Services

AI teams that actually do the work.

We build agents that take the repetitive work off your calendar. Fixed price, live in under a month.

What we do

  • 01

    AI automation and agents

    We set up agents that run in the background, talk to your tools (Tripletex, Fiken, Shopify, Slack), and do the work without anyone lifting a finger.

  • 02

    AI strategy and advisory

    A couple of workshops, a prioritised list of what is actually worth automating, and a plan you can work from. No slide decks.

  • 03

    AI implementation

    We build and ship one concrete workflow end to end. Fixed price, fixed scope, and a runbook you own after handover.

How we work

Four steps, no surprises.

  1. 01

    Discovery

    We walk through your processes, find what is grinding, and score every candidate on hours saved per month.

  2. 02

    Strategy

    We pick the one workflow that pays back the most first, and write a clear scope with acceptance criteria.

  3. 03

    Build

    We build the agent in your environment, connect it to the tools you already use, and test against real data.

  4. 04

    Follow-up

    Two weeks of close support after launch. After that you can take it over yourself, or we can run it on retainer.

What clients walk away with

Numbers from the first pilots.

  • Time saved

    23h

    per week

    Average across three pilots, measured against baseline before the agent ran.

  • Volume

    3x

    more cases handled

    Same team, three times the inbound resolved within the same day.

  • Quality

    62%

    fewer errors

    Fewer typos, fewer forgotten tasks, fewer chase emails from clients.

  • Time to live

    18 days

    from kickoff to live

    Median for a pilot. We have done it in 11, we have spent 25, it scales with scope.

Pricing

Pick the smallest door.

Three tiers. No open retainers. No surprise invoices.

Questions we get

What people ask us.

  • Is this safe? AI hallucinates.

    Every agent we ship has bounded scope, approval gates on anything outbound (emails, invoices, payments), and full audit logs. You sign off on the policy before the agent goes live.

  • Our data is messy.

    That is what the Discovery Sprint is for. We scope what is doable now versus what needs cleanup first, and we price both.

  • We are waiting for ChatGPT Enterprise or Copilot.

    Those are chat assistants, they answer when someone opens them. We install workflows: agents that run in the background as part of operations.

  • Who owns the agent after launch?

    You do. It runs on your account, against your tools, on your data. If we walk away tomorrow, it keeps running.

  • Who actually builds this?

    A small Norwegian team in Bergen, supervised by an accountable principal. We take a limited number of engagements per quarter.

  • Does the agent speak Norwegian?

    Yes. We explicitly test on Norwegian documents, emails, and customer data, not just English.

  • Which industries fit best?

    Most of our delivery has been in accounting, law, and e-commerce. The common thread: many small repetitive tasks with clear rules.

  • What is the first step?

    A 30-minute call. We ask what is grinding, you tell us, and we say honestly whether an agent is the right shape for the problem.

Ready to put agents on the work?

30 minutes on a call. We scope the problem and tell you honestly whether this is the right path.