INTERNAL · strategy notes by Poe for Armitage Rosen
Internal · the thinking behind these pages

Sell one door.
Let them find the house.

The beta already reveals itself one app at a time. The marketing should do the same thing. People don't buy platforms. They buy one job done well, and find the rest later.

The problem

Too much at once

11 apps, 6 clients, 5 infrastructure pieces, ranks, Credz and a funding ask, all on one page.

A vocabulary test

Nearly every name is a codename. Visitors have to learn the lore before they know what anything does.

Asking too early

Pledges up to $10,000 before anyone has used anything.

The test

Can you describe it in one sentence without "agent," "AI" or "chat"? If you can, it's a product. If you can't, it's still a platform.
Way inOne sentenceAudienceStands out
Agent seats ★"Your next hire still shows up to standup."Teams, agencies, startupsHigh
Standing orders"Write it once, and it gets done every week."Founders, consultantsMedium–high
Provenance"Proof of who wrote what, line by line."Schools, news, legalVery high
Your own Sprawl"Your own corner of the internet."Privacy-minded devsHigh
Crucible"Repos and packages, with reviewers that never go home."DevelopersLow–medium
Agent-to-agent"The meeting about the meeting, handled."Remote teamsVery high (later)

Recommendation

  1. Lead with agent seats. It's the clearest pitch, the most different from what's out there, and it naturally takes people through Cicero → Kairos → Alacrity → Crucible.
  2. Provenance is the trust story behind it, plus its own sharp way in for schools and newsrooms.
  3. Standing orders are the "what it does" for a seat, and the home screen shows results, not a chat.
  4. Agent-to-agent is the upgrade once a whole team is on.

Homepage fixes, whatever you choose