ESTIMATEDraft · setup TBD
Weekly pipeline report agent
Auto summary of outbound KPIs
Codex/Claude + Instantly & HubSpot MCP
$1–5 per run LLM (ESTIMATE)
ESTIMATESetup supplied · untested
Codex list QA and dedupe
Clean a lead CSV before sending
The stack
Codex CLI + verifier
Stack alternatives and non-tool inputs are preserved above; linked tools do not imply every integration has been verified.
Prompt / setup
Supplied setup, not tested by GTM Cookbook. Review permissions and keep outputs as drafts.
Prompt Codex to dedupe, normalize titles, flag role/free emails, output clean.csv.Cost breakdown
ESTIMATE≈$1–10 LLM only (ESTIMATE) + verifier cost (unsourced)
Preserved from plays.json. The title follows the per-1,000-leads template; use the actual unit above when this play counts accounts, replies, pages, runs, or months. Not a measured total.
Unit prices not supplied. This cost cannot be independently reconstructed from the dataset. Verify vendor pricing before budgeting.
Unpriced components: Codex. No applicable unit price is supplied for this play. Do not interpret missing prices as free.
Plan allocation assumes the stated volume. Minimum subscriptions, unused credits, taxes, human labor, mailboxes, domains, and verification can change the total. Supplied totals are not silently recalculated.
Expected results vs benchmark
PLACEHOLDER RESULTS · ESTIMATE / TBD
No measured outcomes are supplied for this play.
ESTIMATE/TBD; benchmark: AI email $420–780 per AE-accepted meeting (coldcalls.ca, 2026, vendor-published)
Benchmark source: coldcalls.ca (2026, vendor-published; exact report URL not supplied). $420–780 is per AE-accepted meeting, not per lead. It is context, not a forecast or proof this play works; it may not apply to non-email workflows.
Source notes
Dataset checked: . Recorded claims have not been freshly verified.
- Clay MCP GA in Claude, ChatGPT, Claude Code, Codex (clay.com/blog/mcp-for-reps); Apollo/HubSpot/Instantly/lemlist/HeyReach official MCP (knkoutbound.com/blog/gtm-cli-tools, Oct 2026)