Interactive walkthrough · synthetic data · nothing computes or sends

The Agentic GTM Machine

A real workflow automation recreated for a fictional client, Calyx Diagnostics. The production inputs are deliberately plain: a folder of call transcripts and Google Calendar. Agents run the pipeline; humans hold the judgment gates. Press run and watch a morning happen.

AGENT, runs automatically HUMAN GATE, judgment required
Every name, company, and number on this page is synthetic. The workflow is real: in production it reads a transcript folder and Google Calendar, and posts to Slack.
Monday · idle. The agent runs at 07:00; the coordinator reviews when they arrive at 08:45.
07:00
Calendar scan
07:02
Meeting prep
07:06
Pipeline sweep
07:09
Channel posts
07:12
Follow-up drafts
08:45
Coordinator arrives, reviews

Morning brief

source: google calendar + transcript folder + crm
agent · 07:00

Calendar read: Monday

google calendar · read-only scope
  • 9:30 Demo: Fernbrook Wellness · Maya Okafor · Meet
  • 11:00 Onboarding kickoff: Averline Health · Dana Reyes · Meet
  • 2:00 Internal: weekly pipeline review · conference room

Three events, two external. The agent sees the day; it cannot move, create, or decline anything. Prep cards for both external meetings follow.

agent · 07:02

9:30, Demo: Fernbrook Wellness

Maya Okafor, Head of Product · 2nd meeting
  • Last touch: asked about CLIA lab turnaround at intro call (May 2)
  • Signal: posted last week about launching a metabolic panel
  • Open item: pricing tier for <5K kits/yr was never answered
  • Prep: deck updated with turnaround SLAs; pricing one-pager attached
agent · 07:04

Call coaching: the 9:30

sources: past transcripts + calendar + public posts · goal set from the deal playbook
Call goal: a verbal commit to a 500-kit pilot, and a named technical contact for results-API scoping.
  • Open with the debt: her pricing question from the May 2 transcript was never answered. Lead with the Starter tier before she asks twice.
  • Tie turnaround to her launch: she posted about a metabolic panel launch last week. The 5-day CLIA SLA is a launch-timeline fact for her, not a spec of ours.
  • Plant the integration: offer sandbox API keys during the pilot so results land in her product from day one. That is what makes the pilot sticky.
  • Likely objection: turnaround at peak volume. Show the SLA-by-volume table; worst band is 7 days, and say so before she finds it.
agent · 07:03

11:00, Onboarding kickoff: Averline Health

Signed Thursday · first kit order pending
  • Contract: countersigned, 2,500-kit initial order
  • Blockers: their compliance lead hasn't returned the BAA
  • Prep: onboarding checklist generated; BAA nudge queued for approval →
agent · 07:04

2:00, Internal: weekly pipeline review

Agenda drafted from CRM deltas
  • This week: 2 deals advanced, 1 stalled (Halewood, 21 days silent)
  • Decision needed: requalify or archive Halewood

Pipeline sweep

source: crm spreadsheet + enrichment + inbound forms
crm.sheet · the agent reads it and writes backrow 4 written 07:06
CompanyStageScoreDaysNext
2Fernbrook Wellnessdemo_scheduled784demo 9:30
3Averline Healthonboarding922kickoff 11:00 · BAA
4Torrey Peak Supplementsnew_inbound840sequence step 1
5Halewood Nutritionstalled41212pm decision
green: row written this cycle · red: stall rule fired
agent · 07:06

New inbound: Torrey Peak Supplements

Scored 84/100 · routed to demo track

DTC brand, ~$12M est. revenue, launching gut-health line. Form mentioned "white-label testing." Enriched, deduped, sequenced.

agent · 07:07

Stall flag: Halewood Nutrition

21 days no reply · owner: Sam

Rule fired: 3 unanswered touches. Re-engagement draft prepared, routed to human gate, not auto-sent.

Agent channel

slack · the agent posts to its own channel; humans reply in thread
#gtm-machine3 new · 07:09-07:11
gtm-machine APP7:09 AM
Morning summary
Meetings: Fernbrook demo 9:30 (prep card ready) · Averline kickoff 11:00 (BAA pending)
Pipeline: +1 inbound scored 84 (Torrey Peak) · Halewood stalled 21d, decision at 2pm review
Awaiting approval: 2 drafts in my queue
gtm-machine APP7:10 AM
Call analysis: Averline pre-kickoff sync (Fri 4:00, 18 min)
Trigger: new file in /transcripts: averline-presync_fri-1600.vtt
Read: tone positive, kit volume reconfirmed at 2,500. Their compliance lead was absent; the unsigned BAA is the only open risk to first-order timing.
Goal ladder (integration live before first order): contract ✓ · BAA ○ · sandbox API keys ○ · test round-trip ○
Rank: 62/100 toward goal; the call moved nothing up the ladder but cleared the path (volume settled, agenda accepted). BAA is the bottleneck rung.
Next steps: BAA nudge queued for approval; book their CTO for a 20-minute API-scoping call this week, sandbox keys issued the same day it happens.
gtm-machine APP7:11 AM
Call analysis: Halewood, last call 21 days ago
Trigger: stall rule re-read /transcripts: halewood-disco_jul25.vtt, 21 days old
Read: pricing objection never resolved ("waiting on budget"), champion went quiet after. Their CEO posted about retail expansion yesterday, a fresh reason to talk.
Goal: requalify or archive. Ranked 15/100 toward reactivation; no rung has moved in three touches.
Next steps: recommendation is one signal-based touch citing the expansion post, then archive if silent 14 more days. Judgment call, held for the 2pm review, not auto-sent.
↳ thread · decision lands at the 2pm review

Approval queue

nothing external sends without a human
human gate · draft ready 07:12

Follow-up: Fernbrook pricing answer

to: maya.okafor@fernbrook.example
✓ sent 8:47 AM, logged to CRM, thread linked
human gate · draft ready 07:13

Nudge: Averline BAA

to: compliance@averline.example
✓ sent 8:48 AM, onboarding checklist updated
human gate · flagged 07:13

Decision: Halewood re-engagement

recommendation attached, not auto-sent

Agent recommends one final signal-based touch (their CEO posted about retail expansion yesterday), then archive if silent 14 more days. Judgment call, tone risk on a cold thread. Held for the 2pm review.