The AI operating-model diagnostic for companies of 20–200 people

Get full clarity on how your company creates value, so you can prepare it for the AI-first future

The first step in any AI strategy is getting complete clarity on who does what, how. Mapping information, identifying opportunities and being 100% clear on how the work gets done inside your team. Iris interviews everyone, builds the operating-model map no one has ever had, and gives you a costed, ranked path to your AI-native future.

About 7 days end to end 20 minutes per employee The report is yours to keep

A working model of your company.

Every workflow, activity, system and handoff, rebuilt from what your team tells us.

Today’s operating model

24.0 FTE · £1.14m
Finance
Operations
Sales
Service

Corroborated by 8 colleagues · 2 discrepancies resolved

“

From the demo report

Accounts payable consumes 3.6 FTE today.

AI-accelerated residual need: 0.9–1.4 FTE for exceptions and supplier relationships.

Capacity released: £102k–£128k / year

Confidence 84% · every figure traces to interview evidence

Iris interviews your team where they already work

Web chat Email reply Slack DM Microsoft Teams Web chat Email reply Slack DM Microsoft Teams Web chat Email reply Slack DM Microsoft Teams Web chat Email reply Slack DM Microsoft Teams

The org chart is not how work gets done.

Most AI plans start from job titles and guesswork. The real operating model lives in hundreds of small activities nobody has written down. That is what Overture maps.

Coverage

100%

Every employee interviewed. Not a sample or a survey. The whole company, corroborating each other.

Effort

20 min

Per person, once. A relaxed conversation on web, email, Slack or Teams. No workshops, no prep.

Deliverable

3

Costed futures. AI-Capable, AI-Accelerated and AI-Native, each with a ranked build roadmap.

Overture models your company as a recipe. We map every workflow, every handoff, every decision and every goal, so you can see where your people are needed most.

What you get

One diagnostic. A working model of the whole company.

Not a slide deck of use-cases. It is a working model you can explore, with the economics behind every claim.

Org map

TodayAI-Native
Invoice matchingAgent
Supplier escalationsHuman
Month-end reportingHybrid
Credit decisionsHuman

The living org map

Explore your organisation activity by activity, then morph it between today and each AI future. Click anything to see who does it, how long it takes, and why.

Capacity released / yr

Ranges
AI-Capable£85k–£160k
AI-Accelerated£180k–£290k
AI-Native£260k–£395k

Adjustable assumptions · deterministic arithmetic

Board-ready economics

AI judges the qualities. Plain arithmetic does the money. You get ranges, not false precision, with every assumption on a slider.

Iris interviews everyone, like a colleague, not a form

A real conversation that adapts to each answer and digs into how work actually happens.

  • Web chat, email, Slack or Teams. The conversation follows people across channels
  • Colleagues corroborate each other; conflicts get clarified automatically
  • Employee-safe by design: Overture maps the work, never the person

Iris

You mentioned chasing sign-offs. Roughly how many hours a week does that take?

Honestly? Three or four. Fridays are mostly approvals.

Logged: approvals bottleneck · 3–4 h/wk · needs corroboration

Opportunity

Automate invoice matching

£102k–£128k

Activity

Match supplier invoices to POs · 3.6 FTE

Evidence · AP team

“Most of my Tuesday is copying invoice numbers into the ERP.”

Every number traces down to something a person said

An auditable model, not a hallucinated deck. Click any figure in the report and follow the chain all the way to the interview transcripts.

  • Opportunity → activities → the exact employee statements behind it
  • The AI never does the arithmetic. Scoring and economics run on deterministic code
  • Weak evidence is labelled as weak, not laundered into confident numbers
How it works

From kickoff to roadmap in about a week

1. Iris interviews

Everyone gets a 20-minute conversation in the channel they prefer. Iris adapts her questions to each role and follows up where answers conflict.

2. The model reconstructs

Answers become activities, workflows, systems, decisions and handoffs, cross-checked between colleagues until the picture holds together.

3. The futures get priced

Deterministic scoring and economics turn the model into three costed operating futures and a ranked list of build opportunities.

AI-Capable

Your structure, properly equipped with AI tools and training.

8–15%

capacity typically released

AI-Accelerated

Workflows redesigned around AI; people supervise and handle exceptions.

25–45%

of mapped work typically addressable

AI-Native

Designed from scratch around outcomes, as if founded today.

Day 1

a different company, fully costed

These are typical ranges from the model. Your report shows your own numbers, with the evidence behind them.

FAQ

Common questions, straight answers

Something we haven’t covered? Ask the humans at . We reply within one business day.

Contact us
What do we get at the end?

A board-level report and an interactive model of your organisation, plus a ranked list of opportunities. Each one ships with its implementation scope already written. can build them for you, but the report is yours to keep either way.

How long does it take?

About seven days end to end for a company of 20–200 people. Interviews run in parallel across channels, reconciliation and modelling happen as answers arrive, and the report is ready at the end.

What do our employees actually experience?

A single 20-minute chat with Iris about how their work actually happens, in the channel they prefer. It is never about performance, and nobody is being evaluated. There is a page they can read first: "For employees".

Do we need to connect our systems?

No. The diagnostic needs nothing but the interviews. Slack, Teams and email are optional channels for convenience. There are no integrations to install and no data pipelines to approve.

How is the data handled?

Interview data stays inside your workspace and is used only to build your model. Every fact is evidence-linked, so you can see exactly where a claim came from. We delete everything on request.

What happens after the report?

You get a ranked list of opportunities, each priced with a build scope. Build them with , with your own team, or with anyone else. The model does not lock you in.

In seven days, you’ll know what your company could become.

Deploy Iris, get the working model, and see the three futures priced against your own operations.

Start a Report
Built for companies of 20–200 people No systems to connect