Skip to main content Skip to contact

AI advisor for founders and operators.

I help founders and operating teams find where AI pays back, then help them ship it.

Not ready for a call? Estimate what AI could save your team

Portrait of Stefan Nilsson

Trusted by

  • Bohus Betong
  • Klottrets Fiende
  • Sweplan
  • Ljungheden
  • Norconsult
120+
Systems shipped
12,000+
Newsletter readers
3,500+
Community members

Results

Results you can put a number on.

Recent work, measured in hours back and money recovered.

Construction
28 h
a week saved on quantity takeoff
$500K
a year in added revenue

Cut quantity takeoff time by 28 hours a week and grew revenue by $500K a year.

What we built: AI-assisted quantity takeoff for the estimating team.

Field services
~50 h
a week saved across the organization
~$25K
a month recovered from errors and missed work

Saved ~50 hours a week across the organization and recovered roughly $25K a month previously lost to errors and missed work.

What we built: Automated workflows across the organization's day-to-day admin.

Construction
35 h
a week of admin cut
20 h
a week turned into billable work
$200K
a year in change orders captured

Cut admin time by 35 hours a week, turned 20 hours of previously unbilled work into billable hours, and captured $200K a year in change orders that used to fall through the cracks.

What we built: Admin automation and change-order tracking across projects.

Engineering
~1 h
saved per person, per week, so far
10
people in the Copilot rollout

Early-stage Copilot rollout across a 10-person team already saving ~1 hour per person per week, with most of the runway still ahead.

What we built: A Copilot rollout and hands on training for the team.

Services

How I work with teams.

01

Strategy and sparring

Second opinions, advisory calls, and AI roadmap reviews. For founders and leadership teams who want a thinking partner who has shipped this work before, not a deck builder.

02

Implementation guidance

Architecture decisions, vendor selection, build versus buy calls, prompt and agent design. Hands on enough to write the code when that's what the work needs.

03

Team enablement

Workshops and training that get teams fluent in Copilot, Claude, and modern AI tooling. Focused on what actually changes how people work, not generic literacy slides.

Tools

I build with the tools you already use.

AI goes into the systems your team already opens every day.

  • Claude
  • Copilot
  • n8n
  • Zapier
  • Make
  • HubSpot
  • OpenAI
  • Lovable
  • Gemini
  • AWS

AI Savings Calculator

What is manual work costing your team?

Three numbers, about a minute. You get a simple estimate of the hours and money AI could give back to your team each year.

Team
12 people
Manual work
6 h/week
Hourly cost
$60
Estimated annual value
$59,600
Example estimate. Try your own numbers.

Fit

I take a small number of engagements at a time.

Which means being honest about who I'm useful to.

Right fit
  • Founders and operating teams who already have something shipping and want to put AI inside it.
  • Leaders making a build versus buy or vendor selection call this quarter.
  • Teams that have piloted AI and need help turning experiments into something durable.
Probably not
  • Teams looking for someone to write the deck or rubber stamp a strategy that's already decided.
  • Companies expecting hands off enterprise consulting with a pyramid of juniors behind it.
  • Anyone who wants a vendor to manage, not a partner to think with.

How I think

Principles I work by.

01

Pick the problem worth solving.

Most AI projects fail because they pick the wrong one. Choosing right is half the work, and the half nobody puts on the roadmap.

02

Boring beats clever.

The basic solution that ships beats the clever one that doesn't. Most teams overestimate what AI can do this quarter and underestimate what it'll do in two years.

03

Ship narrow, then widen.

Demos lie. Users don't. Get something narrow and useful in front of them before chasing something broad and impressive.

About

Builder first, advisor second.

I've spent the last few years architecting and shipping more than 120 AI systems, for companies from early stage startups to public sector. I write Sweden's largest AI newsletter and run a 3,500 member community where operators learn to use AI seriously. I'd rather build something that works than write a thought leadership post about it.

Stefan Nilsson
AI advisor, Stefcode
“Stefcode helped us with the architecture and the results are amazing. We ship things very quick and keep continuing to do so.”
Jacob Wiksell
CEO, Flera AI

FAQ

Questions, answered.

Something else on your mind? Ask me directly.

What exactly does Stefcode do?

I help companies find where AI actually pays back, then build it with them. In practice that means three things: strategy and sparring for leadership teams, hands on implementation inside the systems you already use, and training that gets your team fluent in tools like Copilot and Claude. You work directly with me, not an account manager.

How is this different from a typical AI agency?

I start with the problem, not the technology. Most of the first conversation is about where your hours and money actually go. The person who scopes the work is the person who builds it, so nothing gets lost in a hand-off to juniors. And I measure success in hours saved and revenue captured, which is why every case on this site comes with a number.

Do we need to already be using AI in some form?

No. Plenty of teams start from zero. What matters more is having a repetitive process that eats hours every week, like reporting, admin, or quantity takeoff. That's usually where the first win is.

What does an engagement cost?

It depends on the scope, so there's no price list. A focused first project costs a fraction of a larger build, and you'll know the scope and the cost before any work starts. The intro call is where we figure out whether the numbers work, before you commit to anything. For a rough sense of the upside, try the savings calculator.

Can we start small?

Yes, and I'd recommend it. A good first project is one narrow workflow: one process, one team, one number to move. Once that's working and measured, we widen it.

Do you only work with construction companies?

No. Construction is where some of my clearest results are, but I work with founders and operating teams across industries, from early stage startups to engineering firms and the public sector. The same approach works anywhere a team loses hours to repetitive admin, reporting, or manual data work.

Are you based in Sweden or can you work internationally?

Both. I'm based in Sweden and work with teams internationally, in English or Swedish. Most of the work happens remotely, and on-site days can be arranged when they're worth it.

Work with me

Not every team needs me.

If yours does, you'll know on the first call. One short conversation is usually enough.