SOPs were written for humans. Your AI needs AOPs.
The procedures that run your business were written for people to read and follow. As AI agents start doing the actual work, they need a version written for them. That version has a name now: the AOP.
The procedure your business runs on was written for a person
A standard operating procedure assumes a human is reading it. It assumes that person can interpret a vague step, ask a coworker when something is unclear, and use judgment to fill the gaps the document leaves open. That is fine when a person runs the process. It falls apart the moment you hand the same document to an AI agent and expect it to act.
An agent cannot run an SOP the way a person does. It needs the intent stated, the decision points named, and the actions made explicit. The procedure has to be written for the thing that will execute it. For a growing share of routine work, that thing is now software.
What an AOP is
An AOP, or Agent Operating Procedure, is the agent-era successor to the SOP. Where an SOP tells a human how to run a process, an AOP is a natural-language procedure that compiles into a validated workflow an AI agent executes. It captures the intent, the decision points, and the actions, things like process a refund, verify identity, or update a record, rather than only describing them in prose.
It is an emerging term, pushed mainly by Decagon and also used by Skan AI and AthenaIntel. It is not yet a universal standard, and we do not pretend to have coined it. We use it because it names something real that is showing up across the AI space: the agent-runnable version of a procedure.
Why this is the natural next step for us, not a rebrand
Our whole approach already runs on one idea: your SOPs do double duty. Each SOP becomes a training course your team learns from and an automation that runs the task itself. The AOP is simply the third output from the same source. No rebrand required.
- 1Train. How your people learn the process. An AI-generated course with video, quizzes, and certification.
- 2Automate. How the routine task runs without them. A working automation in n8n, Zapier, or Power Automate.
- 3Operate. How your AI agent runs it. An AOP, the agent-runnable version of the procedure. agent-era
So the line evolves cleanly. Your SOPs do triple duty: train, automate, operate. Same source, three outputs.
AOPs are the bridge to owning your AI
This is where it stops being a buzzword. The private, company-owned AI system we are building toward in Own Your AI needs procedures to follow. An agent running on a box you own does not invent how your business works. It runs the AOPs you gave it. So the AOPs you build now, from the SOPs you already have, are the same procedures that later run on the system you own. The work compounds instead of resetting.
How to create an AOP
You do not start from scratch. You start from the SOP you already have and make it explicit enough for an agent to run. The steps:
- Start from the SOP. The written procedure your team already follows is your source of truth.
- State the goal. Name what a successful run looks like in one sentence, for example "issue a refund the customer is entitled to."
- List the decision points. Write out every place a person would normally pause and judge: the checks, the thresholds, the if-this-then-that. An agent cannot guess these the way a person can.
- Define the actions and tools. Spell out the concrete actions and the systems they touch: look up the order, verify identity, update the record, send the confirmation.
- Set the guardrails. Decide what the agent must never do on its own and where it has to hand off to a person.
- Test and validate. Run it on real cases, compare against how a person handled them, and tighten the steps until it holds. The AOP compiles into a workflow the agent executes, not a document it interprets.
The short version: an SOP describes the work for a person, and an AOP makes the same work runnable by an agent, with the judgment written down instead of assumed.
The honest part
AOP is early. The term is partly vendor-driven, and full autonomous agent operation is a direction, not a finished product you can buy off a shelf today. We are clear about that on purpose. The practical move is the same one we make everywhere: turn the SOPs you already have into agent-runnable procedures, start with the routine majority, automate what is safe to automate, and keep a person on the hard cases. The agent earns more of the work as it proves it can run the work.
That is the difference between chasing a trend and building on one. The trend is agents. The thing you build is the procedure they run.
Sources
Common questions
What is an AOP?
An AOP, or Agent Operating Procedure, is a procedure written for an AI agent rather than a person. It is the agent-era successor to the SOP. Where an SOP tells a human how to run a process, an AOP is a natural-language procedure that compiles into a validated workflow an AI agent executes, capturing the intent, the decision points, and the actions, not just describing them. It is an emerging term, pushed mainly by Decagon, and not yet a universal standard.
How is an AOP different from an SOP?
An SOP assumes a person reads it, interprets it, and uses judgment to fill the gaps. An AOP removes the ambiguity an agent cannot resolve on its own. It states the decision points and the actions explicitly so an AI agent can execute the process reliably, with a human handling the cases the procedure does not cover.
Do I need to throw out my SOPs?
No. Your SOPs are the source. QuarterSmart turns each SOP into three things from one source: a training course your team learns from, an automation that runs the routine task, and an AOP your AI agent can operate. Same source, three outputs. Train, automate, operate.
Is AOP a real standard yet?
Not a universal one. AOP is an emerging, partly vendor-driven term as the market shifts from workflows to agents. We adopt it because it names something real, the agent-runnable version of a procedure, but we do not claim to have coined it, and we frame full autonomous agent operation as a direction, not a finished product.
What does AOP stand for?
AOP stands for Agent Operating Procedure. It is the agent-era version of an SOP, a Standard Operating Procedure: a procedure written so an AI agent, not a person, can run it.
How do you create an AOP?
Start from an SOP you already have. State the goal, list the decision points a person would normally judge, define the concrete actions and the tools they touch, set the guardrails for what the agent must not do alone, then test it on real cases. The result is a procedure an AI agent can execute, not a document it has to interpret.
Are AOPs the same as agent prompts or workflows?
Not quite. A prompt is a single instruction and a workflow is a fixed sequence of steps. An AOP sits between them: a natural-language procedure with explicit decision points that compiles into a validated workflow an agent runs, so it handles branches a rigid workflow cannot and stays consistent in a way a one-off prompt cannot.
Can you give an example of an AOP?
A refund AOP: verify the order exists, check it is inside the refund window, confirm the customer's identity, decide full or partial based on the policy, issue the refund, log it, and notify the customer, escalating to a person on any exception. Each step is a stated decision or action rather than a sentence a human has to interpret.
Turn your SOPs into AOPs your agents can run
Bring one process your team runs by hand. We will show you how the same SOP can train your people, automate the routine, and become a procedure your AI can operate.