Main working contact at direct-booker (theresa@directbooker.ai). UK-based (Oxford area / Marlborough). Exact surname/title unknown from the transcripts. (“Teresa” / “Christa” in some transcripts = same person.)
Background as relayed by Elliot: Senior Product Director at TripAdvisor, Head of Product + Head of Engineering at Time Out Group, then co-founder of Direct Booker.
Owns Direct Booker’s product/connector side: Claude/ChatGPT tool releases + descriptions, the directory-page copy, the Claude dashboard, the UI-element rollout, and the prompt/UX research. Technical and hands-on (pushes releases, probes Claude on its own behavior).
Drives DB’s discovery strategy → top-of-funnel dateless hotel search (see direct-booker): her thesis is to be surfaced as early in the journey as possible, before dates. Ran a UserBerry survey (25 real-user responses) to capture first-prompt phrasing and found almost no one includes dates — the evidence behind the dateless-tool hypothesis.
Cares about closing the loop: making DB the first proof case for agent-discovery optimization, attributing visibility gains to specific actions, and ROI. Maintains DB’s session/click metrics + Looker dashboard.
Sharp on prompt realism: flagged Ghostteam’s early benchmark prompts as artificial; wants natural user prompts from survey/live data; will refine + prioritize the working prompt set.
Executes immediately: pushed the recommended one-liner keyword change live during the 2026-07-06 call (form-based connector fields need no code release). Runs her own competitive intel: pulls competitors’ tool metadata visible in ChatGPT into a Claude project for analysis (ChatGPT exposes release version numbers; Claude hides them). Owns DB’s chain-contract risk judgment on brand-term keywords (Google-Ads-era anti-bidding clauses; knows which chains have live apps and which are litigious).
Product instinct: Direct Booker’s “sniping” use case is to surface an official/direct hotel price alongside OTA results, without needing to own the whole travel-planning conversation.
Open questions
Exact surname and title.
Timeline
[2026-07-06] (call) Recommendations session with vincent-mcleese + elliot-garreffa + sanj-directbooker: accepted the four-change sequencing (~1 week each), pushed change #1 live in-call, confirmed connector field caps (one-liner 200 chars / long description 2,000), committed to a fresh-install test of the 2-vs-3 tool discrepancy, and scoped the chain brand-clause risk (Hyatt/Hilton/IHG/Wyndham have live apps; test less-sensitive brands, possible revert). Next check-in Thu 2026-07-09, 13:00 her time. Detail → direct-booker.
[2026-06-26] (call) Ran the 3pm client findings read-out with vincent-mcleese + elliot-garreffa. Articulated DB’s top-of-funnel dateless goal, shared her UserBerry first-prompt research, surfaced the tool-health anomaly + the Claude UI held-back rationale, and agreed to refine/prioritize the prompt set ahead of a ~Tue 30 Jun recommendations session. Full detail → direct-booker.
[2026-06-17] (call) Joined the Direct Booker optimization call with elliot-garreffa, vincent-mcleese, and sanj-directbooker. Shared Claude dashboard context, the review-lock issue after changing tool definitions, planned before/after tool-description sharing, and feedback that Ghostteam should collect more realistic travel prompts. Asked for the prompt-summary deck and detailed written answers to open questions.
[2026-06-19] (internal discussion) Elliot noted Theresa’s senior product/engineering background: TripAdvisor senior product director, Time Out Group head of product + engineering, then Direct Booker co-founder. Reinforces that she is a high-quality product counterpart for the controlled optimization experiment.