Startups & Funding

Stockholm's Wilgot Raises $3M Seed to Make Products Legible to AI Shoppers

Stockholm's Wilgot raised $3M (SEK 30M) led by Industrifonden for 'Agentic Product Optimization' — turning reviews and support tickets into product data AI shopping assistants can act on.

Wilgot Raises $3 Million Seed Round to Improve Product Visibility in AI Shopping
Wilgot Raises $3 Million Seed Round to Improve Product Visibility in AI Shoppingjustynafaliszek / pixabay
By Sophie Lindqvist6 min read

Why it matters

  • Wilgot raised $3 million (SEK 30 million) in Seed funding led by Industrifonden
  • Daniel Wellington product pages optimized by Wilgot saw net sales rise 20%; 7% of Myrqvist customers now arrive via ChatGPT
  • The platform uses reinforcement learning and supports content in more than 100 languages
  • Founded in 2024 by Erik Wikander, Dr. Mhairi Dunion, and Oleg Danylenko; customers include Daniel Wellington, J.Lindeberg, Myrqvist, and Signify
  • Anders Nygren (Foundry Ventures) and Andreas Söneby (Industrifonden) join the board

Stockholm startup Wilgot has raised $3 million in Seed funding — a round it also describes as SEK 30 million — to help retailers make their products easier for AI shopping assistants to find, understand, and recommend.

Industrifonden led the round. Luminar Ventures, Foundry Ventures, Ventures Together, and Inovexus also participated. Angel investors included Johan Bäckarlin and Johan Hjelte, founders of the customer-engagement platform Voyado, plus Niclas Mollin, founder of inriver and Occtoo. Two investors will take board seats: Anders Nygren of Foundry Ventures and Andreas Söneby of Industrifonden.

The company will spend the capital developing its platform and expanding its commercial reach.

Why does this market exist at all?

The backstory is a shift in how people shop. A shopper asking an assistant for a gift for someone who enjoys hiking is describing an intended use, not naming a product. For a retailer to appear in a useful answer, its catalog needs to explain which items fit that situation and why. A photograph and a short marketing description often leave important questions unanswered.

That gap is what Wilgot calls Agentic Product Optimization, or APO. The software turns customer knowledge into richer product information and continually adjusts that content as shopping behavior changes. The stakes are real: as shopping interfaces become more conversational, retailers that can't articulate why a product fits a use case simply won't surface in the answer.

How does the platform actually work?

Wilgot starts with information retailers already hold: search queries, customer reviews, support tickets, and returns. Those sources reveal purchase motivations, recurring questions, and mismatches between what customers expect and what a product delivers. The platform turns those findings into product pages, feeds, FAQs, and structured data.

The mechanism is simpler than it sounds. Suppose shoppers repeatedly ask about the fit of a shoe. That pattern points to a gap in the product's sizing information. Making verified guidance explicit gives both a shopper and a search system more useful evidence. The value lies in extracting a relevant fact and attaching it to the right item — not in producing a longer description.

Wilgot positions APO as an active improvement layer on top of existing tooling. Visibility-monitoring tools can show whether a brand appears in an assistant's answer, while product information systems store and distribute catalog records. Wilgot's focus is changing the information those systems and assistants ultimately work with. In short: it aims to fix the inputs, not just watch the outputs.

What does the feedback loop cover?

The company's website describes three stages: discovering demand signals, generating product content, and measuring the results. Content outputs include titles, descriptions, attributes, and FAQs, generated in the relevant language. The optimization loop then tracks visibility, engagement, and conversion performance across Google and AI search, and uses those signals to revise content.

Wilgot says the approach is designed for catalogs with tens of thousands of products. According to the funding release, the platform uses reinforcement learning to improve product information from performance feedback and supports content in more than 100 languages.

The practical effect is that the system can learn which changes are useful after content is published. But there's a limit worth noting: the release does not disclose the reward design or model architecture, so those details remain outside the public description. The claimed learning takes place inside Wilgot's own optimization process — it does not mean a retailer can directly retrain ChatGPT or determine what an independent assistant recommends.

That distinction matters for buyers evaluating this kind of tool. Structured product information already has a concrete role in search: Google's merchant listing documentation explains how Product and Offer markup can make details such as price, availability, shipping, and returns available for shopping experiences. That illustrates the value of making product facts explicit — though eligibility for a search feature is different from a guaranteed recommendation.

The takeaway: Wilgot can optimize a retailer's content, but it can't control what independent assistants ultimately say.

What's the accuracy risk?

The broader issue is familiar to anyone working on product data preparation for AI shopping: fluent answers still depend on accurate underlying records. A content system must preserve the distinction between a verified specification, a customer's opinion, and an inference about suitability. Otherwise, making an item more visible can also make an unsupported claim travel further.

This is the tension at the heart of the APO category. The same generative machinery that fills gaps in a product page can blur the line between documented fact and plausible-sounding inference. Wilgot's public materials don't detail how it enforces that distinction, which leaves the question open for due diligence.

In plain terms: better visibility only helps if the facts traveling with it hold up.

Who are the customers, and what do the numbers show?

Wilgot names Daniel Wellington, J.Lindeberg, Myrqvist, and Signify among its customers. According to the funding announcement, net sales increased 20% across Daniel Wellington product pages optimized by Wilgot, while 7% of Myrqvist customers now arrive through ChatGPT.

Those figures describe different outcomes, and the announcement doesn't provide the measurement periods or experimental design. One concerns sales on selected product pages; the other concerns a customer acquisition channel. They offer early commercial signals rather than a single benchmark that applies across retailers. Buyers should read them as promise, not proof.

Who built the company?

Founded in 2024, Wilgot brings together three co-founders. Erik Wikander serves as CEO and previously held the role of chief marketing officer at Lendify. Dr. Mhairi Dunion is the company's AI Scientist, holding a PhD focused on deep reinforcement learning from the University of Edinburgh. Oleg Danylenko is CTO, with a background in data science at Klarna, IKEA, and Teemyco.

The mix is telling: marketing leadership paired with reinforcement-learning research and retail data-science experience maps directly onto the problem — turning messy customer signals into content that converts.

Wikander framed the company's argument in one line in the announcement: "Brands are spending too much time measuring whether AI assistants mention them, and not enough time giving AI a reason to recommend their products."

What comes next?

The Seed round gives Wilgot resources to expand its proposition beyond its early customer base. Its opportunity is to connect the knowledge scattered across a retailer's customer interactions — queries, reviews, tickets, returns — with the product records that influence discovery. The open questions are the ones the release doesn't answer: how the reinforcement learning is designed, and whether the early sales and traffic gains hold up under measurement with disclosed timeframes. As conversational shopping scales, making that knowledge accurate, accessible, and useful becomes a commercial task in its own right.

Source: unite.ai

Share this article:

More from Sophie Lindqvist

Sophie Lindqvist

Show full bio

Staff writer covering regulation, policy and their impact on artificial intelligence and machine learning at AI In Context.

777 articles

Related articles

  1. Delphina raises $7.5M to scale its AI context system
  2. Mirakl Bets on AI Agents and ChatGPT Enterprise to Rebuild Commerce
  3. Floqer passes $1.2M ARR and will open-source its AI prospecting database
  4. Bloomreach acquires AthenaHQ to connect AI search and personalization
  5. Mirendil Raises $1B Series A to Scale One-Researcher AI Development

« Previous article