RACI Chaos
Build a RACI matrix for a messy cross-functional project — assign Responsible, Accountable, Consulted, or Informed for each role on every decision, including an AI-agent role — then press run. The project plays forward and freezes at every stall your matrix actually produces: a decision with no owner, two owners in a standoff, or an owner with nobody to do the work. The reveal traces each freeze to the exact gap in your grid, and adds up the days it cost. Runs entirely in your browser (0 uploads, works offline).
How to use it
- Read the grid: each row is a decision, each column is a role (PM, Eng, Design, and an AI Agent).
- Click any cell to cycle its code — empty → R → A → C → I → empty — until it matches who really owns that call.
- Press Run the project. Rows that can't move light up red.
- Read the health score, the count of stalled tasks, and the days lost, then open the reveal to see the literal decision-rights gap behind each freeze.
What this clears up (the fundamentals)
- Accountable is not the same as Responsible — exactly one person is A (the single owner who answers for the outcome); one or more people are R (the ones doing the work). Confusing the two is the most common RACI mistake.
- A decision with no A never gets made — with nobody accountable, everyone assumes someone else will decide, and the task quietly stalls forever.
- Two Accountables is a deadlock, not a backup — when two people both think they own a call, each waits for the other; it's a collision, not redundancy.
- An owner with no R is a bottleneck — someone is accountable but no one is assigned to do the work, so it sits on a desk. Naming a doer unblocks it.
- AI-agent roles obey the same rules — giving an agent a lane doesn't remove the need for a single human A; it just adds another row that can stall if the rights aren't clear.
Where it's used
A leadership and operations drill for spotting decision-rights gaps before a project stalls instead of after. It's a miatz build-lab concept playable — a teaching widget paired with a build challenge, so you can learn to build the matrix analyzer yourself.
FAQ
What's the difference between Accountable and Responsible?
Accountable (A) is the single person who owns the outcome and answers for it — there should be exactly one per decision. Responsible (R) is whoever actually does the work; there can be several. A task needs one A and at least one R to move.
Why does two people marked "A" count as a problem?
Because accountability doesn't split. If two people are both "the owner," each assumes the other will make the call, and the decision deadlocks in a standoff. RACI wants exactly one A per row.
How is the health score calculated?
You start at 100. Each high-severity gap (no Accountable, or no Responsible) costs 20 points; each medium gap (two or more Accountables) costs 10. The score is clamped at 0. Stalled-task count and days-lost are derived from the same gaps.
Is anything uploaded?
No. The matrix, your clicks, the simulation and the reveal all run in your browser — nothing is transmitted or stored on a server. Turn off your Wi-Fi and it still works.
Limits
A teaching model: it checks the two rules that cause most stalls (one A, at least one R) and scores gaps with fixed weights. Real projects also care about who is accountable, escalation paths, and how consulted/informed roles are kept in the loop — none of which a matrix alone can judge. Use it to build the instinct for clean decision rights, then apply that judgement to your own org.
Related
Part of the Demystify Playgrounds. Explore the rest from the Playgrounds home.
Bookmark this page (Ctrl+D, or ⌘D on Mac) — it works offline the next time you need it.