One process.
Sixty minutes.
Then you know.
Sixty minutes of conversation, becoming a process map.
The shared screen from a live session. Every step, system and handoff appears as the people who run the process describe it, using plain language, their words, structured as they speak.
Your team, interviewed live. Your process, mapped in real time — how it really works, not how the org chart says it works. A clear verdict on what AI agents can take on, and what they can't. You decide what happens next.
Illustrative example — we don't publish client sessions.
Pick a time and you come straight back here to finish the details.
What an hour like this produces.
Before anyone signs anything.
That's what Round Table is for.
Everyone in the room sees the same thing. The people who describe the process receive the same record as the people who commissioned it. Same map, same findings, same numbers. There is no second version for upstairs.
Describe your process
The one eating the most time, or hiding the most risk. Tell us in advance so we can prep the interview.
We run the session
Sixty minutes. Our AI agents interview the people who actually run it, live, while a shared screen maps every step, system and handoff as they talk.
You get a verdict
What agents can genuinely take on, what they can't, and the hours sitting behind each one — every unresolved question named rather than glossed over. The Autonomy Map is yours to keep either way.
Then we build, without touching your systems.
Round Table maps the process. Construct is where NorseAI builds the agents that take it on, and governs them for as long as they run, wherever you choose to run them.
- 01
Mapped with your people
Touches your people, not your systems
One hour with the people who run the process. What they describe becomes the Autonomy Map the build starts from.
- 02
Built and tested in Construct
Touches nothing of yours
Agents are built and tested against your Autonomy Map, using synthetic or anonymised data. Live data never enters, and Construct never connects to your systems.
- 03
Deployed by your team, where you choose
Touches your systems
Your people take the agents into your test environment, then into production — hosted in Construct, or in your own cloud or datacentre. Running agents report to Construct, but Construct has no way to act on them: the controls stay with you.
Already a client? Sign in to Construct
It runs inside your boundary, not ours.
Runs where your data already lives
On-premise inside your own infrastructure, or in your own EU cloud tenancy. Air-gapped deployment supported. The deployment boundary is yours, and nothing crosses it without your sign-off.
Processed in the EU, under a signed DPA
Your data is processed in the EU under a signed DPA with named subprocessors. On-prem and air-gapped tiers go further: open-weight models running on your own hardware, inside that same boundary — the flagship delivery path for clients who need it, not a claim about how every session runs.
Engineered in-house, agent architecture you can inspect
NorseAI AB designs, builds, and operates the full agent stack — no outside partner in the room or in the pipeline. The architecture follows published agent-design patterns, and every claim made in the room is traceable to a line in the session transcript.
Contracted before anything is processed
A GDPR Art. 28 DPA is signed before a single record is touched. Agent participation is disclosed by design, per EU AI Act Art. 50. Defined retention, defined deletion, named subprocessors.
AI was never the hard part.
The process was. Most of our clients have theirs documented — mapped, ISO-certified, signed off. But the documented process and the running process are never quite the same thing. Between the official steps sit the spreadsheets, the email threads, the shared folders and the manual workarounds that never make it into the description.
That gap is where agent projects fail — the agent automates the documented path, then meets the real one in production. Round Table captures what actually runs, from the people who run it. It is why the hour is spent interviewing your operators rather than demoing our software, and why most of the risk comes out of an agent project before the build starts.
And we build and run the agents ourselves, so the verdict comes from the team that would have to deliver it.
Not a generic AI demo built for any business. Built for yours.
Transcript
NorseAI builds AI agents for enterprise operations. But agents are an investment — and you want to know it's a good one before you commit. That's what Round Table is for. One process. Sixty minutes. With the people who actually run it. Then you know.
One fixed price to find out. Nothing after that unless you decide.
The hour is priced on its own. Everything past it is scoped from what the session measured, so you are never quoted against a guess.
Then, in Construct.
Construct is our platform for what comes next. It is where we build your agents, and how they are governed for as long as they run. You decide at each step whether to go on.
We build the agents your Autonomy Map points to and test them on synthetic or anonymised data, never on your live systems. It starts with one to three weeks of detailed mapping: the volumes, touch time and exception rates that set the Baseline you hold us to, and the scope and price of the build.
Your agents run where you choose: hosted in Construct, or in your own cloud or datacentre. They report to Construct, which has no way to act on them, so the controls stay with you. Priced per completed process cycle, not per agent or seat, with a base fee for the SLA and monitoring.
Ready when you are.
Book at norse-ai.com/book