Vessla Raises $3.3 Million to Run Landlords’ Back Office

Stockholm startup Vessla left stealth on October 19, 2026 with $3.3 million in pre-seed funding led by Seedcamp. The round was more than twice oversubscribed. Inception Fund, Tiny VC, operators from Lovable and OpenAI, Tandem Health founder Lukas Saari, and Fredrik Hjelm of Voi and Pit joined, along with several angels.

The money pays for AI agents that run on the software residential and commercial landlords already use, with a later aim of tying leasing, building operations and investment planning into one system.

Vessla’s Agents Sit on Software Landlords Already Run

Vessla, founded in 2026, sells a shared AI layer rather than a new property-management suite. Agents pull records, apply a firm’s own operating rules, and carry work across the tools that tenants, managers, suppliers and finance teams already live in, while people keep the calls that need judgment or a signature.

Chief executive Orman Gazvini put the product in those terms, and he did not pitch a vacant back office.

We believe AI will fundamentally change how property companies operate, from maintenance and costs to leasing and rental income. The hard part is making it work at scale across the systems and processes those companies already rely on. That is what we are building. As AI takes on more of the execution, people’s judgment, relationships and responsibility become even more important.

Orman Gazvini, CEO, Vessla

Seedcamp already files the company as an agentic OS for facility management, which is a larger claim than ticket routing. The near work is coordination. The later work is income and spend across a whole portfolio, which is where a $3.3 million pre-seed either becomes an operating system or stays a clever inbox.

WHAT WE KNOW

• The round: Seedcamp led $3.3 million of pre-seed capital that was more than twice oversubscribed, with Inception Fund, Tiny VC, Lovable and OpenAI operators, Saari, Hjelm and angels in the syndicate.
• The design: Agents sit beside current software, read context, follow house rules and execute workflows, with staff still approving binding steps.
• The clients: Vessla says it is already working with some of the Nordics’ largest property owners and operators, unnamed in the launch.

WHAT IS UNCONFIRMED

• Named landlords: No portfolio, unit count or contract has been disclosed.
• The stack: Which ledgers, lease tools or invoice systems the agents actually write to is not public.
• The price of the company: No valuation shipped with the round.

That silence is normal for a 2026 pre-seed. It is also the gap that larger rivals will press, because an agent that cannot name its systems of record is still a demo.

Why a Maintenance Ticket Crosses Four Systems

A leaky tap is not one task. Someone has to check the unit file, read the contract, raise a supplier job, then match the invoice to the right entity, and each of those steps often lives in a different programme. Staff copy the same facts by hand, which is slow and how a charge ends up on the wrong building.

THE PATH OF A MAINTENANCE REQUEST

• Property records: The agent has to see which unit, which owner entity and which service history apply before anyone is called.
• Contracts: Warranty language, service-level rules and who pays the bill sit in leases and supplier agreements, not in the work-order screen.
• Invoices: Finance still has to match a job to a cost code, a VAT line and a closing pack.
• Suppliers: Booking the trade, chasing attendance and closing the loop with the tenant is a separate thread, usually on email or chat.

Vessla’s pitch is that one agent can walk that path, write back into the connected tools, and stop when a human has to say yes. The product that will stick is the one that files the invoice and updates the lease, then gets out of the way. Another dashboard is extra load for teams that already hop between systems to finish a single job.

Janehag Built Agents After NASA Work

Kasper Janehag founded Vessla in 2026 and is chief technology officer. He earlier worked on NASA’s Mars 2020 programme, then led large AI builds at QuantumBlack, McKinsey’s AI unit, which is a consulting bench more than a property one.

Chief scientist Arash Dabiri tested agentic applications at Andon Labs and at QuantumBlack. Gazvini advised software firms at McKinsey and worked strategy and mergers at Pophouse. Other founding colleagues came through Legora, Stanford and Netlight, so the first payroll is consulting, legal AI and engineering rather than a facilities directorate.

Seedcamp partner Felix Martinez has pointed to that enterprise deployment record as the reason large property companies showed interest so quickly. For a pre-seed that has not named a customer, the team is the due diligence.

The In-House Build Goes Stale in Three Months

Oliver Molander, founder and partner at the Inception Fund, invested on a narrower thesis than “AI for real estate.” He argued that a landlord’s own build is already ageing by the time it ships, and that the winner in vertical AI will function as the customer’s outside research bench.

“AI is advancing faster than most companies can keep up. What is built in-house today could be outdated in three months. The winners in vertical AI will become their customers’ external AI R&D department, continuously turning the latest technology into practical value. Vessla has the technical depth to help property companies stay ahead. That is why we invested,” Molander said.

Operators from Lovable and OpenAI in the cap table fit that idea. People who have just shipped consumer and developer agents are putting personal cheques into a dull coordination problem, which is a tell that the next agent wave is vertical, messy and full of rules. Seedcamp leading a round more than twice oversubscribed, with Nordic angels who have already scaled Voi, Pit and Tandem Health, is the same bet in a different wrapper.

A $4 Billion Phone Agent Already Has the US Lead

Vessla is not walking into an empty category. Housing AI in the United States already has a scaled winner, and Nordic buildings already have agents in production. The $3.3 million cheque has to buy a slice that those products do not own.

THE AGENT STACK VESSLA IS WALKING INTO

Those are four different products sharing one slogan, and the differences matter more than the word “agent.”

EliseAI Already Covers One in Six US Apartments

EliseAI said on September 29, 2026 that it had raised $350 million at a $4 billion valuation, led by Andreessen Horowitz and Bessemer Venture Partners, with Ontario Teachers’ Pension Plan, Sapphire Ventures and Navitas Capital in the round. The company said it passed $200 million of annual recurring revenue in June 2026 and that its platform now powers one in six U.S. apartments.

That is a phone-and-admin machine for multifamily housing, now stretching into healthcare. New York’s Uniti raised a $12 million Series A, led by Pathlight, for agents that handle leasing, maintenance and collections for self-storage and multifamily operators. Both firms prove landlords will pay for agents that take work off a desk. Neither is a Nordic commercial-and-residential layer sitting on local ledgers and supplier files.

Nordic Buildings Already Run Hundreds of Agents

Closer to home, ProptechOS says agents on its platform have completed more than 12,000 tasks so far in 2026, with more than 300 agents in production. At 15 minutes of expert work per task, that is more than 3,000 hours. Vasakronan built an agent called SpotOn Control on the same platform. KLP cut energy use 31% at its Eufemia property. ProptechOS also says agents onboarded more than 6,000 buildings in the third quarter of 2026, against 2,000 buildings in the six years from 2019 to 2025.

That stack starts with building data, meters, work orders and an open model called RealEstateCore. Vessla starts with the office side of the same firms: rules, invoices, suppliers, leases and, later, investment planning. The overlap on work orders and leases is real. So is the opening. A landlord can already buy a building agent and still lack a shared brain for finance and leasing.

Same Headcount, Bigger Portfolios, Then the Asset Layer

Buildium’s 2027 industry report, based on more than 930 property managers, found 83 percent using or investigating AI, a combined bucket that breaks into 33% already using it or automation, 18% planning to adopt in the next year, and 32% still looking. The labor math underneath is the second-order point. 70% plan to keep teams the same size in the coming year, even as 64% plan to expand their portfolios.

WHAT PROPERTY MANAGERS TOLD BUILDIUM

• Adoption: 83% are using AI, planning to adopt it, or actively investigating it.
• Headcount: 70% intend to keep team size unchanged over the next year.
• Growth: 64% plan to take on more doors anyway.
• Jobs AI already touches: Resident communication 48%, internal task management 45%, maintenance coordination 36%, leasing 34%.

More doors with the same payroll only works if coordination stops eating the day. Morgan Stanley Research estimates 37% of tasks in real estate can be automated. ProptechOS’s own study from early 2026 puts the share for building operations at 48%. Those are different bases, and neither is a forecast of Vessla’s revenue.

Vessla’s longer goal is the piece most ticket tools never reach. Agents would coordinate work, hunt for ways to raise income and cut costs, and support asset performance across a portfolio, which is an investment product wearing an operations badge. That is also where a bad agent does real damage, because a wrong lease read or a misposted invoice hits cash, not a chat log.

AI can look impressive in a demo. The real test is whether it can handle the messy, consequential work of a complex business. That requires deep integrations with vast amounts of context, rigorous evaluation and clear limits on what agents can do. We are building Vessla to meet that standard across entire property portfolios.

Kasper Janehag, founder and CTO, Vessla

The Nordic owners already on the platform, still unnamed, are the proof that standard either holds or it does not. Until those names and those systems are public, $3.3 million has bought Vessla a seat in a race that EliseAI and ProptechOS are already running, on opposite sides of the Atlantic, with a lot more miles on the clock.