Consulting & enablement

One week in your business. Your team running AI by Friday.

Your people have already tried AI. A few are quietly using it for real work, and nobody is quite sure whether that's allowed, whether it's safe, or whether it's any good. Meanwhile the work that actually eats the week — the quoting, the scheduling, the reporting, the inbox — looks exactly like it did two years ago. We come in for a week, find where AI genuinely fits, teach your team to use it properly, and build your first agents alongside them.

Talk to us about a week See how the week runs

On-site, Monday to Friday · workshops in groups of 10–15 · you keep everything we build.

“They made big promises, but now I can't imagine working without the agent Claudia and Anton built for me.” Justina — The Book Photographer, Amsterdam

~8 hrs/weekof admin taken off one founder's plate by the first agent we built for her.

85,000+works in the searchable library we built and handed over to author Celia Hammer.

The situation

You don't have an AI problem. You have a "where does it actually fit" problem.

Almost every company we speak to is in one of these four places at once. If more than two sound familiar, a week is probably worth your time.

Every proposal arrives as a licence to buy

Vendors want to sell you seats. None of them start by asking what your people actually do all day — because that isn't what they're selling.

Shadow AI is already happening

Some of your team are using AI daily and not mentioning it. That's not a discipline problem — it's an unmanaged one. Nobody has told them what's safe to paste in.

The strategy is a slide, not a workflow

There's a deck about AI. There's possibly a committee. What there isn't is one person whose Tuesday got measurably easier.

Someone ran a training day and nothing changed

A generic session on prompting, everyone nodded, and on Monday they went back to exactly what they were doing before. Training that isn't tied to real work doesn't survive contact with a normal week.

The expensive part isn't the tooling — it's the year you spend deciding. Most AI budget gets burned on licences nobody adopts and pilots nobody finishes. The point of the week is to end the deciding: you finish it knowing what's worth doing, what isn't, and who in your building can do it.

What you keep

What's different the Monday after.

The test of a week like this isn't how it felt on Friday. It's what's still true a month later.

A ranked map you can act on

Where AI pays in your business, in plain language, ordered by effort against payoff — including the places we think it isn't worth it.

A team you don't have to worry about

People who know what good output looks like, what to check, and what never leaves the building.

A safe-use standard in writing

Short, specific to your business, and written so people actually follow it rather than route around it.

Named champions, not a committee

The two or three people who will carry this after we leave — identified from how they actually work, not from job titles.

Working agents you own

On your accounts, with your data, handed over completely. Nothing locked behind us.

The ability to build the next one

The whole point. You shouldn't need to call a consultancy every time you want to automate something small.

How it works

The week, hour by hour.

Five days on-site. Management on Monday, your people mid-week, workshops at the end — then, if it's useful, we come back and build.

Day 1 Monday With leadership

What the business needs, and what it will actually allow

We sit down with management and get the honest version: where the business is going, where time and money leak, what's already been tried, and what the real constraints are — budget, compliance, the tools you already pay for, and the things that are genuinely not up for discussion. We also agree what "worth doing" means to you, so we're ranking against your definition rather than ours.

OutputA shortlist of candidate workflows, and the constraints everything else gets tested against.

Days 2–3 Tuesday & Wednesday With the people doing the work

The workflow on the org chart is never the real workflow

We talk to your team one-to-one and in small groups about what they actually do — not the documented process, the real one, including the spreadsheet they maintain on the side and the report nobody reads. This is where the useful findings come from. We're looking for two things: which work is genuinely suited to AI, and which people have the most to gain and the appetite to run with it. Those people become your internal champions, and this is how you find them.

OutputA ranked map of the real work, effort against payoff — plus the names of the people to build around.

Days 4–5 Thursday & Friday Workshops · 10–15 people per group

Not a tool demo. Your team, working on their own tasks.

Two days of hands-on workshops, run in groups small enough that everyone actually participates. We cover how to get good output rather than plausible-sounding output, how to spot the moments AI is confidently wrong, what should never be pasted into a public model and why, and where the genuine limits are. Everyone works on their own real tasks — the ones we found on Tuesday and Wednesday — so nobody leaves with notes they'll never open again. Larger teams run in consecutive groups.

OutputA team that can use AI on their own work, and a written safe-use standard that fits how you operate.

Optional · week 2 Building With your champions

We build your first agents with your people, not for them

If the map points at something worth automating, we come back and build it — sitting next to the people who'll own it rather than disappearing and returning with a finished box. Two things come out of that: the first agent works properly, because we made sure it does, and the second one gets built without us. This is the part most consultancies skip, and it's the reason most AI projects don't survive the handover.

OutputWorking agents, running on your accounts, owned by you — and someone in-house who knows how they were made.

Who runs the workshops

The room is the whole job.

Anton runs them — and he's also the person who builds the agents afterwards. That matters more than it sounds. The moment that decides whether a room takes any of this seriously is when someone asks a question the slides don't cover.

Your work, not a curriculum

Everyone spends the two days on the tasks we found on Tuesday and Wednesday. Nobody sits through a demo of somebody else's job.

Nobody has to become technical

The person at the front is an engineer. The room isn't, and doesn't need to be — it's run in the language your team already uses about their own work.

The sceptic is the useful one

There's usually someone who thinks this is all nonsense. They ask what everyone else is too polite to, and they get a straight answer rather than being talked over.

Anton Deken — engineer, has used AI in his own work daily since 2022, and builds what you end up owning. Judged on presenting by international panels at three international student engineering competitions, scoring 224 out of 225.

Why our advice is worth having

We don't sell licences. We have nothing to push.

This matters more than it sounds. Most companies offering AI advice are resellers, implementation partners, or tied to a platform — and their recommendation was decided before they met you.

Genuinely tool-agnostic

We are not affiliated with Anthropic, OpenAI, Microsoft or anyone else. No reseller agreement, no commission, no partner tier to hit. If you've already committed to a platform we'll make it earn its keep; if you haven't, we help you pick on the merits — and sometimes the honest answer is that the tool you're already paying for is fine.

Including when nothing may leave the country

If you're in a position where data cannot go outside the Netherlands or the EU, that rules out most of the obvious options — but not AI itself. Through a strategic partnership with an AI company specialising in high-security environments, we can put local, GDPR-compliant capability in front of your team without the data ever leaving your jurisdiction.

Honest fit

When this works — and when it doesn't.

We'd rather tell you now than take a week of your time and your budget to arrive at the same conclusion.

This is built for you if

You have enough people that their work genuinely differs — roughly 10 to 50 is the sweet spot.
Someone in the room can actually change how work gets done — not just recommend it.
You'd rather your own people ended up capable than keep a consultancy on retainer.
You have real administrative weight — quoting, scheduling, reporting, client correspondence, compliance paperwork.
You want a straight answer about where AI won't help, as well as where it will.

Probably not, if

What you want is a platform recommendation and a deck. Plenty of firms will sell you that, and more cheaply than we would.
Nobody is permitted to change a process without a long approval cycle. The week will produce good findings that then sit still.
You're looking to reduce headcount. That's not what we do, and your team will work out the real agenda by Tuesday afternoon regardless.
You need one specific agent built and nothing else. That's a smaller, cheaper job — just ask us for that instead.

Before you ask

The questions we get most.

Our data can't leave the Netherlands. Is this even possible for us?

Yes, and it's a more common position than people assume — healthcare, legal, public sector, finance. It rules out most consumer AI tools, not AI. Through our partnership with a company specialising in high-security environments, we can run local, GDPR-compliant models where the data never leaves your jurisdiction. We'd establish exactly what your constraints are on the Monday, before anything gets recommended.

We already pay for Copilot / ChatGPT for the team. Is this redundant?

The opposite — it's usually the strongest starting position, and one of the most common things we find is a company paying for seats that maybe a third of the team meaningfully uses. You've removed the procurement question already. The week is then about making what you own actually earn its keep, and we'd have no reason to move you off it.

Half our people are already using AI without telling anyone. Is that a problem?

It's the single most common thing we find, and it's better acknowledged than policed. Those people are usually your fastest adopters — they've just been doing it without guidance on what's safe. The workshops are where that becomes visible, supported, and safe, rather than something people hide.

How do we know it won't fade out after you leave?

It's the main risk, and it's why the week is shaped the way it is. Training that isn't tied to someone's real Tuesday doesn't stick, which is what Tuesday and Wednesday are for. And an agent built for people gets abandoned, whereas one built with them gets maintained — so we build alongside your champions rather than handing over a finished box. We're honest that we can't guarantee adoption: what we can do is remove every excuse for it not happening.

The timing isn't great. Can we look at this next quarter?

Sometimes that's the right answer, and we'd rather you came back in three months than booked a week you're too stretched to use. The question worth asking honestly is whether the quarter after this one will actually be quieter. If it won't, then the work you'd be fitting this around is the work we'd be trying to take off you — and every month it stays manual costs the same hours it costs now. Either way, tell us and we'll check back in when it suits you rather than chase you.

I'm not the only person who has to sign this off.

Then bring them to the first call — it's half an hour, and it saves you relaying our answers second-hand. The questions that actually decide this (what it costs, what changes, what it means for people's jobs) land far better asked directly than passed along. If that isn't practical, tell us what they'll want to know and we'll put it in writing in a form you can forward.

What does it cost?

It depends on the size of your team and whether you want the build week, so we quote it properly after a short conversation rather than putting a number here that's wrong for you. What we can tell you: it's a fixed scope agreed in advance, and there are no licences, per-seat fees, or ongoing costs owed to us afterwards — anything you run, you run on your own accounts.

Do you have to be on-site?

For the middle of the week, strongly preferred. The Tuesday and Wednesday conversations are where the real workflows surface, and people are markedly more candid about how work actually gets done in person than on a call. Monday and the build week can be remote if that's easier.

Get in touch

Start with a conversation, not a proposal.

Tell us roughly how big your team is and where you think the time goes. We'll tell you honestly whether a week would be worth it — and if it wouldn't, what we'd do instead.