OrgTP User Guide

OrgTP is two things in one place:

1. An operating system for running your company. It gives your team the weekly rhythm that high-performing companies use: clear quarterly goals, one weekly scoreboard of numbers, a tight weekly meeting, a running issues list, and shared to-dos. If you've heard of EOS (the "Entrepreneurial Operating System" from the book Traction), OrgTP runs that playbook for you in software.
2. A home for your AI agents. Most companies now have AI tools doing real work (drafting, analyzing, prospecting). OrgTP puts those agents on the same org chart as your people — each gets a seat, an owner, and a job — and gives them a shared, reusable memory called an OOS (Organizational Operating System) so they coordinate instead of each starting from zero.

You can use just the first half (run your team) and grow into the second half (coordinate your agents) whenever you're ready.

The five things you'll touch every week:

OrgTP uses Google sign-in (and email/password). Go to orgtp.com, click Sign In, and choose Continue with Google or enter your email and password.

• If someone invited you, click the link in your invite email first — it ties your login to your seat in their organization.
• Your login identity (email, password, two-factor) is managed on the Account settings page later (see §13).

The first time you create an organization, OrgTP walks you through seven quick steps. Every step after the first has a "Skip for now" link and a "Skip → take me in" escape to the dashboard, so you can do the bare minimum and fill the rest in later. You can always return to any of these later from inside the app.

Step 1 of 7 — Your profile (this creates your organization)

Name your company (company name, industry, team size) and claim your own seat (your name and your role). Submitting this step is what brings your organization into existence.

Step 2 of 7 — Your team

Add coworkers by name, role (Team member / Manager / Admin / Observer), and email. Each one gets an email invitation and shows up on your chart marked "Invite sent."

Step 3 of 7 — Your goals (Rocks)

Add the one or two big outcomes the company is driving toward. Give each a title, a timeframe (This year / This quarter), and an owner. These become your Rocks.

Step 4 of 7 — Your KPIs

Define the numbers that each seat watches. For each KPI: a title, an optional target, an owner, and a cadence (Weekly / Monthly). This is the start of your Scorecard.

Step 5 of 7 — Your agents

Give your AI tools real seats. For each agent: a name, what it does, its runtime (Claude / ChatGPT / Zapier / Make / Other), and an owner. This is what makes OrgTP different from a plain EOS tool.

Step 6 of 7 — Your teams

Group seats into teams (e.g. "Leadership," "Web Team"). Each team gets its own meeting and scoreboard. Name the team and tap the members who belong on it.

Step 7 of 7 — Your first meeting

A recap of everything you set up, plus a preview of your weekly meeting (defaults to "[Company] Leadership Team," Mondays, 30 min) and its four-step agenda. Click "Start your first meeting" or "Explore your chart first" to enter the app.

Where: /dashboard — the screen you land on after signing in.

This is your personal cockpit. It greets you by name and stacks everything you need for the day into panels:

• Meeting — your next (or most recent) meeting, with a Live / Completed / Cancelled badge.
• Headlines — recent customer/employee news.
• Quarterly Priorities — your Rocks and how they're tracking.
• My To-Dos — your action items, with a quick-add box right there.
• My KPIs — your numbers in a small scoreboard table.
• Issues (IDS) — the issues you own or that need attention.
• My Agents — the AI seats you own.
• Leader insight panels (if you're an owner/integrator) — "Waiting on others," "Founder Dependency," "Accountability Gaps," "Hand-off candidates," and a "Cascading message" box.

Controls: switch between organizations or teams (top of the page), toggle "show only mine," and open any meeting or to-do directly.

This is where the whole system pays off: one timed screen your team runs its weekly meeting from, so meetings start and end on time and nothing falls through.

5a. The meetings list

Where: /l8 (the old /l10 address redirects here).

Lists your past and upcoming meetings, each with a status badge (scheduled / in progress / completed). Use the New Meeting form to create one (title, date/time, type, recurrence, team), then Create + Open. If you're on more than one team, a team filter appears. Click the "?" for an explainer of the meeting types (Weekly Leadership, Departmental, Quarterly, Annual, One-on-One).

5b. The live meeting runner

Where: /l8/meeting/[id] — opens when you click a meeting or create one.

Everything for the meeting lives on one page: the title and time, Start / End controls, an "Add to calendar" panel, a "Driving toward" strip (your longer-term targets), the attendee list, an optional Agent Reports panel (a summary from your AI seats), and a sticky agenda bar with a per-section timer.

Tip: Click Start meeting at the top. This freezes a snapshot of your scorecard and Rocks so the meeting becomes the official record of where things stood. A completed meeting is read-only.

The agenda runs in seven sections, in this order:

1. Check-in / Segue (5 min) — each attendee shares a personal best and a business best. Warms up the room.
2. Scorecard (5 min) — read this week's KPIs, each shown on-track (green) or off-track (red) against its goal. No discussion — just read the numbers. Anything off-track becomes an issue later.
3. Quarterly Priorities / Rocks (5 min) — each Rock with its owner, due date, and an On Track / Off Track toggle. Mark complete, reorder, or add a new one. Off-track Rocks become discussion material.
4. Customer & Employee Headlines (5 min) — short good/bad news items. Mark each "Addressed" or convert it "→ Issue."
5. To-Dos (5 min) — the 7-day action items from last meeting, with done checkboxes. Quick-add new ones (owner, due date, priority, recurrence).
6. Issues / IDS (60 min — the core of the meeting) — your issues list. Work each one through Identify → Discuss → Solve. Set its priority, assign an owner, convert it to a to-do, or solve it (which prompts for a resolution note and an optional follow-up to-do).
7. Conclude (5 min) — a recap of to-dos created today, a 1–10 meeting rating from each attendee (aim for an 8+ average), and a Cascading message (auto-drafted from the meeting) you can send to the wider team.

Where: /me/todos — your personal action queue, separate from the dashboard.

Sections you'll see: a create box at top, "⚠ Verify these happened" (items you delegated that are awaiting your sign-off), your open queue, "Waiting on others" (things you delegated out), an L10 section (to-dos tied to a meeting), and "Closed last 24h."

What you can do: create a to-do (title, optional description, priority P1–P4, due date, assign to one or more people, recurrence, optionally attach it to a meeting); mark Done; "→ L10" to promote it into the leadership meeting; Edit; Delete; and Verify or Reopen delegated items. On mobile there's a sticky quick-add bar and swipe-to-complete.

Where: /dashboard/kpis — the full company scoreboard.

Every measurable number in the company, organized by the seat that owns it. A tab strip switches between Trends / Weekly / Monthly / Quarterly / Annual views; a filter row lets you scope by team, owner type (All / Human / Agent), and date range. KPIs appear in tables showing Title, Owner, Goal, and either editable per-period value cells (green = met, red = missed) or a latest value with a sparkline trend.

What you can do: type actual values straight into the period cells to record results each week, assign a KPI to a person with a goal, and jump to the org chart to add new KPIs.

OrgTP doesn't have a separate "Rocks page." Your Rocks show up in three places:

• During setup (Step 3 of the wizard).
• On your Dashboard, in the "Quarterly Priorities" panel.
• Inside the weekly meeting, section 3, where you review and update them.
• On each person's profile page (see §10), listing the Rocks they own.

So to manage Rocks day-to-day, open your weekly meeting or your dashboard panel.

Where: /dashboard/ids — a cross-team master list of every open issue (visible to owners/admins/managers).

Filter by team and by IDS status (open / identified / discussed / solved). Each row shows the issue, its status, owner, team, and date. You can reassign an issue to a different team. Most issue work happens inside the meeting (§5b, section 6); this page is the bird's-eye view for leaders managing several teams.

Org chart

Where: /dashboard/team — a visual, editable chart of your humans and AI agents and who reports to whom.

Filter by All / Agents / Humans. Click a tile to edit a seat or see its meetings and accountability. Add a human or agent, invite by email, or import people via CSV. Edits save to a draft until you click Publish — publishing turns the chart into an OOS (see §11).

Open seats (and claiming one)

A human seat can be marked open: a real role with nobody in it yet, like a CFO you are still hiring. Tick Open seat (no person in it yet) in the seat editor and the tile carries a quiet open pill. Marking a seat open clears its email, phone, and Slack ID, because an open seat is a role, not a person.

When a person comes on board, open their seat and click Claim this seat, then pick them from your members. Their name and email land on the tile and the seat stops being open. If they are not a member yet, invite them instead: the invitation holds the seat until they accept. When someone leaves, Vacate seat does the reverse: the role stays on the chart as an open seat and its contact details are cleared.

Open seats are what a Blueprint carries (see §16), so a chart made template-ready is simply a chart whose seats have all been vacated.

Members & access

Where: /dashboard/members — invite people and control what each can see and do.

Shows active members and pending invites. Invite by email with a role, team assignment, and fine-grained Feature / Data / Agent access toggles. Edit, resend, or revoke. Roles range from Owner and Admin down to Observer and Free — the dropdown explains each.

Teams

Where: /dashboard/teams — create sub-teams so meetings, issues, and KPIs can be scoped to a group instead of the whole company.

Create a team (name, short slug, description), then add or remove members.

A person's (or agent's) profile

Where: /team/[seat-id] — open by clicking a name anywhere.

Shows what that seat owns: Quarterly Priorities, Responsibilities, Seat Fit (the "right person, right seat" rating: Gets it / Wants it / Capacity), open To-Dos, open and solved Issues, and their meeting history.

People Review

Where: /team/review — rate your direct reports on seat fit and your company's core values (the EOS "People Analyzer").

A grid: one row per person, columns for Understands / Wants / Capacity plus each core value, and a computed verdict (Solid / Needs conversation / Wrong seat). Click a cell to cycle its rating.

This is the part that turns your operating data into reusable intelligence for your AI agents and even across organizations.

Key idea — the OOS (Organizational Operating System): a structured file of your company's coordination rules and knowledge — who owns what, how decisions get made, what's failed before, and the lessons learned. Each rule in an OOS is a Claim, tagged with how confident and how evidence-backed it is. Agents read your OOS so they start informed instead of from scratch.

Operating Plan

Where: /dashboard/oos-operating-plan — a guided business-planning workspace (vision, traction, strengths/weaknesses) that turns your strategy into a 90-day execution list and into OOS claims. Create a plan, fill in the tabs, generate a 90-day plan, and (for admins) push it to your OOS.

CEO view

Where: /dashboard/ceo (owners/visionaries only) — the whole company on one screen: long-term direction, scoreboard health, this quarter's execution, the team, and what needs attention. Driven entirely by the Operating Plan above.

Browse coordination knowledge (Claims)

Where: /claims — a public, browse-by-topic index of coordination knowledge published across every organization on the platform. Click a section to read its claims (each with confidence, evidence, and a rule / why / failure-mode).

Intelligence graph

Where: /graph — an interactive map of how organizations connect through shared patterns and best practices. Switch modes (Intelligence / Infrastructure / Best Practices), click nodes for detail, and in Best Practices mode you can ingest a practice into your own org as a claim. (The graph data requires sign-in.)

Source documents

Where: /dashboard/source-documents — upload your own raw docs (a playbook, SOPs, a CLAUDE.md) and OrgTP extracts coordination intelligence into OOS files. Upload/paste, click Process, then open the generated OOS.

Switching from another EOS tool? OrgTP imports your data.

Where: /import/ninety (for Ninety.io) or /import/bloom (for Bloom Growth). Both work identically.

1. In your old tool, export your Rocks, To-Dos, Issues, Headlines (as spreadsheets) and your Scorecard (as CSV) — one team at a time, and widen the date range first.
2. Drag those files into the drop zone (up to 12 files).
3. You get a free, private preview — a "What we found" count and a rebuilt roster showing who owns what. Nothing is stored yet.
4. Click Import into my workspace to commit. (If you're not signed in, it'll prompt you to sign in, then you re-drop and re-click.) Re-running won't create duplicates.

Reach these from the settings menu. All but one are fully functional.

This is how an agent (e.g. Claude Code, ChatGPT) actually reads and writes your OrgTP data.

Recommended: Remote MCP connection (no install). On a paid plan, open /settings/api, click Create connection under Remote MCP connection, and copy the hosted URL (it's shown only once). Paste it into your client — in Claude, Settings → Connectors → Add custom connector. No terminal, no config file. Turn on Remote MCP connection in Settings → Labs first.

Where: /settings/api

1. Click Generate New Key, give it a name.
2. Copy the key immediately — it's shown only once.
3. The page includes a copy-paste MCP setup snippet for Claude Code so your agent connects in one step.
4. Revoke any key anytime.

Read-only tools work without a key; publishing to your OOS requires one. MCP (Model Context Protocol) is the standard that lets an AI agent talk to OrgTP's server safely.

Want the fast path? The Connect Your Agent page has a copy-paste AI prompt that generates your first OOS in about 60 seconds, plus a one-line install that gives Claude Code the OTP MCP server and slash commands.

Meetings evaporate without capture. Ollie Insights™ reads your meeting transcript and turns it into the team's follow-ups automatically.

Where: the Follow-ups page (/l8/meeting/[id]/followups), reached from the Meetings list after a meeting ends.

1. End the meeting. A recap email lands in every attendee's inbox.
2. Add the transcript on the Follow-ups page — paste it, or upload the file from Plaud, Fireflies, Gemini, or your own notes.
3. Press Generate Ollie Insights. Ollie pulls out the to-dos, the issues, and the headlines, and writes a short insight that opens your next meeting.
4. Review and accept. Each piece Ollie found is accepted with one click.

Credits: your first Ollie Insights generation is free. After that, each run draws down credits from your organization's prepaid wallet — top up or set auto-recharge on the Billing page (/settings/billing).

Where: /blueprints. Blueprints is currently a Labs (beta) feature: switch Blueprints on in Settings, Labs and it appears in your sidebar.

A Blueprint is a versioned package of how an organization runs, captured from a real working account and applied to others. Configuration travels. Data never does.

What travels

What never travels: people, live KPI values, meeting history, to-dos, keys, and billing. Seats travel as roles, never as the people sitting in them.

Open seats: how the roles arrive

Because seats travel as roles, an organization created from a blueprint lands with a chart full of open seats: every seat is there, nobody is in one. That is the honest state, and it is the whole point.

As each person comes on board, open their seat on the org chart and click Claim this seat (see §10). Their name and email land on the tile, the seat stops being open, and any placeholder the chart was holding for that role folds into their account.

Who can use it

You can apply a blueprint to an organization you own, one you coach (an active coach grant), or one that sits in your portfolio. Nothing else is reachable.

Creating one, and versions

Create a blueprint from a source organization: a real working account, not an empty one. That capture becomes version 1. When the source moves on, capture a new version to re-read it. Every version carries a summary of what was added, changed, and removed.

Applying

Applying always starts with a preview. It classifies every asset as new, unchanged, a safe update, edited locally, or a conflict. Safe updates land. Local edits are preserved unless you choose Overwrite. Every apply can be undone: the undo restores the assets it overwrote, and it does not delete assets the apply created.

Linked organizations

Each organization you apply to becomes linked to the blueprint. When a new version is ready, you can push it to many linked organizations at once: one set of choices, one result per organization.

Sharing

A share link is permanent. Whoever opens it gets a frozen copy of that version in their Imported tab. An imported blueprint cannot capture new versions, because it has no source organization of its own.

Publishing as a template makes the blueprint visible to every OTP organization, and a recipient can even create a brand new organization straight from it.

The portfolio door: an operating model

A portfolio (a group of companies you own or run) can designate one blueprint as its operating model: the way every company in the portfolio runs. Every company the portfolio seeds is created from it, and every company it owns is measured against it. You manage it on the portfolio page (/portfolio), not here.

Offered or enforced. An offered model preserves each company's local edits: updates are proposed, and only items the company never touched are updated. An enforced model converges: on the next push, an item a company edited is replaced with the model's version. Enforcement can be scoped to specific asset types, so a portfolio can enforce one scorecard while each company still phrases its own values. Either way every push is recorded and can be undone.

Compliance. The portfolio page shows one row per company: the version it sits on, whether it is behind, and which asset types it has drifted in. Resync all companies pushes the latest version everywhere at once; Update pushes it to a single company.

Why the numbers line up. Companies born from the same operating model share their KPI definitions, so a portfolio roll-up does not have to be mapped by hand: Auto-map from operating model finds the matching KPI in every member company and wires it into the consolidated metric.

Numbered steps and a short FAQ live on the Blueprints guide page.

• Agent — an AI tool given a real seat on your org chart, with an owner and a job.
• Cascading message — a short summary passed down from a meeting to the wider team.
• Claim — a single coordination rule inside an OOS (a Rule, why it matters, its failure mode, and its scope), tagged with confidence and evidence level.
• EOS — the Entrepreneurial Operating System, the business-management framework (from the book Traction) that OrgTP's operating side is built on.
• Headline — a short customer or employee news item shared in the weekly meeting.
• IDS — Identify, Discuss, Solve: the three-step way OrgTP works an issue.
• Issue — anything that needs to be resolved as a team.
• KPI — Key Performance Indicator: a number a seat watches each week.
• L8 / L10 / Level 10 Meeting — the structured weekly leadership meeting. (OrgTP labels it "L8"; "Level 10 Meeting" is the EOS term.)
• OOS (Organizational Operating System) — a structured file of your company's coordination rules and knowledge that AI agents read.
• Open seat — a seat on the org chart with nobody in it yet: a real role you have not filled. A Blueprint carries seats as open seats; you claim one for a person when they come on board.
• Rock / Quarterly Priority — a major goal committed for the next ~90 days.
• Scorecard / Scoreboard — the weekly table of your KPIs vs their goals.
• Seat / tile — a position on the org chart (human or agent), with responsibilities and an owner.
• Seat Fit (GWC) — "Gets it, Wants it, Capacity": whether a person is right for their seat.
• To-Do — a small action item due within about 7 days.
• Workspace — a shared space for collaborating with other organizations on OOS files.