case_study · ecm_group

Case study · ECM Group

From "we're just a bit afraid of it" to a company that redesigns its processes around AI.

the company

Contract engineering — supervision on airports, museums, roads, rail and energy projects.

starting point

Not one technical person on staff. "IT" was an outside firm managing Microsoft licences.

the stakes

A single error in a bid lost a multi-million tender. Lack of data structures crippling operations.

the span

Twelve months on the map — three stages, walked in order.

futureorgdesign.comcase_study · ecm_group — 2026

section: where_it_started

where it started

A company running on Microsoft,
binders, and memory.

Low tech maturity — the tech stack was, in practice, just Microsoft.

Some of the work still lived in binders.

A lot of the company's knowledge lived nowhere in particular — scattered across people's heads, email threads, OneDrive folders and local computers.

A couple of people using ChatGPT, occasionally.

They decided to look into AI to solve problems that had been haunting them for years — and found that the solutions on the market were either nonexistent or too expensive.

Instead of buying a solution, they decided to build a capability — and learn how to leverage AI on their own. This is that journey, stage by stage, including the part where it almost got cancelled.

stage_01 · foundations

1stage 01

Foundations — the whole team, together.

The whole team was trained on the fundamentals of AI:

what it is,
what it means for the business,
what it means for each of them personally — for their careers, for the education of their kids, for where the world is going.

They needed to understand that the shift is bigger than they think.

That last part carried the stage. The first associations with "AI" in the room were robot and fear — fear of being automated away, but also fear of an uncertain future. The work of this stage was helping people realise that the future is uncertain, yes — but it also holds great opportunities. AI is not something that replaces us. It's something that lets you do the things you want to do, better — an enhancement of your own capabilities, not a substitute for them.

"My fear grew today — but a fascination was born beside it."

workshop participant, after the first foundations workshop

the honest part — what was hard here

The stage started in the summer — holidays made coordinating everyone genuinely difficult. And the team spanned multiple generations with very different starting levels of AI understanding, so everything had to be paced to bring the whole group to roughly the same level of competence — because the point was to learn collectively, as one team, not to produce a few fast individuals.

stage_02 · automation_and_the_value_trough

2stage 02

Automation — and the value trough.

The focus moved to automating the tasks that take the most time and are simplest to build — to reinforce the foundational AI skills the team had just learned. But the majority of the real work in this stage was structural: mapping the processes, understanding how the company actually operates, and figuring out how to make it better with AI.

And here is the part every company should hear before it starts: the automations, by themselves, weren't providing visible business value. Yes — a report now took half an hour instead of four hours; an offer took two hours instead of eight. But the results of the work were still the same results. Faster, not better — and with no net effect on the business.

That's because automation is not where the value of AI is. Automation buys back time — so people can raise the quality of what they produce, or go after the things that grow the company and fix what's broken inside it. That value only arrived in stage three, as people started to use AI to learn and improve — not just automate.

the honest part — what was hard here

This is where the board almost cancelled the whole project. Their reasoning was fair: people doing things faster didn't justify the investment, because it wasn't affecting business results. Frustration peaks exactly where the curve is about to bend — and this was that point.

stage_03 · the_compounding_value

3stage 03

The compounding value —
business problems start falling.

Everything built in the first two stages:

the shared competence,
the mapped processes,
the working habits,
the improvement mindset,

started paying off at once. Real business problems — some of them carried for years — began getting solved with AI, at a fraction of the cost, time and resources that would otherwise be needed. The team's mindset shifted towards building, as they recognised they now have the power to fix the things that had been bugging them for years and to improve the quality of their work.

the tender checker

Every bid, checked before it leaves

Every year the company was losing millions to lost tenders — including disqualifications over a single mistake in a submission. The team built a solution that checks every tender before it's sent and catches the errors.

millions in opportunity, protected

the board dashboard

Full visibility, for the first time

The board gained visibility into what is happening across 30 multi-million-dollar contracts — a dashboard, updated on a regular basis, built with AI. Something that simply wasn't possible for them before.

30 contracts, one view

the personnel database

A $50,000 quote, answered in-house

Two years earlier they had been quoted $50,000 for a personnel database system. The team built it themselves — using a standard Claude licence.

$50,000 not spent

the startups who came selling

"We already built that"

Startups approached with solutions to sell — one for tenders, one for legal documentation. It turned out the team had already built both themselves, inside the Claude licences they were paying for anyway.

two vendors, not needed

80% adoption

across the company

A team of power users

able to fix the company's problems and build the things that grow it

A transformation leader

redesigning operations and turning capabilities into habits

section: the_transformation_leader

the transformation leader

The leader was built too.

Throughout the journey, the internal transformation leader — Michał Męczkowski, head of legal — was building his own capabilities alongside the company's.

He set up his own Claude Code. He built his own personal OS. He learned how to map the organisation, and how to redesign its processes in an AI-native way — and how to put them together into a coherent operating system.

Today he leads the company's continued evolution: redesigning operations, and building the processes that make sure the capabilities the team built become habits.

section: what_really_changed

what really changed

The people. The team. The business.

the people

Work worth doing

They let AI execute the work they don't want to do — and focus on the work that actually moves the business forward.

the team

Capability, not tools

The team built the capability to fix problems and build new solutions — and keeps using it, on the next problem and the one after.

the business

Improving itself

The business is now able to continuously improve itself — redesigning how it operates instead of waiting for a vendor or a consultant to do it.

section: the_discovery

one thing we discovered on the way

The value of AI is not in automation. It's in the improved capability of the people you already have. Allow them to use AI and grow with it — and your business grows with them.

See the system they walked it with.

The full story — why this exists, how the method works, and what it looks like inside a real company.

Explore the System → Start the conversation