Join OTP the operating platform for people and AI agents
Back to Blog
Founder Notes 2026-05-22 · David Steel

The Process Component of EOS®, where SOPs become executable

The Process Component is the one EOS® component most companies under-invest in. Vision is fun. People is uncomfortable but rewarding. Data shows up in the L10®. Issues get the room. Process gets the leftover hour at the end of the quarter, if any.

This is the component that quietly determines whether AI integration works at your company.

Gino Wickman's framing: identify your handful of core processes (sales, marketing, operations, HR, finance, customer service, and any other unique to your business), document each one at the 20% that drives 80% of the outcome, and get them Followed By All™. The whole point of the Process Component is consistency, scale, and the freedom that comes from a business that does not depend on memory.

Agents need exactly the same thing. An agent reading a sharp, current SOP can execute the process. An agent reading a missing or stale SOP improvises, and improvisation by an agent at scale produces the wrong outputs faster than humans can catch them.

The Process Component is therefore not just a Wickman discipline anymore. It is your agent layer's source code.

What changes when SOPs become executable

Three changes, ordered from cosmetic to structural.

Cosmetic: SOPs move from Google Docs to plain text. Most EOS® companies keep SOPs in Google Docs or Notion. That is fine for humans. For agents, the easiest format is markdown or plain text in a folder the agent can read. Migration takes a day per process. The content barely changes.

Practical: SOPs get more specific. Human-readable SOPs can say "qualify the lead." An agent-readable SOP has to say "a qualified lead is one where (a) the company size matches the ICP list, (b) the contact title matches the buyer titles list, and (c) the form fields contain a stated budget over $X." Both are SOPs. One is executable. The forcing function of agents makes your SOPs sharper, which improves the humans following them too.

Structural: SOPs become living artifacts. A Google Doc SOP gets updated every six months, maybe. An executable SOP gets updated when the process changes, because the agent will start producing the wrong outputs otherwise. The SOP is no longer a binder on a shelf. It is the company's running instruction set.

This last one is the most underrated upgrade in EOS®. Many companies say their SOPs are Followed By All™ when in reality the SOPs are followed by no one because they were written two years ago and the team has been improvising since. Agents do not improvise. Agents follow what is written. The SOPs come back to life.

The six fields every agent-executable SOP needs

A human SOP can be a paragraph and a checklist. An agent SOP needs structure.

  1. Identity. Which agent owns this process. Reference to the seat on the Accountability Chart.
  2. Job. What the agent is supposed to do, in one or two sentences.
  3. Scorecard. The measurable output. Numbers, not adjectives.
  4. Data sources. Where the agent reads from to do the job. APIs, files, channels.
  5. Escalation rule. When the agent stops and asks a human. Specific triggers.
  6. Off-limits actions. What the agent must never do. Especially important. Most agent failures come from missing this field.

This is the format we use at Sneeze It for every agent SOP. The Integrator can read any agent's SOP in under three minutes and know exactly what the agent does and does not do. The Visionary can read the off-limits actions and feel confident the agent will not embarrass the company.

How to migrate your existing SOPs in a week

Most EOS® companies have eight to twelve documented core processes. Migrating them to executable format takes about a week of focused work.

Day 1. List every documented SOP. Print each one. Read them in one sitting. Note which ones are current and which are stale.

Day 2. Pick the three most-used processes. Rewrite each one in the six-field format above. Do not improve the content yet. Just restructure.

Day 3. Read each of the three through the lens of "could an agent do this." Identify the gaps. Most will be in the scorecard field (no measurable output) and the data sources field (process assumes the human can find the data).

Day 4. Fill the gaps. Sharpen the scorecard. Specify the data sources. Decide which actions are off-limits.

Day 5. Have the human who currently owns each process read the new SOP. Ask: "If I gave this to someone on day one, could they execute it." If yes, the SOP is ready to go to an agent. If no, refine until yes.

By Friday you have three executable SOPs. Do the next three the following week. By the end of a quarter you have all your core processes in agent-ready format.

You do not have to deploy an agent for each process. You just need the SOPs in the shape where you could.

What this does for the human team

The under-told story of executable SOPs is that humans benefit more than agents.

A sales rep reading an agent-ready SOP can execute the sale with full clarity. A new hire reading agent-ready SOPs on their first day can ramp in days instead of weeks. An Integrator can audit any process in three minutes instead of an afternoon. An Implementer can coach the Process Component meaningfully because the SOPs actually mean something.

The Process Component stops being the boring component. It becomes the leverage component.

Followed By All™ in a hybrid workforce

The classic EOS® line is "Followed By All™." Document the core processes, then make sure everyone follows them.

In a hybrid workforce, the line extends to "Followed By All™, humans and agents." The humans follow the process because the process is clear and current. The agents follow the process because the process is executable. Both halves of your workforce are operating from the same source of truth.

This is what real AI integration looks like. The same SOP runs both halves of the team.

FAQ

Do we need a tool for this? No. Plain text files in a folder will do. Notion, Google Docs, Obsidian, GitHub repos, all work. Pick a format the agent platform can read and the humans can edit.

Who owns SOP maintenance? The owner of the seat that runs the process. The Integrator owns the overall discipline. Same as classic EOS® Process Component ownership.

How often should SOPs be reviewed? Quarterly at minimum. Anytime the process changes in reality. If the agent starts producing wrong outputs and the process has changed, the SOP is behind.

What about confidential or sensitive SOPs? Same rules as any sensitive document. Use enterprise tiers of Claude or ChatGPT with zero training on customer data. Access-control your SOP files. The format does not change.

EOS®, Entrepreneurial Operating System®, Process Component, Followed By All™, V/TO™, Level 10 Meeting®, L10®, Rocks™, Scorecard, Accountability Chart, and EOS® Implementer® are concepts and trademarks of EOS Worldwide, LLC. This article is an independent practitioner perspective and is not affiliated with or endorsed by EOS Worldwide.

DS
David Steel

Founder of OTP. Runs an AI agent army at a digital agency. Building OTP because nobody else seems to be building it. Notes from inside the build, not from the conference circuit.

More about David →

More posts on the blog index.

All posts