Jenz Raises €540,000 to Glue a Store’s Systems Together

Jenz has raised €540,000 in pre-seed funding from Antler and Ada Ventures to run retail back-office work with an AI agent.

Innovation Norway added a NOK 1 million grant (€92,000). The Oslo company is launching the product at Antler’s demo day on October 21, 2026.

The cheque pays for a shared data layer on top of the tools a shop already runs. An agent then works that layer and pings a person when money is about to move.

They Won in Seven Hours, Then Ada Wrote a Cheque

Remi Bjørnestad, Jo Schou Sandåker and Natnael Belete Seifo arrived in Mandal still hunting for a company. Sandåker, now chief technology officer, had just shut the AI startup Bicone AI. Bjørnestad, now chief executive, had spent three years as a consultant building AI agents for firms. Seifo, now chief commercial officer, had been building an agent platform of his own after stints at Foodora and Elkjøp Nordic.

Ada Ventures and Startuplab picked them as one of eight teams for the Norway AI Hackathon at The Foundry on June 3, 2026. The brief was blunt. Build a working product in seven hours, on open-source models only, with no closed frontier APIs doing the heavy lifting. Ada’s published prize was a £150,000 investment for the winning team, structured as a convertible for 5% with a post-money cap, if the whole team went full time.

They built one of the first pieces on the Jenz roadmap and won. Ada became the first investor. The three then left Kristiansand for Oslo. Norwegian company records show Jenz AS was formed on June 15, 2026, with Bjørnestad listed as chair and managing director from July 6, and headcount in the 1-4 band.

Michael Tefula, principal and head of product at Ada Ventures, said the three came together as a team and in seven hours built something that would otherwise have taken several days. He also pointed at a UK path from the start, which is now written into the company’s plan.

FROM MANDAL TO MESH YOUNGSTORGET

1. June 3, 2026: The trio wins the Norway AI Hackathon at The Foundry in Mandal.
2. June 15, 2026: Jenz AS is formed, with Bjørnestad later listed as chair and chief executive.
3. August 2026: The team joins Antler’s eight-week Oslo residency and starts onboarding pilot shops late in the month.
4. September 2026: Antler joins the cap table.
5. October 21, 2026: Jenz launches the product at Antler’s demo day and discloses the €540,000 pre-seed.

Anna Munthe-Kaas, a partner at Antler, said Jenz shows what a hungry team can do when it moves fast. The speed is real. The open question is whether a 7-hour open-source demo becomes a product that can sit on a live till, a warehouse, and a set of ad accounts at once.

What the Agent Does Once It Is Inside a Store

Jenz’s own site calls the product an agentic back office for commerce teams. It asks a shop to plug in Shopify and the tools around it, then keep one agent in the chat apps the staff already open.

The pitch is not a new dashboard. Sandåker has said most software is built to keep people inside it, and that Jenz is built the other way. Staff hand the agent a job. It watches the connected systems around the clock and tells the right person when something is wrong, needs follow-up, or needs a decision.

Most retailers we talk to want to use AI, but don’t know where and how to start. Many have tried generic AI tools and run into the same problems: they start from scratch every time and can’t handle the amount of data a store actually has. When AI is making decisions that cost real money, you have to make sure it acts on real and trusted data. That’s why we built the Jenz agent.

Remi Bjørnestad, co-founder and CEO of Jenz

On the product site the work is scheduled, not conversational. A restock check can run at 07:00. A margin report can land in Slack on Mondays. Return replies are drafted for a human to send. Ad spend can be watched every hour. A purchase order still waits for a tap on Approve.

One sample restock ping on the site warns that a wool coat in size M will sell out in six days and puts a purchase order on the table. That is demo copy, not a named customer. It still shows the job the company wants to own: the unglamorous work that sits between a stock figure and a supplier portal.

WHAT JENZ SAYS IT ALREADY KNOWS HOW TO DO

• Inventory planner: Days of cover, reorder points and a weekly forecast from the shop’s own orders.
• Ads manager: Spend, return on ad spend and creative fatigue across Meta and Google, with budget moves waiting for approval.
• Norwegian VAT file: MVA per period from Shopify and Fiken, prepared for Altinn.
• Klaviyo recap: Open rate, click rate and revenue per campaign over the last 30 days.
• Return reasons: Group returns by product and reason, then name the top three.

The same page lists a Norwegian VAT skill beside a supplier price-list check. That is a local tell. The first shops are Nordic. The agent is being taught the boring files those shops already file.

Jenz says the agent can live in Slack, email, Microsoft Teams, WhatsApp or iMessage, with one memory across those channels. Hosting defaults to the EU, with Azure in Amsterdam listed as the active region and Google Vertex in the Netherlands as a second option. The company says shop data is not used to train shared models.

The site also claims 3,000-plus apps and integrations. That is a platform figure, not the same 3,000 as the pilot task count, and it is the company’s own marketing claim.

Pilot Shops Have Already Thrown 3,000 Jobs at It

Jenz began onboarding pilots in late August, mixing pure online shops with chains that run several physical stores. The company has not named them.

Those shops are the only proof that the data layer is more than a hackathon demo. The figures below are Jenz’s. They have not been audited in public, and “up to” and “one retailer” are doing a lot of work.

PILOT RESULTS JENZ HAS PUBLISHED

• Task volume: Pilots have assigned the agent more than 3,000 tasks.
• System spread: Those jobs ran across more than 20 back-office systems.
• Time saved: Some pilots reported saving up to 100 hours of work in the first month.
• Capacity: One retailer recorded a 31% increase in team capacity after putting the platform in.

The useful read is the mix, not the round number. A shop that can hand an agent work in more than 20 systems has already done the painful part, which is letting a vendor sit on till data, stock, ads and accounts at the same time. That is also the part generic chat tools never finish, because each new question starts from a blank context window.

The first jobs on the product site are the ones a small ops team already does on a timer: restock in the morning, a Monday pack of numbers, return drafts, an hourly glance at ads. That is a narrower product than “replace the back office.” It is also a clearer one.

Shopify’s Own Assistant Stops at the Admin

Any store already on Shopify has an official assistant. Sidekick lives inside the Shopify admin, reads that admin’s data, and can run tasks there rather than only suggest them. Shopify’s own help pages treat Sidekick as the in-admin assistant and say a Microsoft Copilot plugin does not replace it, and does not replace the admin itself.

That is the incumbent Jenz has to live beside, not a rival in another country. Sidekick is free to a merchant in the sense that it is already in the building. Jenz is asking that same merchant to pay for a layer that reaches the tools around Shopify: point of sale, ads, finance, mail, returns.

JENZ VERSUS SHOPIFY SIDEKICK

The gap Jenz is selling is real only if a shop’s pain lives outside the admin. Plenty of it does. Stock in a warehouse tool, spend in Meta, books in Fiken, mail in Klaviyo. A Sidekick that cannot see those systems will keep answering a partial question. A Jenz that cannot stay cheaper and simpler than a bundle of native assistants will get squeezed as soon as Shopify, Meta or the accounting tools grow their own connectors.

Bjørnestad’s line about generic tools starting from scratch each time is the cleanest version of that gap. It is also the test. If Jenz forgets a shop’s rules overnight, it becomes one more generic tool with a nicer login.

The Cheque, the Grant and a Matching Promise

The closed round is €540,000 from Antler and Ada Ventures, about 5.9 million kroner. On top of that sits the NOK 1 million Innovation Norway grant, which the company converts at €92,000. Counted together that is €632,000 of fresh capital, with the grant as a separate public cheque rather than equity.

Ada’s first money was the hackathon convertible. Antler came in later, after the team entered the Oslo residency in August and after Antler took equity in September. Antler’s published Nordic terms are not a window into Jenz’s exact split, and the company has not broken out how much each fund put into the €540,000.

Those terms still shape what happens next. Antler says it puts €200,000 wired after committee approval, split as €100,000 for 8.5% and €100,000 on an uncapped SLIP with a most-favored-nation clause. It also commits up to €300,000 of matching capital if the company raises from outside investors within nine months of that first cheque, and it says the terms are not negotiable.

Munthe-Kaas and fellow partner Kristian Jul Røsjø are named on that fact sheet as the Oslo partners who stay close after the cheque, with a weekly cadence aimed at customers, revenue and the next raise.

The public money is a Startup Grant 2 cap of NOK 1 million, and 2026 is a harder year to win it. Innovation Norway says funding for the scheme has been cut by 60% this year, with tighter triage and more rejections. The grant can cover up to 50% of approved project costs, and the company must put in at least as much of its own money. Jenz still got the full cap.

The agency also says it will not distort markets that already work, and it lists traditional retail, including e-commerce, among those markets. Jenz is software for shops, not a shop. The grant still had to clear a bar that got higher this year, which is why the NOK 1 million is not a footnote.

Britain Is Next on a Nine-Month Clock

Jenz wants the product to become fully self-serve and is already testing demand outside Norway. After the Nordics, the next market is the United Kingdom, with a launch planned within nine months of the October 21, 2026 announcement. Ada is a London fund. Tefula has already said the firm wants to help the team into Britain.

That UK timer is not Antler’s matching timer, even though both are nine months long. Antler’s match, on the published terms, runs from its own September investment. The UK plan runs from the October launch. They sit in the same season. They are not the same clock.

The product is being put on stage at the Antler Demo Day at Mesh Youngstorget on October 21, 2026, from 09:30 to 12:00, during Oslo Innovation Week, which runs October 19 to 23. The listing is invite only. Luma showed 120 people going. The address is Møllergata 6.

What the room will see is a three-person company with a €540,000 pre-seed, a NOK 1 million grant, more than 3,000 unnamed pilot tasks, and a promise that a human still has to approve the purchase order. The next test is whether a British shop, on a different till and a different set of books, will hand over the same keys.