Kovant wants to become the nerve center of enterprise agentic AI operations after pre-seed funding – News

Swedish agentic operations startup Kovant, officially known as Kovan AB, today is launching its platform globally after closing on $1.5 million in a pre-seed funding round led by J12 Ventures.

The round also saw participation from Ampli, Green Ventures and a number of prominent angel investors.

Stockholm-based Kovant says it wants to help industrial enterprises find a way to put artificial intelligence agents to work, so they can automate various business functions and enhance productivity, while ensuring they remain in full compliance and don’t expose sensitive data to unnecessary risks.

The startup has created a platform that’s focused on managing “agentic workforces.” It’s a much more ambitious approach to agentic AI than the kinds of highly specialized, one-off copilots and bots deployed by technology giants like Microsoft Corp. and Salesforce Inc. Instead, it has created a platform for managing entire fleets of specialized AI agents that can autonomously carry out essential business tasks across functions such as procurement, inventory management, supply chain operations, customer success, marketing and compliance.

To do this, it has developed a novel architecture that acts like an AI chatbot-style concierge. Enterprises can deploy and manage entire swarms of agents based on small language models, rather than more resource-intensive large language models.

Each of its SLMs has been trained to be good at one particular function, as opposed to LLMs that are designed to perform many different kinds of tasks. For instance, in procurement, it has built individual agents for coordinating suppliers, automating workflows and managing spend, risk and performance. Meanwhile, for sales and quoting operations, it provides access to agents that specialize in handling complex requests, tenders and other aspects of this work.

By using SLMs, Kovant says, it can reduce the risk of AI hallucinations. Each of its agents operates within clearly defined guardrails, making them compliant, enhancing their accuracy and ensuring full auditability. Yet despite their small size, they’re also capable of learning continuously as they operate, improving their accuracy over time, the startup said.

J12 Ventures founding partner Emmet King said Kovant’s approach is more suited to industrial enterprises because they’re still extremely reliant on fragmented systems and manual processes. “The rigid and siloed nature of these environments makes AI deployment very difficult, which is why most enterprises have seen very little ROI from the technology,” he said. “Kovant has a credible path to change this and drive real scalable impact on enterprise operations.”

Kovant may not have attracted much capital so far, for $1.5 million is just a drop in the ocean compared with the billions of dollars invested in some other AI startups, but it claims it doesn’t need a huge amount of funds either. Since launching in stealth mode a year ago, it has already signed up a number of enterprise customers in countries such as Sweden, Norway and Denmark, capturing more than $1 million in revenue to date.

Its founding team also has plenty of pedigree in the AI industry. Co-founder and Chief Executive Ali Sarrafi previously led a startup called Silo AI that was later acquired by Advanced Micro Devices Inc. Before that, he headed up many of the AI initiatives at Spotify Inc., while other members of its team include veterans from Meta Platforms Inc., Amazon.com Inc. and Google LLC.

“The team has fantastic prior experience in building data products and data infrastructure in large companies, as well as transforming enterprises with AI, and they are moving with rare speed and precision, already shaping how enterprises operationalize AI and gain huge value from it,” King said.

Saraffi explained that enterprises require a system where AI handles task execution so that people can focus on the things that require human judgment and creativity. “What matters… is real profit and loss impact, not building more agents or running endless proof-of-concepts,” he said. “[We are building] something that can be achieved in weeks, not months. This funding is about taking that approach to a wider scale and show that enterprises can get genuine, measurable results from AI.”

Ultimately, Saraffi is hoping to grow Kovant to become the next in a long list of Swedish “unicorns” that have achieved valuations north of $1 billion and reinforce the nation’s reputation as a kind of “Silicon Valhalla.” But to do that, it will need to expand beyond the Nordic markets it currently serves. And while it’s launching its platform globally, its initial marketing efforts will be focused on expanding in Europe, especially in Switzerland, Belgium and the Netherlands.

Image: Kovant