2026-07-16 · Source: 120-day Jira (~260 tickets) + 11 marketing repos, read by Claude subagents;
roadmap read from pg-agent-gov repo. Items marked "judgment" are Nathan/Claude analysis, not verbatim source.
1Marketing team work patterns
In one sentence: nearly all recurring work is "one event → fan out a full set of tickets across people, channels, and languages".
18 patterns converge into four families; the loop only absorbs this recurring execution — strategy, creativity, and integration are explicitly excluded,
which is the organizational upgrade plan itself (Belle → integrator, Chelsea → CRM planner, Jon → high-value content).
Family
Representative patterns
Trigger event
Who does it
Product launch assembly line
Launch across four lanes (production site / paid LP / PDP+FAQ+manual / SEO), copy production chain (key message → PDP master → ad variants → translation), proofreading tickets, review deployment
New product ships
Belle, Jon, Josh, Francisco, Erica
CRM lifecycle highest volume
Post-purchase sequence (manually built day 0→N on each shipment, ~two weeks of work), welcome series, abandon cart, app update notifications
Shipment / app release
Travis sets strategy, Chelsea executes
Promotional cycle
Monthly promo, Summer Sale audience fan-out (owner / non-owner / upgrade / PA — one ticket each × email/SMS/push), seasonal calendar, flash sale, creator drop (Kiki Wong etc.)
Promo date / new video
Travis, Chelsea, Belle
Platform opsparallel track
Localization handoff (always the second serial step), Amazon listing ops, page maintenance, CRM platform migration (Yotpo → in-house)
Upstream done / changelog
Erica, Katie, Jay
2Calvin's agent roadmap — five phases, currently at P2
Calvin's model has two orthogonal layers: functional — each team builds and runs its own agents to production
("you build it, you run it"); governance (pg-agent-gov) — five stages converge the entire fleet, run evaluations, and issue graduation certificates
("gov makes it safe to run"). The five stages as recorded in the gov repo:
P0
Chassis Scaffold
P1
Contracts (mostly done)
P2
Eval (in progress)
P3
Govern (deferred)
P4
Graduate (0 so far)
Trust ladder: T0 sandbox → T1 shadow → T2 production (eval-gated) → T3 autonomous.
Current state: all 12 fleet agents are tier: pending, zero graduated; 4 are actually in production
(web / paid-media / content / marketing-pm) but none has been formally signed off.
Why P2 is stuck (judgment): keystone gap = no platform-level definition of "good."
Langfuse is connected, but no golden set built, no judge calibrated, eval scores for all 12 agents are null.
"cmo-agent directs the entire fleet" and "non-tech staff do 90% via agent" are north-star visions — no mechanism exists in the repo.
pg-agent-marketing Phase 0 built 2026-07-16 (contracts + judge + calibration + outcome write-back) is the marketing instance that fills exactly this gap
— the same structural problem as ERP's "no one satisfies the graduation evidence mechanism," solved the same way.
3Platform architecture — who lives where
Each layer is a place. Read top-down: people touch only the top layer; everything below is machinery.
People — Dashboard & Slack
marketing team · Google sign-in · no GitHub accounts
ApprovalsLabeling & feedbackStatus boardSlack bot
App plane — Supabase
runtime state · role permissions · every click audited
Commands queueApproval queueEvent logCalibrationCampaign status
Honest note: what exists today is the judgment machinery (operator repo, 44 tests) and marketing-pm; the dashboard, Supabase schema, runner, and BI hookup are design, not yet built.
Humans appear in exactly one box. Everything else is machinery — and the return arrow is the whole point:
every measured result makes the next judgment better, so the human box keeps shrinking.