AI today makes one person faster.
We help people
think
better together.
We are Rainmaker, a Nordic learning lab for augmented intelligence. Technology is an amplifier: it grows whatever it is applied to. We apply it to people thinking together, and keep human judgment in the loop. The data and the decisions stay yours, always and by design.
01The space
We run spaces where people and AI think together.
Most AI works for one person at a time. Ours works for the whole group. It becomes part of the conversation, reflects the group back to itself, and holds on to what the group learns. So each time the group meets, thinking together comes more easily.
A space starts as many voices.
Everyone who has sat in a meeting knows what happens when strong views pull in different directions. Many voices are not yet a group that thinks. It takes trust, a shared experience, and voices that resonate before difference becomes a strength. Our methods and technology create those conditions, and our AI system hears every voice at once and highlights what really matters. When the conditions are right, highly effective teams follow, and so do organizations.
02Dreamer
Human judgment and machine intelligence, in one loop.
Dreamer is our name for that loop. It is positive-sum: each side makes the other better.
Humansdecisions
AI & Datapatterns
Humans know what matters and what is at stake, and we make the decisions.
AI and data take in more than any group can, and surface the patterns beneath it.
this is Augmented Intelligence
Each loop that closes leaves both better equipped for the next.
03How we work
We test, prototype, run it in a real space with real people, take the feedback, and go again.
Every project starts with an idea and a problem or possibility.
Every project starts with an idea and a problem or possibility.
We build a prototype: an AI agent set up with your context, and a session plan with an agenda. The first one ran in the autumn of 2022.
It runs live inside your organisation, with a real group such as a leadership team. Trust, and what each person has at stake, only show when the decisions are real.
The AI reflects back what the group said, and the group decides what to change. That feedback goes straight into the work.
Steps 1 to 5 make one learning loop. Then we go again, and each loop starts from what the last one learned. So the gains compound: the group gets going faster each time, and every insight builds on the ones before.
03Case study
“…what we think.”
For a year, one team at IKEA Sweden brought an AI into their work, as a companion in every meeting. In multiple sessions, without comparing notes, people said the same thing: it’s not a tool, it’s a colleague. Then they began to say “what we think” about conclusions the AI had helped shape. And the groups with the highest psychological safety thought best together.
The “we” was the whole group, including the AI. That is what we mean by augmented intelligence.
04Why now
The infrastructure of the intelligence age is being decided right now.
Every age is shaped by whoever owns its infrastructure: steel and rail in the industrial age, the internet backbone in the information age. The intelligence age is being decided right now, and the default is that most of us rent intelligence from a few large companies—the AI labs.
We offer an alternative, where you stay in control. Here are six things happening right now, and Rainmaker’s take:
-
3 months
Machines now work for hours alone.
METR finds that the length of a task an AI agent can finish doubles every three months.
What we offerWhat the machines work on is still a human decision. We help your people make it together.
-
9–900×
Answers are getting cheap.
Epoch AI measures the price of a fixed AI capability falling 9 to 900 times a year. METR finds agents far weaker at strategic judgment than at technical work.
What we offerWhen anyone can buy answers, your people’s judgment is what sets you apart. We help them use it together.
-
$800B
The labs are buying all the hardware.
Hyperscalers will spend about $800 billion on AI in 2026, analysts estimate. Nvidia doubled its revenue in a year.
What we offerAccess follows ownership. So we are building our own machines in Sweden, and what you build with us stays yours.
-
−20%
The default aims to remove people.
Amazon cut 16,000 corporate roles in January 2026 and tied the cuts to agents. Stanford finds employment of developers aged 22 to 25 down nearly 20% since 2024. Nvidia’s Jensen Huang says leaders who cut people for AI are “out of imagination”.
What we offerWe build the other way round: AI that helps your people take on more, and grow.
-
4 months
Open models are only four months behind.
Epoch AI finds the best open-weight models trail the closed frontier by about four months, and the gap is closing. A frontier-class open model fits in 17 GB on a laptop.
What we offerYou can own the AI you use. The conversation can stay inside the building.
-
€30B
Sovereignty is now policy.
Sweden has its first national AI strategy, and it funds sovereign Swedish language models. The EU has called for AI gigafactories, about €30 billion in all.
What we offerSovereignty by contract is sovereignty you can lose without being told. We build it into the infrastructure instead.
We build for groups thinking together, not individuals querying alone. Your team’s conversations become a memory that grows for years, on infrastructure you understand and control.
05Our story
People matter more in the age of intelligence.
We have worked from that idea since 2022. More than 30 pilots later, it still holds.
The record goes on below, moment by moment.
Since 2022The white paper came out in September 2022. Its argument: technology should grow our wisdom as fast as it grows our power. The first prototype followed that autumn.
2022At least 30 pilots by October 2026. Three cases written up: a Nordic standards organisation, IKEA Sweden and Yggdrasil.
30+→ 15–20Minutes to think together
At IKEA Sweden, the time the team needed to start thinking together fell from 35–40 minutes in the first session to 15–20 by the fourth. The pattern underneath: the higher the team’s psychological safety, the stronger its collective intelligence. More than twenty breakthrough moments were documented.
35–40 → 15–20Three cohorts have completed our AI whisperer programme. Each practitioner builds agents for their own work and runs it live with their own customers. A partnership starts with the training, not with a contract.
32022 to now. Drag or scroll sideways, and open any moment.
06What grows
Your people and the AI learn together, and what you learn together compounds.
Teams that think better together.
A team trained in Dreamer runs its own sessions, and each one gets going faster than the last.
The implicit, made visible.
The AI surfaces patterns people sense but seldom say, in one team and across the organisation.
A memory the organisation owns.
From the first session, the AI records what the group says and decides, and why. That record grows into a memory your organisation owns.
Every loop adds to the memory, sharpens the method and improves the platform, so the next loop starts from more. All of it runs on one stack, from your data up to the wisdom that informs it.
8Wisdom
Human wisdom—the ability to make long-term and mutually beneficial choices—informs the data, and the flow rises again. As machines take on more of the detailed work, we believe human skills matter more than ever: judgment, trust, and a sense of what truly matters.
7Partner · practice
A partner is a practitioner with their own customers, trained in our methodology and using our systems. They take the platform and the process to those customers, build the contextually aware agents each one needs, and run sessions themselves.
Garden · new ventures
Where new ventures grow, on everything below: ideas get access to the method, the platform and the people needed in order to thrive and grow.
6Applications built on top
Full applications for professional work, assembled on Deep from a library of ready and tested components, like Lego bricks, rather than written from scratch.
5Deep · the meta application layer
You describe what you need, and Deep assembles it from parts already in the stack, or creates them on the fly. Every part it creates becomes a component the next application can use, so the library grows with each one.
4Prism · infrastructure
Open models and secure systems in Stockholm. Your data stays where you are in control, and the models are run on sovereign hardware.
3Hardware · our own, in Stockholm
Our own machines in Stockholm, Sweden. Where the data physically sits decides who can reach it, so we own this layer rather than rent it. Owned machines stay under Swedish law, where we can point to them, run by people we can name. Rented capacity can change its terms without asking.
2Ontology
Rainmaker’s ontology, sourced by River. It guides how the system behaves, toward augmented intelligence. Most stacks skip this layer, yet it is the one that makes the most difference, and it stays when the models are changed or updated.
1Data · owned by you
You always own your data. It is the foundation the rest of the stack stands on.
07Partnership
We work through partners.
A partner is a professional practitioner, team or organization.
Trained in our methodology and philosophy, partners bring the platform and the process to their customers, and build and run AI in their contexts. We provide the platform, the training and the support.
Garden is our innovation hub.
It is where new ventures are seeded. An idea that starts in a session, with a partner or in our own team gets what it needs to take root: the method, the platform, and the right people around it. The ventures that grow become businesses of their own, and what each one builds, the others can build on. Garden is being built now.
08People
One team, in Sweden.
The people who run the spaces and the people who build the platform work as one team, from Sweden.
We are Rainmaker,
a Nordic learning lab for
augmented intelligence.
We are sourced by River, a commons we helped start but don’t own.