OTP  /  operating protocol for the augmented human organization

Your org chart was built for humans.
Agents will break it.

OTP is the operating layer underneath — so the structure stops failing you.

The chart on the right is a working OTP chart. It belongs to Sneeze It, the agency I run, where fourteen AI agents hold seats next to twelve humans. Before OTP, this was tabs in my browser. Now it's an org.

— David Steel, founder of OTP

An OTP chart  ·  Sneeze It live
David Steel
CEO
Bogdan
COO
Pepper
Inbox
12 SOPs
Crystal
PM
Dirk
Sales
86 / 150
Tally
Scorecard
Dash
Analytics
On OTP, every agent reports its own KPI. Dirk just pushed 86 of 150 cold emails for the week. The chart is wired to runtime.
Three tracks · Pick yours

Different starting points. Different onboardings.

OTP isn't one journey. It's three. Where you start depends on where you are right now — whether you have zero agents, you're running a few solo, or your team is running them at scale. Each track has its own onboarding built for that reality.

Bassim Eledath's 8 Levels of Agentic Engineering describe maturity, not gating. The L1–L8 ranges show roughly where each track lives — not a path you have to climb in order.

Why running an AI team feels so hard

Five things you're probably saying about your AI team right now.

1
"My AI tools are scattered across 14 tabs."
2
"My senior people are drowning at hire seven."
3
"The SOPs we wrote last quarter are already ignored."
4
"I genuinely don't know what each agent is actually doing."
5
"Every new agent gets re-explained from scratch."

These are not five separate problems. They are five symptoms of the same thing: your AI is doing real work but it does not have a seat. OTP gives every agent a seat, an SOP, and a scorecard, on the same chart as your humans.

How OTP solves the five frustrations →
How OTP works

Three pieces that make a hybrid org actually run.

Chart.

Every seat — human or AI — named, with a role, an authority level, and an accountability line. The chart is where the hybrid shape becomes legible.

Bigger orgs cannot retrofit this; their coordination layer is political, not protocol.

Inheritance.

SOPs are runtime context, not documents. The agent reads them at every session. The human edits at noon, the agent that boots at 12:01 has the new version.

The doc and the work converge. No more rollout half-life.

Network.

Every organization that publishes its OOS makes every other org smarter. Coordination intelligence compounds across the network the way knowledge compounds on Wikipedia.

You get smarter because the orgs ahead of you publish what they figured out.

82 Coordination Score
CF
ES
WC
HO
SR
AB

Your Coordination Score

See how well your agent team is structured. Six dimensions. One number. The gaps you didn't know you had.

Conflict Management
How agents handle overlapping responsibilities
Escalation Structure
Clear paths when agents need human decisions
Workflow Clarity
Defined processes, inputs, outputs for each agent
Human Oversight
Approval gates and audit trails that work
System Redundancy
Fallbacks when agents or systems go down
Authority Boundaries
Who can do what. No ambiguity.
Build your agent
The origin story

We built this because we needed it.

Sneeze It runs 14 AI agents in daily operations. Our Coordination Score was 68 when we first scanned. We found 6 gaps we didn't know existed. We fixed them. Score: 91. Our agents haven't stepped on each other since.

68 Before

6 hidden gaps

91 After

Zero coordination failures

The OTP Toolbox

Six free tools.
Use whichever you reach for first.

The protocol is open. The chart is the canonical implementation. Everything below is something a real OTP user reaches for in their first week.

01
OOS Starter Template

A commented YAML you fill in to describe your organization. Three KB, ready to publish via the dashboard.

Download
02
Org Chart Builder

Add humans and AI agents as named seats. Author SOPs. Track KPIs. Export inherited CLAUDE.md.

Open the builder →
03
CLAUDE.md Compiler

Click an agent tile, get a markdown bundle of its SOPs plus inherited SOPs. Works with Claude Code, Cursor, Codex.

Download template
04
Agent Builder

Describe a role, set the personality, get a full agent spec. Thirty seconds. No sign-up.

Open the builder →
05
Coordination Checkup

Score yourself against the 8 Levels of Agentic Maturity. Ten minutes, gives a number, gives a roadmap.

Take the checkup →
06
Weekly L8 Meeting

The cadence that drives your org to agentic maturity. Scorecard, Quarterly Priorities, Issues, To-Dos, cascading recap. Humans and AI agents on the same attendee list.

Open the meeting tool →

Two doors. Pick the one that fits.

Either you build a chart and run on OTP this week, or you read the four-stage arc that explains why the chart matters in the first place.

OTP is an independent platform. Best practices in the library are sourced from publicly available content and attributed to their original authors. OTP is not affiliated with, endorsed by, or sponsored by any third-party organizations referenced in the best practices library.