Your data isn't ready for AI.
Your numbers live in several systems, and nobody has written down what they mean. An agent that reads them has to guess.
Agentic consultancy for finance, accounting and procurement
AI is ready.
Your systems aren't.
we deliver with the speed of AI
We make sure AI can access your systems and data safely, so your AI agents can start doing real work in your company now. Experienced domain experts lead the work, and Gathena's agents do most of the building work. You pay for the result, not for the hours.
THE REAL BLOCKER
You're not behind on ideas. You're blocked on access.
Companies that want AI agents usually get stuck in the same three places.
AI can't reach your systems.
Your ERP, your bank and your spreadsheets were not built to let an agent in, and nobody wants to hand it every password.
Nobody has written down how the work is done.
So an agent cannot do a task again the same way, and nothing can be automated.
In a 2026 survey, more than 60% of finance leaders named data governance, security and ERP integration as their main blockers to AI, far ahead of cost. Source: Savant Labs, 2026 Finance AI Automation report.
THE AI FOUNDATION
What is an AI foundation?
Everything your agents need to know about your company: access to your data sources, your business language, what your numbers mean, and how you work and get tasks done, written down so AI can use it. Getting your data in order shows your agents where to look and which numbers to trust, so they do better work.
With the right AI foundation, your agents:
- Have access to the data they need.
- Understand your business: which data to use, and what each number means.
- Write down how they did each task, and learn from your corrections.
- Run the work that repeats on its own, on a schedule or when something arrives.
- Show where every number comes from.
THE AI INFRASTRUCTURE
What is AI infrastructure?
It is the setup that lets your agents do real work in your systems, and do it safely. Their tools and data stay in your own cloud, they work inside limits they cannot cross, and your team approves the work that needs a decision.
When your infrastructure is in place, your agents:
- Work toward a goal you set, and bring the work that needs a decision to your team for approval.
- Stay inside the limits you set. A control outside the agent enforces them.
- Use tools and data that stay in your own cloud, with access your own people control.
- Share one AI foundation, so the next agent is set up like the first.
- Stop and raise an alert when something breaks, so a person sees it.
- Use the model that fits the job, and switch when a better one comes along.
YOUR AI AGENTS
What is an AI agent?
It does a piece of your company's work with your team, and writes down how it did it: where it found the information, what it decided and what it did. That record lets it do the job again, and shows when a job is ready to be automated.
Your agents start with the work as it comes: a question, a check, a draft. As they work, they write it down. A task that repeats and is well defined can then run on its own, and it is monitored.
Every agent has three things: your language, access to what its work needs and no more, and your team, who check its work and approve what needs a decision.
Three examples, one from each stage
- BudgetingYou ask, the agent helpsYou ask, "What if the new hire starts in March?" Your finance agent shows the effect on the year.
- AccountingThe agent prepares, you approveA supplier invoice arrives. Your accounting agent matches and checks it and proposes the posting. Your bookkeeper approves it.
- ReportingIt runs on its own, and you check the resultsEvery month end, your reporting agent drafts the report on its own, with a source for every number. You check the result.
Money leaving the company, anything going outside it, a commitment and an unclear rule always go to a person.
YOUR STACK
The right stack for your needs, built on what you already run.
We build on the systems and AI tools you have, and we advise on what fits your needs and your values. You own everything we build, and nothing moves onto a platform of ours.
Built on what you run
Microsoft, Google Cloud, OpenAI, Anthropic or your own ERP and accounting systems. We start from what your company already has and knows.
In the EU, where you want it
If you want your data and tools in the EU, we show which parts can run with EU providers or in EU data regions without lowering the quality of the solution.
Lower running costs
Each stack is strong at different things. We place each part where it works best and combine stacks where that keeps your running costs down.
Where each part runsEUUS
- Your dataStays in your own cloud, in an EU region.
- Pipelines and warehouseRun in the same EU region as your data.
- Your agents' toolsSit next to your data, in your own cloud.
- LogbookWhere the record lives is your decision.
- AI modelRuns at the provider you choose. An EU option where the quality holds, a US model where it is clearly better.
A NEW KIND OF CONSULTANCY
Domain experts lead the work. Gathena's agents do most of it.
Consultancies have always sold hours. When our AI agents do most of the work, hours stop being a fair way to price it.
| — | A traditional consultancy | Gathena |
|---|---|---|
| Who does most of the work | Junior consultants | Gathena's agents |
| Who leads the work | Senior partners, part of the time | Domain experts in finance, accounting and procurement, all the way through |
| What you pay for | The hours spent | The result agreed before work starts |
| What you keep afterwards | Reports and slides | The foundation, your own AI agents, and a written record of how they do the work |
The domain experts make the decisions, our agents do the work, and a person checks every result before you see it. Projects that used to take months are now sometimes a matter of hours, days or weeks.
WHO LEADS THE WORK
Lasse Feldt-Rasmussen, founder
MSc in Supply Chain Management and BSc in Business Administration from Copenhagen Business School. Has advised companies since 2006: as a managing consultant at ABC Softwork (now Inact), leading more than 50 inventory and working-capital projects; in a working-capital team at EY; and as chair of the board and interim CFO of MIILD through its sale to Matas in 2022. Today a fractional and interim finance leader for growing companies, with AI in the work from the first day. When a task needs more, Gathena brings in senior specialists, always agreed with you first.
A case: DuffBeauty
THE CASE
A growing beauty brand, and a finance function that kept up.
Since 2023, Gathena has been DuffBeauty's part-time CFO through a period of fast growth, with AI in the work from the first day. In 2026 we took it much further: a data platform in the company's own Google Cloud, an accounting agent, and answers in the AI tool the team already uses.
| — | Before | Now |
|---|---|---|
| A question about costs | Someone exported data from several systems and pieced the answer together, which took days. | The employee asks in Claude Team and gets the answer, with the records behind it. |
| Bookkeeping entries | Typed and matched by hand, one line at a time. | The agent prepares them, and the external bookkeeper approves a batch inside the accounting system. |
| A new data source | An IT project with a quote. | The agent that builds and maintains the pipelines writes a new one from a short brief, and a person at the company approves the plan. |
| Growth | More sales, through retailers and online, meant more entries and more questions. | Still no finance team of its own: a fractional CFO, AI agents and an external bookkeeping agency keep up. |
Read the full case: how it was built on Google Cloud →“We have grown the company while keeping our finance and accounting costs low, even as our operations have become more complex. Gathena has pushed AI from day one. Today I ask an AI agent for the financial information I need, and I get it when I need it.”
EARNED AUTOMATION
Agents help with the work. Automation is earned.
Most work in a company is small and ad hoc. Your agents help with it, and write down how they did each task: where they found the information, what they decided and what they did. When a task repeats and is well defined, we automate it. It runs on its own, and it is monitored.
Money leaving the company, anything going outside it, a commitment, an unclear rule and a step the law gives to a person, such as a posting in the books, always go to a person. If something changes, it comes back to you for approval.
Three stages
- You ask, the agent helpsAd hoc work
- The agent prepares, you approveWork that repeats, written down
- It runs on its own, and you check the resultsAutomated and monitored
An AI is a machine and cannot be held accountable. People can: the person who approves the work answers for it, and the company stays accountable for it. We help your people build the AI literacy to carry that with confidence.
THE FIRST STEP
Start with a conversation. Leave with a plan.
- 1
A conversation
You tell us what you want AI to do and what stands in the way. It is free.
- 2
An AI access review
We map your systems and data, who has access to what, where your data should live, and which tasks in finance, accounting and procurement your AI agents should take on first. You receive a written plan: the first agents, what each one needs, where it runs, what it costs to run, and a fixed price for building each one. It takes two weeks and has a fixed price, which we credit against the build if you continue.
- 3
The foundation and the first agents
We build them in your systems, and your team approves what needs a decision.
- 4
The next agent
Each new agent builds on the same foundation, so the next one is easier to set up.
Not everyone starts here. Some start with training, a strategy session or the data pipelines.
YOUR QUESTIONS
Is our data safe?
Your data, pipelines and your agents’ tools stay in your own cloud, and you control access. The AI model runs at the provider you choose, which processes only what each step needs.
Can our data stay in the EU?
Often, yes. Your data, pipelines and your agents’ tools can stay in EU data regions, and some AI providers are European or process data in the EU. We show what can stay in the EU without lowering the quality, and what the trade-off is where it cannot.
Will we be locked in?
No. You own everything we build, it runs on the stack you already have, and nothing moves onto a platform of ours.
How do you price the work?
We agree the result and its price before the work starts. You pay a fixed price for the result, never for the hours our AI agents work. Ongoing work, such as running the agents or a part-time CFO role, is a fixed monthly fee.
What if an agent gets something wrong?
Your agent writes down every step, so you can see where it found the information and what it did. While a task is new, the agent prepares and your team approves. Only a task that repeats and is well defined runs on its own, and it is monitored. If something changes, it comes back to you for approval.