Protocols for Business · Kickoff

Kickoff · Monday 2 November 2026

Protocols for Business

The next year’s learning plan, and putting protocol vision to work as agents join business operations.

Discovering New Nature

Watch the film full screen

TL;DR

Let's get good at seeing protocols.

  • WhyAgents are making the formal and informal rules of business matter, and very few people can see them well yet.
  • WhatEvery other Monday, 15:30 UTC on Discord, 26 sessions in all. One reading each time, read together. Recorded.
  • You getA skill, a case study from your own work, and a small group of friends who care about this too.
  • We askCome when you can, post what you notice, and carry one case if you have one. That's it.

Where we're coming from

We've spent a year and a half reading wide.

Tap every season you were there for.

What it taught us

Five lessons shaped this year's plan.

Tap the question you most want to dig into this year.

  • Some tensions are handled in the moment; others are built into structure.→Which rules should be hard, and which can stay soft?
  • Good protocols produce non-events, and nobody counts what didn't happen.→What can incident reports show us about everything that usually goes right?
  • Protocols show up when tempo changes: Slack in the office, Boom's R&D against its supply chain.→What shows up when agents speed everything up?
  • Standards spread through incentives and narrow scope, then lock in.→What do our systems emit, and how do we keep rules open to change?
  • Working with agents is management; the scarce work is gathering context.→Who gathers that context where you work, and could it be you?

What's different now

Agents won't ask for manager approval.

A small team can now run thousands of agents. An approval step can't keep up, so work flows around it. What holds is what's built in: into every agent, and into the ground they work on.

This year's question

How will AI-native organizations run their operations?

Tap the themes you're most curious about.

  • ThemesSkillDate
  • I. Working alongside agentsReading what agents can see, do and spendNov–Dec 2026
  • II. Natural coordinationSpotting coordination without a managerJan–Feb 2027
  • III. EmissionsMapping what systems publishFeb–Apr 2027
  • IV. Seeing hardnessReading failures and successes for what heldMay–Jul 2027
  • V. Engineering hardnessChoosing what to harden, and howJul–Sep 2027
  • VI. Operational livenessKeeping rules open to changeOct–Nov 2027

What you'll build

By next autumn, you'll see protocols everywhere.

  • SeeSpot the rules a business runs on, written and unwritten.
  • ReadRead incidents and agent traces for where rules held or broke.
  • DesignDecide what to harden, what to keep soft, and try it.
  • WriteDescribe a messy system so the people inside it recognise it.

Your case study

Bring one case: a client, or your day job.

Pick an organization you can see into, and one data operation AI is changing. Carry it for about a year: choose, brief, intervene, gather evidence, publish.

That's how last year's cases started: water rate data, construction bids from PDF drawings, a brand kit agents build from, and how people actually co-work with AI.

Introduce yourself

Your AI assistant drafts it from what it knows about you: name, website, and a case you might carry. You check it before anything is sent.

Put it to work

If you want, turn it into paid work.

Protocol Vision Advisory is an independent network of consultants and teams, built on this group's open research. Any member can become a network consultant; nobody has to.

  • Work comes inA coordinator takes each inquiry and picks a lead by fit. The lead builds the team.
  • One ruleA field log from day one, shared with the client.

Interested?

Next steps

Let's get started!

Write a blyg, blog or Substack?

Add it to the community feed, so the group reads your posts alongside ours.

Introduce yourself

Your AI assistant drafts your intro from what it already knows about you. You check it, then send it to the hosts.

  1. Paste it into ChatGPT, Claude or any assistant that knows you.
  2. Paste its answer below, check it, and send.
Read the prompt
I'm joining the Protocols for Business research group (https://protocolsforbusiness.com/). We read together about how organizations coordinate through protocols as AI agents join the work. Each member can carry a case study for a year: an organization they can see into (a client, or their own employer) and one data operation AI is changing there.

Using what you know about me (your memory of me, our past conversations, my work and projects), draft my intro for the hosts in exactly this format:

Name: my name
Website: my personal site, blog or profile (leave blank if you don't know one)
Case study idea: two or three sentences on the organization, the data operation, and what I'd want to find out

If you don't know enough for the case study idea, ask me one or two questions first. Leave out anything confidential about my employer, clients or other people. Show me the draft and wait for me to confirm.

Only the hosts see this.

Add your blyg, blog or Substack

Your posts join the community on the Feed. We check that the address has a blyg or an RSS feed (every Substack does), then the hosts add it.

Your name, contact and address go to the hosts in the group's Discord channel.