Dear readers,

This Wednesday we tell the story of Cluing, the Serbian startup that has evolved from a collaborative writing tool into something far more ambitious: a system for structured memory in an AI-saturated world.

Beyond the main story, we track what’s happening across region: Greek VCs-backed defense startup Alta Ares raises a $60 million Series A to accelerate autonomous air defense systems built from battlefield experience in Ukraine, while Romania’s Naratix secures fresh seed funding to expand its AI-driven e-commerce data platform globally.

We also zoom out with an ecosystem take on how Western Balkan SaaS founders can challenge weak local GTM foundations and build globally competitive distribution strategies in a market where positioning matters more than ever.

Enjoy the new edition!
Bojan Stojkovski
Editor-in-Chief, IT Logs

On a quiet evening in Novi Sad, the product looks almost deceptively simple: a stream of saved highlights, snippets pulled from articles, fragments of insight captured from meetings, books, newsletters, and conversations. But beneath all of that is a company that has already lived through a full architectural rewrite of its identity. What once was Collabwriting is now Cluing, with a shift that represents a fundamental repositioning of what knowledge work software is supposed to do in an AI-saturated world.

The company was founded in 2021 by Sandra Idjoski and Ivan Ralic in Serbia, originally as Collabwriting, a tool designed to solve a problem that felt obvious at the time: researchers and knowledge workers were drowning in information. 

The early product reflected that assumption, as a Chrome extension allowed users to highlight text across the web, save it into structured collections, and collaborate with others in shared workspaces. On top of that, an AI assistant branded as “Research Copilot” helped users discover academic papers and accelerate discovery.

The early traction validated the pain point, and following a win in Bucharest at How to Web’s Spotlight 2023 competition, venture backing worth €1.1M followed, including support from Poland-based SMOK Ventures, coupled with a grant support from Serbia’s Katapult Accelerator.

But as the product scaled, something less obvious started to emerge for its founders - users were not only collecting information, they were trying to preserve meaning.

The moment the product stopped fitting its name

The turning point did not happen in a boardroom with investors, but in repetition. Hundreds of conversations at events, most notably TechCrunch Disrupt, where Collabwriting participated in Startup Battlefield, forced the same explanation over and over again.

“We had maybe 300, 400 people we talked to in a few days. And every time we had to explain that we were not a tool where five people just write together.” Sandra recalls.

For Sandra, the pivot was more structural than cosmetic. The company realized it was not building a collaborative writing tool, but something closer to cognitive continuity infrastructure.

That distinction became the core philosophy of Cluing. “It is not just about storing things,” she adds. “It is about how your brain connects them, and making sure those connections are not lost.”

Cluing, in her words, is about interpretation rather than storage. “Cluing is like pulling people into information. Helping them find the clues in everything they are going through and curate them in a sensible way.”  she explains.

Who actually needed Cluing

Once the product shifted, so did the users. While the original positioning leaned toward researchers and content teams, real-world usage revealed a different pattern: consultants, lawyers, senior marketers, and operators working under information overload conditions.

The requirement is not summarization, but precision retrieval. “Sometimes it is about finding that third paragraph in a very long contract. Or a specific line in a report. You cannot lose that level of detail.” she says. 

That insight pushed Cluing away from generic AI productivity tools and toward structured memory systems anchored in human judgment. In a market where everything is being rebranded as AI-first, Cluing takes a more restrained position.

Sandra is explicit: “I am not sure our main value is AI. For us, it is about giving the best input so you can get a better output.”

AI exists in the system, but not as the decision-maker. It assists with onboarding, suggests structure, and helps transform collected inputs into usable outputs, while keeping human interpretation central.

Built between Novi Sad, Belgrade, and beyond

Cluing’s development is also shaped by geography. Built primarily in Novi Sad, the company is deeply embedded in Serbia’s startup ecosystem, with constant engagement in Belgrade’s more active founder network.

“We are very often in Belgrade for events. Three or four times a month we are connecting with people building different things.” Sandra says.

Exposure to other ecosystems has sharpened positioning, as time in Berlin and San Francisco helped clarify a gap in the market.

“Very few solutions are aimed at people who are not super technical but are very knowledgeable. People who do not want complex tools, but still need to preserve their knowledge.” she says.

That insight reinforced Cluing’s design principle: low friction, high cognitive value. “You should not have to learn it. It should just be there as your assistant.” she adds.

By 2026, Cluing reports around 40,000 registered users and approximately 20,000 weekly active users.

More important than scale is behavioral depth. Power users are emerging across professional categories, embedding Cluing into daily workflows rather than treating it as a passive archive.

The shift also improved investor perception. “The feedback has been good. The product right now makes even more sense than when we started.”  Sandra points out, adding that the company is preparing for a new fundraising cycle.

From snippets to systems

But the deeper ambition is structural: mapping relationships between ideas over time. “Once we have enough data, we can show second-degree connections between knowledge,” she explains.

That moves Cluing beyond note-taking and into cognitive graphing territory, where the product does not just store information but models how users think. There is also a broader systems-level bet: that human-defined relevance can eventually improve how AI systems prioritize information under rising computational costs.

“If we can understand how humans decide what is important, we can replicate that logic in systems that process information at scale.” Sandra says.

Kalshi is running a free tournament contest. Pick which country wins the World Cup, sign up, and if your pick is right, you split $1 million with every other person who called it.

The markets are already moving. Every country has a price, and that price shifts with every result, every injury, every upset. These aren't opinions. They're real money backing real conviction, updating in real time as the tournament unfolds.

Kalshi is the official regional partner of the Argentine National Team and the only federally regulated prediction market exchange in the US. Make your pick and get $10 free to trade alongside it.

Pick the Winner

_{Trade responsibly.}

IT Logs: The Western Balkans has strong technical talent, but many SaaS startups struggle with GTM. What are the biggest gaps in sales, marketing, and customer acquisition?

IT Logs: Should SaaS founders in the Western Balkans validate locally first or target global markets from day one?

IT Logs: As AI makes building software easier, distribution becomes more important. How can SaaS startups from the region stand out and scale internationally?