On 21 October 2026, Cody Solutions hosted AI in Finance: What Actually Works in Mid and Enterprise Business, an official side event of Oslo Innovation Week. CFOs, finance teams and founders came for lightning talks, direct Q&A and a networking breakfast, and for the official launch of Kairos, the AI treasury built by Cody Solutions.
Why this conversation, why now
Most AI claims in finance come from vendors. This morning was built to test them against evidence: what works, where it falls short, and who stays accountable. Kjartan Tveitnes, Managing Director of NUCC, opened the event. Three speakers followed, each from a different angle: Faruk Bajramovic, Big Four, brought the evidence; Dima Medvedjko, Big Four, with more than 20 years in finance analytics, brought a practitioner's playbook; and Max Gontar, CEO of Cody Solutions, introduced Kairos.
Faruk Bajramovic: the gains are real, but not where they're sold
Faruk Bajramovic, Partner and AI Country Lead at Big Four, looked backwards rather than forwards: at randomized experiments, audited accounts and control groups. The gains are measurable. Accountants with AI close the books 7.5 days sooner. So is the cost: finance teams now spend 26% of their week checking AI output. His verdict: "AI fails at the strategic work it is sold for and wins at the work with an answer key." "Today, a human with a tool beats an agent without one."
Dima Medvedjko: start with a decision
Dima Medvedjko, Director at Big Four, has spent more than 20 years in finance analytics. "Standardized analytics became mainstream with BI. Now customized analytics is becoming cheaper with AI." He showed a weekly management report rebuilt with AI: more than 8 hours of work before, under 1 hour after. But speed alone does not change how a finance team works: "AI potential is winning it a lot of attention. Trust enables adoption." His advice to CFOs was practical: "Start with a decision. Delegate the work. Keep the accountability." Because "AI scales your judgment. It can scale your mistakes, too."
The launch: Kairos
Kairos turns that advice into a product. Max Gontar, CEO of Cody Solutions and co-founder of Kairos, shared the research behind Kairos, done before a single line of code: 25 conversations with CFOs, finance leads and founders at companies from $50M to a few billion in revenue. One question came out of them, where does the money actually go, and three answers became the three modules of Kairos.
Find: where the money leaks. Invoices paid twice, terms that drifted, payments nobody approved. "Most of it is not fraud. It is ten thousand lines a day, and nobody reading them." Kairos reads every invoice before a human opens it. Its first client stopped $137,000 of wrong payments in its first three months, before the money went out. Kairos typically identifies $1.5–4M in working capital for every $300M of spend.
Forecast: how old the number is. 16 of the 25 CFOs forecast cash in a spreadsheet, or not at all. At one design partner, before go-live, the forecast was 27 days old by the time it reached the CEO. Kairos keeps one live 90-day cash forecast across every entity, bank and currency, rebuilt every morning from ledger and bank data.
Invest: the cash that doesn't work. 21 of the 25 do nothing with spare cash. A $50M company typically holds about $1.6M idle above its buffer, roughly $60,000 a year that nobody collects. "Not because anyone is careless. Because nobody can see it." In Norway alone, companies hold NOK 1.18 trillion in bank deposits, and 88% of it is not placed for any term. Kairos aims for a 4.5%+ yield on idle cash, against a 1.5% average deposit rate.
Kairos runs inside the client's own infrastructure, read-only. A person approves every action, and the client's own bank moves the money. Max summed it up: "It was never human versus AI. It is human with AI, versus human without AI."
Kairos is now live for mid-market companies with $50M to $2B in revenue that run cash across many entities, banks and currencies but have no treasury department. The first 20 clients get a guarantee: implementation pays for itself within 9 months. AI does the reading. People make the decisions.
Curious what it would find in your own ledger and bank data? See how it works in a walkthrough with the founders.
What we took away
Three ideas ran through the morning:
- AI wins where there is an answer key. Invoices, reconciliations and ledgers have a right answer, and a machine can read every line where people can only sample.
- Trust comes before adoption. Sources, assumptions and calculations have to be visible before a CFO will act on the result.
- The human stays on the decision. Every speaker kept accountability with a person, not a model.
Thank you to everyone who joined us at Epicenter Oslo, to Kjartan Tveitnes and NUCC, and to Faruk and Dima for an honest, evidence-first morning.
