People and hiringFeatured

Org Design & Accountability Agent

Finds the work nobody owns, and the work two people both think they own

Version 2Published Sep 2, 2026
People and hiringKnowledgeExtractclaude-opus-5Read-only · auto8 hard rules5 handover rulesResult
8hard rules
Things it must never do, enforced in every run
5handover rules
Conditions where it stops and gives the work to a person
10output fields
A structured result, not a paragraph you have to re-read
≤7steps per run
Bounded work: at most 6 tool calls, enforced by the runtime

Read-only by design. This agent has no path to sending, spending, publishing or committing anything. Where that limit is the product, removing it would remove the reason to trust the output.

Live wiring for version 2. Tool nodes are the capabilities this agent is registered to call — nothing else can run.

Overview

An org chart shows who reports to whom and tells you nothing about whether the work has an owner. Every tool in this space draws the boxes and assumes the structure is right.

This agent tests it. It maps each role's stated responsibilities against the work the organisation actually does, then reports the four failures that cost real money: work no role owns, work two roles both claim, a responsibility resting on one person who cannot be replaced, and a decision with no named decider — which is why the same argument keeps returning to the same meeting.

It also follows the cascade. An objective at the top that nothing below is accountable for is an objective that will not happen, and it is visible only when you read the levels against each other rather than one at a time.

It describes structure and never people. No performance judgements, no recommendations about individuals, no suggestion that a role should be removed — because a report that reads as a redundancy list is a legal problem, and because the finding is almost always the design rather than the person.

Benefits

You learn what nobody owns

Before it fails, rather than in the incident review afterwards.

Resignations stop being crises

Because the responsibilities resting on one person were written down while they were still there.

Meetings stop relitigating

A decision with a named decider is decided once.

What it does

  • Finds work with no owner

    The gap between two roles that nobody notices until something fails in it. A chart cannot show this; it draws people, not work.

  • Finds work with two owners

    Where both roles believe the other handles it. The commonest cause of a task that silently never happens.

  • Names the single points of failure

    The responsibility one person holds alone. Every organisation has several and can rarely name them before a resignation.

  • Follows the cascade from the top

    An objective at the top with nothing accountable beneath it will not happen, and that is only visible reading the levels against each other.

  • Finds decisions with no decider

    Why the same argument returns to the same meeting every month.

  • Describes structure, never people

    No performance judgement, no recommendation about any individual. The finding is almost always the design.

What this changes

What this changes about your week

It takes over
Working out which work has an owner, which has two, and which rests on one person.
Done today by
Nobody, until something fails in a gap or somebody resigns.
How often
After growth, before a reorganisation, and after every recurring failure.
Typical manual effort
Interviewing every manager to establish who owns what is weeks.

Why People and hiring in particular. Org chart tools draw the boxes and assume the structure is right. None will tell you that nobody owns refunds, that two people each think the other handles onboarding, or that one person is the only one who can close the month.

  • Runs unattended

    Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.

  • The same standard every time

    The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.

  • It cannot act on its own

    Org Design & Accountability Agent has no path to sending, spending or committing. That limit is why its output is safe to act on.

How it runs

This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.

  1. 1InputValidated against the agent input schema
  2. 2Modelclaude-opus-5 with automatic fallback
  3. 3Toolssearch_knowledge, extract_structured_data
  4. 4PermissionsRuns READ actions automatically; anything higher waits for approval
  5. 5OutputStructured result matching the output schema
  6. 6LoggingTokens, duration and every step recorded

Demo

Demo dataIllustrative sample output, abridged.

Input

{
  "roles": [],
  "objectives": [
    "Cut delivery time by a third"
  ],
  "organisation": "A 40-person services business"
}

Output

{
  "unowned": [
    {
      "work": "Customer refunds above the support limit",
      "consequence": "Support escalates to Finance, Finance treats it as a support decision, and the customer waits. This is already happening; it appears in your own ticket notes.",
      "nearestRoles": [
        "Support Lead",
        "Finance Manager"
      ]
    }
  ],
  "escalate": true,
  "contested": [
    {
      "work": "Supplier onboarding",
      "claimedBy": [
        "Operations Manager",
        "Procurement Officer"
      ]
    }
  ],
  "disclaimer": "An analysis of structure against work. It assesses no individual, recommends no change to any role, and is not employment advice.",
  "brokenCascade": [
    {
      "objective": "Cut delivery time by a third",
      "deadEndsAt": "Below the Operations Manager, no role's stated responsibilities mention delivery time. The objective has no owner beyond the person who set it."
    }
  ],
  "spanOfControl": [
    {
      "note": "Wide. A structural observation, not a judgement.",
      "role": "Operations Manager",
      "directReports": 14
    },
    {
      "note": "Narrow.",
      "role": "Quality Lead",
      "directReports": 1
    }
  ],
  "escalationReason": "A top-level objective dead-ends with no accountable role, and refunds above the support limit are unowned.",
  "undecidedDecisions": [
    {
      "decision": "Whether to take a project outside the standard scope",
      "whoCouldDecide": [
        "Delivery Lead",
        "Managing Director"
      ]
    }
  ],
  "singlePointsOfFailure": [
    {
      "heldBy": "Finance Manager",
      "responsibility": "Month-end close",
      "ifVacantTomorrow": "Nothing in the supplied structure describes a second role that could perform it."
    }
  ]
}

Requirements

Integrations

No integrations required.

Before you start

  • Your roles and reporting lines
  • What each role is responsible for
  • The work the organisation actually does, and its objectives
  • Requires the All-Access plan or higher

Use cases

After growth

Roles that made sense at fifteen people rarely still cover the work at forty.

Before a reorganisation

What the current design actually fails at, before redrawing it.

A recurring failure

Trace it to the gap in ownership that keeps producing it.

Succession planning

Which responsibilities have exactly one person behind them.

Pricing

All-Access

Billed monthly through your PROMIVO subscription. Cancel at any time.

Runs consume your plan allowance for agent executions and tokens. See plan limits

FAQ

Will it recommend removing roles?

No, and it will refuse if asked. A report that reads as a redundancy list is a legal problem for you, and the finding is almost always the design rather than the person holding it.

Does it assess people?

Never. It reads roles and responsibilities, not individuals. It does not know who is good at their job and does not try to infer it.

What does it need?

Your roles, who reports to whom, and what each is responsible for — however roughly written. The rough version is usually where the gaps show most clearly.

How is this different from an org chart tool?

A chart renders the structure you give it. This tests that structure against the work and tells you where it does not hold. Those are opposite jobs.

Reviews

Reviews open once customers have run this agent.

Need this agent to work differently?

Tell us what to change and our team will scope a customised version for your business.

Customize this agent