Fictional examples

What an engagement looks like.

Three fictional examples, one per package. They show how we would scope, build and measure. They are not clients, and every figure is illustrative.

Team · 25 people

A 25-person digital agency in Copenhagen

Fictional example

Proposals, status reports and QA eat the hours that should go to client work. The whole company, in one Team engagement.

2.5 hours, from 5.0 h

Account manager: Proposals and estimates, hours a week (illustrative)

1.5 hours, from 4.0 h

Project manager: Status reports to clients, hours a week (illustrative)

3.5 hours, from 6.0 h

Developer: QA and test writing, hours a week (illustrative)

The situation

Account managers, designers, developers and project managers work in Google Workspace, Figma and Jira. Proposals, weekly status reports, QA and time tracking take a large share of everyone's week, and AI use is ad hoc.

Scope

All 25 people in one Team engagement, with a 1:1 deep-dive and 1:1 build sessions for everyone.

Approach

Map: interviews with the partners and each discipline lead, and a check of Google Workspace, Figma and Jira settings.
Measure: an AI intake for everyone, then four weeks of baseline from Jira, time tracking and short time samples.
Build: a shared proposal and status-report skill library, personal setups, and 1:1 build sessions per person.
Compound: one champion per discipline and a monthly lift report.

Measurement

Jira and time-tracking data where they exist, short structured time samples for proposals and reports, and one agreed output measure per discipline.

What they keep

A working setup and first tools for every person, a measured result against the target, and a library the agency keeps.

Company · 250 people

A 250-person industrial manufacturer in Jutland

Fictional example

Copilot licenses are paid for, but few use them. A Pilot with customer service and inside sales, then two waves.

6.0 hours, from 9.0 h

Customer service agent: Replies in three languages, hours a week (illustrative)

2.5 hours, from 5.0 h

Inside sales: Quotes and order confirmations, hours a week (illustrative)

2.0 hours, from 5.0 h

Supply planner: Supplier follow-ups, hours a week (illustrative)

The situation

Inside sales, customer service in Danish, German and English, supply planning, QA documentation, finance and engineering run on Microsoft 365 and an ERP. Copilot was rolled out to everyone, and usage is low.

Scope

A Pilot with about 40 people in customer service and inside sales, then two waves across the rest of the company.

Approach

Map: leadership interviews, a Copilot usage review, and an ERP and data-access check with IT.
Measure: AI intake interviews in Danish, German and English, and four weeks of baseline from the ticket system and ERP.
Build: Copilot configured with the right data connections, reply and quote skills, and build clinics by role.
Compound: champions in every team, and each wave measured before the next one starts.

Measurement

Ticket and ERP timestamps first, short time samples where systems don't log the work, and Copilot usage reports alongside the lift.

What they keep

Licenses already paid for, put to measured use, with a plan for the next two waves.

Enterprise · 2,000 people

A 2,000-person insurer

Fictional example

A regulated business with employee representatives at the table. A measured cohort of 80 first, then waves of about 300.

3.0 hours, from 6.0 h

Claims handler: Case summaries and notes, hours a week (illustrative)

5.0 hours, from 8.0 h

Customer service: Drafting replies, hours a week (illustrative)

3.0 hours, from 5.0 h

Underwriter: Document review, hours a week (illustrative)

The situation

Claims handling, customer service, underwriting and IT work in Microsoft 365 Copilot and an internal model platform. Data protection, AI and IT-risk rules shape what can be configured, and employee representatives are involved from the start.

Scope

A Pilot cohort of 80 people in claims and customer service, then waves of about 300, with a network of about 80 champions.

Approach

Map: leadership and risk interviews, a review of both platforms, and the measurement plan agreed with employee representatives.
Measure: AI intake for the cohort, and four weeks of baseline from claims-system timestamps and time samples.
Build: governance and permissions with IT and risk first, then personal setups and role clinics.
Compound: champions, office hours and a monthly lift report for each wave.

Measurement

Claims-system data first, results reported only for groups of five or more, and measurement data never used for decisions about individuals.

What they keep

A documented, measured rollout the organization can repeat wave by wave.

Start with one team.

One team. We inform employees first, take baselines, set everyone up and measure again about 60 days later. Same promise, price agreed up front.

Book a scoping call