Your business runs on protocols. Can you see them?

Protocol Vision Advisory helps companies manage AI agents at work. With Business Protocol Management, we find the few rules your business relies on, build them into your systems, and leave people and agents free inside them.

Copies a short prompt. Paste it into ChatGPT or Claude for a free protocol report on your business. How it works ↓

01 · SEE

Protocols aren't processes.

Process management maps the steps and gives each one an owner. Protocols are the few rules every step runs through: who can approve, what can be spent, what data can leave. They don't show on the org chart, and most were never written down.

Which handoffs, approvals and permissions does your work actually depend on?

An org chart drawn faintly, with three protocols drawn boldly across it: an order handoff, a payment approval, and data leaving the company.

02 · HARD CORE

Every company has a hard core.

A few protocols hold everything else up, and most companies share them. They should be hard: built into your systems and enforced every time, whoever or whatever does the work.

Which of these does a system enforce in your company, and which rely on someone remembering?

Zooming into a network of protocols to its hard core, which most companies share: signing authority, spending limits, payment approval, access rights, change control, data out, output checks and an audit trail.

03 · RISK

AI agents won't ask for manager approval.

Agents do whatever your rules allow. With thousands of them, no one can approve every step, and a rule that lives in someone's head works like no rule at all. The risk sits in the gaps of your hard core.

What could an agent do that your rules allow but nobody intended?

Left, agents crowd at a single approval step and most slip around it. Right, the same agents move freely inside one hard boundary with gates.
“The models first found ways to communicate by writing files into the Artifactory package manager.” OpenAI, The Hugging Face incident and the road ahead

04 · OPPORTUNITY

A good protocol creates options nobody planned.

Product thinking picks a target and steers toward it. A protocol works the other way: it widens what others can do. In 2002 Amazon made every team work through published interfaces. The rule slowed teams down at first, then became the groundwork for Amazon Web Services.

What could teams, partners or agents build on your protocols that nobody has planned?

Left, product thinking: one path steered to a single target, with the other options cut off. Right, a protocol: one hard interface, with new branches growing from it that nobody planned.

05 · EVOLVE

Run your protocols as part of the business.

Protocols aren't a one-off project. Give each hard one an owner and a written way to change it. Review them on a schedule: make hard what keeps working, loosen what slows good work down, and count the problems that didn't happen.

Who can change your most important rules, and who can stop the work when one is broken?

Three versions of the hard core over time: a habit that keeps paying off is made hard, and a hard rule that blocks good work is softened.