Contents

What’s in this readout.

From the headline numbers to what comes next. Click any section to navigate there.

1. 01Executive summaryThe headline number and the decision
2. 02What was studiedOne function, four weeks, team-level only
3. 03Coverage & confidenceWhat we’re sure about, and what we’re not
4. 04Where the time goesThe £2.6M work catalogue, drilled to the task
5. 05The Work ProfileThe same work, by area, type and cost
6. 06How tools fragment workFour to five tools to handle one ticket
7. 07Hotspots worth fixingSix problems, £481k, three to do first
8. 08Grow the greenPutting freed time into high-value work
9. 09Opportunity portfolioNine plays, tracked, with detail cards
10. 10Decisions for leadershipFour choices — each with its proof
11. 11Proving it landed — PulseHow we prove it moved · continue in Pulse
12. 12Basis of calculationHow the numbers were worked out
13. 13Privacy & what we didn’t look atTeam-level only, nothing about individuals
14. 14What comes nextThe versioned document system

01Executive summary

Recover £300k–£450k per year, across three common activities.

Northstar’s board approved an AI and transformation budget after renewals slipped. Before the Q3 budget lock, the COO asked a simple question.

02What was studied

One function, four weeks, team-level only.

We looked at how the Service & Commercial Operations function actually works — never at how any one person works.

03Coverage & confidence

What we’re sure about, and what we’re not.

Of the work we observed, 82% cleared the evidence bar and is itemised in this report. The other 18% was emerging, suppressed or excluded — shown everywhere as a held-back allowance, never guessed or costed.

04Where the time goes

The £2.6M, all the way down to the task.

Every pound of the studied work rolls up into six areas, and drills down through 18 activities to 64 costed tasks. Click any row to open it; the bar shows its work-type mix.

The six areas are the reported work — the 82% that cleared the evidence bar (£2.6M itemised). The grey “held back” row is the ≈18% that was emerging, suppressed or excluded: shown for honesty, never itemised. Renewals and Support are the biggest areas — and where most of the recoverable time lives.

05THE WORK PROFILE

The same work — by area, by type, by cost.

This is the £2.6M as a picture: each column is a work area (width = its cost), split into its activities. Click any column to drill into that area’s activities and tasks. Greens are high-value work to protect & grow; ambers and reds are low-value work to shrink.

The six areas are the reported work (the 82%); the grey column is the ≈18% held back — emerging, suppressed or excluded, never itemised. The red isn’t a column of its own — it’s mixed into every area. The visual twin of the table in section 04.

06How tools fragment work

Four to five tools to handle one ticket.

The information already exists — it’s just scattered. People hop between systems to pull it together, and the two un-merged customer systems are the visible cause. About £260k of the £481k friction traces straight back to it.

A typical path through one request

The same customer detail is looked up, cross-checked and re-entered across two CRMs that were never merged after the acquisition — connect them, and a chunk of this disappears.

07 THE HOTSPOTS WORTH FIXING

Six problems, £481k. Three to do first.

Each is a repeated, everyday problem we saw often enough to be sure about — in plain language, with what to do about it. 

Named problems total £481k. That’s the ceiling on the £300k–450k we’d expect to actually recover — itself a slice of the £1.2M low-value load. The numbers stay honest as you zoom in.

08Grow the green

The point isn’t to cut. It’s to grow the green.

Recovering time is only half the story. The prize is putting that time back into the work that wins and keeps customers — moving the team from 52% high-value work to 60%.

Where the freed time goes

Three AI agents put recovered hours back into growth work — with a human leading every customer conversation. Revenue figures are illustrative, a scenario to pressure-test, not a forecast.

10Decisions for leadership

Four choices for the room.

Not an activity audit — a short set of decisions, each with the evidence behind it and the measure that will prove it landed.

11Proving it landed — Pulse

We saved a starting picture. Each quarter we check it moved.

The Study is the baseline. Pulse re-reads the same team the same way every quarter — proving the work shifted, not that a tool got adopted. The north star is more time on high-value work; the three priority problems are the early proof.

12Basis of calculation

The figures are capacity-equivalent estimates, not savings claims.

No black box. Every headline figure traces to costed tasks and a stated assumption you can challenge — built on a loaded team-cost basis (£71k blended), never individual pay or performance data.

What the low and high ends assume

13Privacy & what we didn’t look at

Team-level only. Nothing about individuals.

We only show patterns we saw often enough to be sure about, and only at team level. Nothing about individuals, and no group too small to be safe, is ever shown.

14What comes next · the document system

This readout sits in a versioned system of documents.

The immediate next step is to act on the decisions in §10 and commit to the quarterly Pulse in §11; a second function can be studied in parallel using the same method. Each artefact below answers a defined question about scope, method, evidence, governance, or what comes next — this readout is the leadership artefact, the rest are the system around it, versioned, owned and access-controlled.