Somebody at your company is the integration layer.
They copy figures out of one system and type them into another, and they read the same request in six languages and route it by hand. We find that work, cost it in hours, and automate whatever earns its build.
We scope before we sell. You get a written list: what we would automate, what we would leave alone, and what each is worth in hours. Most firms never write the second list.
What one process costs you
Salary plus employer cost. The salary alone will understate it.
- Hours on it, per year
- 1 104 h
- Cost of doing it by hand
- kr 529 920
- Worth of the hours we would expect to give back, per year
- kr 317 952
Counts a 46 week year, and assumes we remove 60% of the hours. Sixty is a deliberately unexciting default: the real figure comes out of the diagnostic, and for some processes it is zero. Nothing you type here is sent anywhere.
The four places the hours actually are.
The same four shapes turn up in almost every operation we open up. Each one looks like a job for a person because the input is messy, and each one stopped needing a person the moment machines could read a sentence.
Documents somebody has to read
Invoices, delivery notes, claims, contracts, customs paperwork. We read them, check them against what you already hold, and send only genuine mismatches to a person.
Inboxes that need judgement
Shared mailboxes where every message has to be read before it can go anywhere. We sort them on what they actually say, and draft the reply where the answer is already known.
The gap between two systems
Where a human is the integration, copying and reconciling until the ERP and the CRM agree. The transfer runs on a schedule and disagreements come back as a short list.
Questions your documents answer
Policy, specification, price lists, what was agreed in a contract. We answer from your own material and attach the source, so the asker can check it.
A supplier invoice, arrival to ledger. Illustrative.
-
Invoice arrives
-
Read and extract
-
Matched against the order
Stalls here -
Approver notified
-
Posted
Most of these projects quietly return nothing.
Almost none of the reasons are about the model. They are about baselines, data, compliance and cost, which is where we have spent our careers.
95% of enterprise generative AI pilots produced no measurable impact on profit and loss. MIT Project NANDA (opens in a new tab)
We count first
A pilot with no agreed before-number gets judged on impressions. We count the hours and you sign off on the count, so the comparison at the end is arithmetic.
We check the data in week one
Inconsistent documents and missing access are what make this expensive, and they surface later than you would like. We look while it is still cheap to walk away.
Compliance is a design decision
The EU AI Act's transparency duties and full penalty regime apply from 2 August 2026. We have run SOC 2 programmes, GDPR work and audit readiness, so this is not a phase we bolt on.
Cost has a ceiling you approve
Token spend and cloud spend both drift upward without anyone deciding they should. We instrument cost from the first deployment and set a ceiling you approve before anything runs.
What we have already done.
Public sector, at scale
Technical leadership of Datafangst and Vegkart, the national GIS platforms at the Norwegian Public Roads Administration. Both were years into service and carrying real load. We found what was throttling them and rebuilt it from the database up.
Compliance
SOC 2 programmes, GDPR, security review and audit readiness, run as engineering work rather than as paperwork assembled afterwards.
Cost
A production cloud bill reduced by 99.1% through FinOps analysis, database restructuring and service tuning.
Engineering
Go, TypeScript, Java and Spring Boot on the backend, React and Next.js in front, and Azure, AWS, Google Cloud, Kubernetes and Terraform underneath.
Four steps, and you can stop after any of them.
Each step is priced on its own and ends with something you keep. Nobody should have to sign a transformation programme to find out whether this works.
A call
Forty-five minutes, no charge
You describe an ordinary week. You leave with a straight answer on whether anything here is worth automating.
The diagnostic
Roughly two weeks, fixed fee
A written list of what to automate and what to leave, the hours behind every line, and a ranked order. Yours either way.
The first build
Fixed price, live in 90 to 120 days
One process in production, measured against the baseline you agreed, plus the runbook, the documentation and the credentials.
Running it
Monthly, cancel any month
Someone watches output quality, spend and the edge cases, and builds the next process on the list when you want it.
Every fee is fixed and quoted after the call, against a scope you have read first. We do not bill by the hour, because the hours are the thing you are paying us to remove. The longer version of all of this covers what we would refuse to automate and why.
Non-profits do not pay us.
We build and ship websites for non-profit organisations at no charge. Animal shelters, community groups, arts organisations, sports clubs, whatever you run. It is the same work we would bill a company for, on a domain and hosting you own, and it stays yours to edit after we have gone.
Designed and built
A site that works on the phone most of your visitors are actually holding, and that loads on a bad connection.
Usable by everyone
Readable contrast, real headings, keyboard navigation, and text a screen reader can follow. Whoever you exist for, some of them are on an old phone or a slow connection.
Live on your own domain
Set up in accounts you control, so nothing about your website depends on us still being around.
Yours afterwards
You can change the words yourself. No monthly fee, no licence, and nothing to cancel later.
Tell us what your organisation does and what your current site stops you doing. A paragraph is plenty. Working out what to build is our job, not yours.
We would rather do this properly than often, so we cannot run many at once. If we are full we will tell you straight away instead of leaving you waiting on an answer.
Subject line: non-profit
Tell us which process annoys you most.
One sentence is enough. You will get a straight answer on whether it is worth automating, what it is costing you now, and what we would do first.
Forty-five minutes, no charge, and no deck. We will ask you to walk us through an ordinary week, which is the only briefing that has ever been useful. There is nothing to prepare.