sebastian Florian.
  • Carwise — Carwise / getcarwise.app — automotive marketplace ChatGPT app prospect; solo builder seeking organic discovery and ranki
  • Direct Booker — Direct Booker — hotel-booking aggregator startup optimized purely for AI-agent discoverability; currently most-visible i
  • Lucidchart) — Active engagement: ChatGPT-app eval testing under a statement of work (SOW contact was Mallory, prod
  • ManpowerGroup — ManpowerGroup — paid engagement using Ghostteam’s LinkedIn/social signal tracker; CMO Ruth is the champion; weekly diges
  • Manufact — Manufact — manufact.com MCP hosting/devtools platform; both a potential Ghost Team customer and preferred partnership ch
  • Mitchells & Butlers — Consultancy phase wrapped 2026-06-12 (wrap-up call: ewa-turney, [[people/richard-whitehou
  • Monday.com — Monday.com — project/task-management SaaS; SIGNED client (confirmed 2026-07-06 via email) for Claude-connector optimizat
  • Qonto — Qonto — EU business-banking fintech that launched a 52-tool MCP (Jun 2026); warm prospect evaluating ADO platform, waiti
  • Tink — Tink / gettink.ai — recruitment/job-discovery product exploring whether ChatGPT/Claude connector discovery is worth prio
  • Web3Audience (Laurent Chemla) — Laurent Chemla (web3audience.io) — partner/channel thread, not a classic customer; SPV-per-client partnership proposal,
  • Decision: offering, pricing & fundraise posture (June 2026 retro) — Offering shape, pricing posture (no free tier), and fundraise stance set in the 2026-06-15 retro — one narrative, paid-f
  • Decision: Q2 goals & GTM plan (June 2026 sprint planning) — Q2 goals + GTM plan set 2026-06-16 — 5 (stretch 10) paying customers, goals→objectives→KPIs structure, brand/name still
  • live — SUPERSEDED 2026-06-24 → public stays benchmark-only; custom is a private self-contained arena. Was “public = better of (
  • Decision: self-contained custom arena + public stays benchmark-only — Custom score = a self-contained private arena (the whole roster re-scored on the brand’s own prompts); public stays benc
  • Decision: app-build pricing model + freelance delivery strategy — Build estimate (two apps): ~21 dev-days. Architecture/tooling design 5d (deliberate thinking tim
  • Decision: company brain structure (agent-readable) — How the company brain is organized for agent consumption — folders, page format, granularity, boundaries, pipelines.
  • Decision: stop assigning work to Quan; escalate to Han — Dev strategy — Quan contract winding toward an end-it call + the replacement/hiring track (co-located Amsterdam/London d
  • Decision: discovery niche, not ChatGPT-app agency — The window for building ChatGPT apps as a standalone agency offer has passed (“that ship has saile
  • Decision: epistemic-status standard (claim-level maturity tags) — The brain-wide epistemic-status standard — five claim-level maturity tags (idea → hypothesis → evidenced → fact, plus re
  • Andre Broekman — Andre Broekman — solo builder of Carwise/getcarwise.app, an automotive AI-app marketplace prospect.
  • Apeksha Awasthi — PM for Lucid’s MCP team — our main point of contact for MCP server changes on the ChatGPT-app engagement.
  • Barak (Monday.com) — Barak (barakbe@monday.com) — Monday.com; analytical voice on the ADO prospect (A/B-test + competitor-attribution questio
  • Blake Fulford-Brown — Blake Fulford-Brown — Senior Product Manager at Awaze; owns the analytics, the data/ROI-skeptic voice, flags LLM-surface
  • Connor Murphy — Connor Murphy — Oisin’s closest investor and recommended first angel-syndicate intro for Ghostteam’s possible rolling SA
  • Dimitri Ewald — Chief of staff at Alpic — handles outreach/radar; on the partnership thread.
  • Elena Librich (Monday.com) — Elena Librich — Monday.com growth team (product-led acquisition); Ghostteam’s relationship lead for the Monday ADO prosp
  • Elliot Garreffa — Cofounder, ghostteam / agentdiscoverability.com (elliot@ghostteam.ai). Owns BD/outreach, sprint plan
  • Erica Beavers — Our main contact / champion at Alpic — organises the partnership calls; the relationship owner on their side.
  • Ewa Turney — mitchells-and-butlers (ewa.turney@mbplc.com). Our day-to-day contact on the AI apps /
  • Han — Han — manager of quan at the Vietnam dev agency; our escalation contact on contractor performance. Fast to re
  • Harley (Lucid) — Harley — Lucid manager of Mallory (product marketing); interim commercial contact for the Claude discovery SOW while Mal
  • Jens Mannanal — Who is Jens Mannanal (Bluerails founder, ex-GetYourGuide) and what does he care about?
  • John Perry — John Perry — Senior Engineering Manager at Awaze (Sebastian Florian’s manager); owns Awaze’s 12–18-month engineering-pla
  • Kelsey Nield — Lucid PM driving the unauthenticated-user-mode launch (the second ChatGPT submission).
  • Laurent Chemla — Laurent Chemla (lchemla@web3audience.io) — Vincent’s long-time deal partner; enterprise-marketing rainmaker with Stellan
  • Mallory (Lucid) — Mallory — Lucid product marketing; the original SOW contact for Ghostteam’s Lucid engagement; on maternity leave (~until
  • Muzamil Hasan — Muzamil Hasan — host of Beyond the Bubble; interviewed Elliot on agent discoverability and ADO.
  • Nick Ridge — Growth PM at lucid (nridge@lucidchart.com). Our primary working contact on the ChatGPT
  • Oisin — Oisin — founder friend/adviser who recently closed a ~$4M angel-led round and is helping Ghostteam think through a rolli
  • Pietro Zullo — Pietro Zullo — founder of manufact (pietro@manufact.com), their-side contact; elliot-garreffa o
  • Quan — Contract developer in Vietnam, paid hourly, works on the ADO platform/tracker ([[product/ado-platfor
  • Richard Whitehouse — Senior stakeholder / decision-maker at mitchells-and-butlers (richard.whitehouse@mbplc
  • Rodrigo Mallo — Rodrigo Mallo — Outsized.vc investor contact; warm fundraising relationship for Ghostteam.
  • Roslyn — Roslyn — Noetic; Davos-network bridge to ManpowerGroup (Ruth) and a warm investor/adviser-intro source for Ghostteam.
  • Ruth (ManpowerGroup) — Ruth — CMO of ManpowerGroup; champion for Ghostteam’s signal-tracker engagement.
  • Ryan Ferguson — Developer on Lucid’s MCP team — makes the actual MCP server changes we want tested.
  • Sanj (Direct Booker) — Sanj / Sanjay — Direct Booker (likely founder/senior); relayed the Marriott investor signal; Oxford-based team.
  • Sebastian Florian — Sebastian Florian — principal engineer at Awaze; Ghostteam’s ~4-month relationship owner there (met at a Paris MCP event
  • Sophie Cornay — Sophie Cornay — Qonto contact (product/marketing intersection) leading their internal evaluation of ADO; gatekeeper to t
  • Dev candidate — Spibel founder (Amsterdam) — Amsterdam dev candidate (name not yet captured) — founder of the Spibel note app + Keuspodium; via Dennis; first candida
  • Stefano Amorelli — Stefano Amorelli — external engineer at Qonto who built their MCP server; technical counterpart on the ADO evaluation.
  • Theresa (Direct Booker) — Theresa — Direct Booker contact driving connector/product updates and discovery-optimization collaboration.
  • Tom Chevalier — Tom Chevalier — Tink/gettink.ai contact exploring whether AI-agent connector discovery is worth prioritizing for recruit
  • Valentin Beggi — Developer at Alpic who built/contributed to Atlas and Beacon — the technical voice on the partnership.
  • Vincent McLeese — Vincent McLeese — Ghostteam cofounder; owns much of the product, scoring, and platform implementation work.
  • ADO Platform (product hub) — The product hub — what the ADO platform is, current focus, and the index into product knowledge. Start here.
  • ADO Score Model — How the ADO score is computed — the four 25-pt buckets, what feeds them, and the categories/keywords prerequisite.
  • Architecture — How the platform is built — the durable architecture invariants + the canonical repo docs to read. Links, does not resta
  • Build offering — discovery-optimized (USPs + philosophy) — The discovery-optimized BUILD offer — the six USPs, the tools-first build philosophy (“start with the tools that win the
  • Discoverability framework (the 4 R’s + the loop) — The discoverability framework — the four R’s (Reachability, Relevance, Retrievability, Rank) as the measurement spine, t
  • Idea Triage — Idea triage funnel — raw product ideas to discuss, accept onto the roadmap, or reject (with the why). Pre-Linear.
  • Roadmap (strategic) — High-level strategic roadmap (Now/Next/Later) — themes, the why, and links to Linear epics. Not the backlog.
  • Agent discovery market & competitive landscape — Two named competitors, opposite shapes. Ora (web-readiness tooling, below) and Wani Wani
  • Claude connector routing (measured) — How Claude actually routes a connector recommendation, measured on ~4.3k of our own runs — prompt→keyword compression (3
  • Claude directory rank — decoded (the browsable listing over time) — How Claude’s browsable connector directory ranks connectors over time — decoded from 63 days of scrape history. Freshnes
  • Claude registry search — decoded (the backend “opaque box”) — How Claude’s backend search_mcp_registry actually matches and ranks connectors — decoded and live-validated. Which metad
  • Company memory & agent infrastructure — **The brain is LIVE (2026-06-12): ghostteam-brain repo, cloned to each founder’s home-directory to
  • Bluerails — Who is Bluerails and how does their agent-discoverability product (400-prompt website scans, €99/mo) compare to our ADO
  • MetaTuner (competitor) — MetaTuner — Rotterdam-based (likely solo/part-time) MCP-optimization tool; optimizes invocation quality POST-connection
  • Noodle Seed (competitor) — Noodle Seed — website↔AI-app connection layer with persona generation and a conversion/lead focus; “Halo” turns a websit
  • Wani Wani (competitor) — Wani Wani (waniwani.ai) — “Make AI your #1 distribution channel”; builds AI storefronts across ChatGPT/Claude/Gemini, in
  • Discovery analysis method (how we decode a selection) — How and what we analyze to EXPLAIN an agent’s selection — the three data layers (run artifacts / tool calls + responses,
  • Per-client organic surfacing (Critical · in motion) — How does each major client organically surface tools the user has NOT connected — the triggers, surfaces, and data sourc
  • Signal set & weights (Critical · in motion) — What is the complete set of signals an agent uses to select a tool, and how are they weighted? Inventory grouped under t
  • Tool-description optimisation (Critical · in motion) — MCP-specific — What is the best way to optimise tool descriptions for agent discoverability/selection — and does agent-optimal diverge
  • Product Thesis — Agent Discovery Optimisation (ADO) — Our apex bet — Agent Discovery Optimisation (ADO) becomes a defining discipline, and the team that builds its measuremen
  • Research Priorities (ADO) — The live ADO research agenda (set 2026-06-16) — Priority 1 (tool-call triggers, agent loyalty/memory, connector segmenta

Graph View