Kohort raises €5.9M for UA agents

Mobile game marketing is where a lot of AI product pitches go to either prove themselves or die fast. The work is repetitive, numbers-heavy, and brutally measurable. If a tool helps studios buy users more efficiently, the payoff shows up quickly. If it does not, nobody cares. That is the backdrop for Kohort’s new €5.9 million Series A, announced on May 8, with The Raine Group leading the round to fund a push into AI user-acquisition agents for mobile studios. ### What does Kohort actually sell? Kohort is not starting from scratch as an “AI agents” company. It already sells mobile gaming analytics, forecasting, and user-acquisition optimisation software, built around machine-learning models and cohort-based forecasting. The company says its tools are used by studios, operators, finance teams, and even investors trying to judge game performance and UA efficiency. That matters because the new agent layer sits on top of an existing prediction stack, not just a chatbot wrapper. (eu-startups.com) ### Why is user acquisition such a hard niche? In mobile games, user acquisition is basically paid growth. Studios buy ads across channels, watch install costs and player behavior, and try to predict whether a player acquired today will still spend money months from now. The hard part is that the signal arrives late but the spending decisions happen now. A team can easily optimize for cheap installs and still lose money if those users churn or never monetize. Kohort’s whole pitch is that better long-term forecasting makes those decisions less blind. (thesaasnews.com) ### So what are these “agents” supposed to do? The company is aiming at three concrete jobs: campaign optimisation, deep research, and automated reporting. In plain English, that means software that can help tune spend, pull together context on performance, and generate the reporting loop UA teams already do by hand. Kohort has been blunt about the ambition here — its CEO Dan Marcus said the company is building agents “not just a Claude wrapper,” with prediction systems underneath that let the software act on context instead of vague prompts. (eu-startups.com) ### Why does the predictive layer matter so much? Because in this market, “AI assistant” is cheap but trusted automation is hard. A mobile studio will not hand budget decisions to a model unless the model has a decent grasp of lifetime value, payback windows, and cohort quality. Think of it less like asking a bot to summarize a spreadsheet and more like giving software partial control over a trading desk. Marcus himself framed top UA teams as operating more like high-frequency traders than marketers. (gamigion.com) That is a strong claim, but it gets at the point — speed only helps if the prediction engine is good. ### Who backed the round? The lead investor is The Raine Group, which had also invested in Kohort’s seed round. That repeat backing is probably the most useful signal in the announcement. It suggests the investor is not just buying the 2026 “agents” label, but the company’s progress between seed and Series A. The announced raise was €5.9 million, or about $7 million, and multiple industry writeups published on May 7 and May 8 carried the same core details. (markets.financialcontent.com) ### Why would investors like this now? Turns out this is one of the cleaner places to fund applied AI. The buyer is obvious. The workflow is frequent. The ROI is measurable. And the distribution path is narrower than consumer AI chaos because the product is aimed at UA teams inside mobile game studios. That does not make the business easy, but it does make the pitch legible: save or reallocate ad spend, improve forecasting, and become embedded in a revenue-critical function. (eu-startups.com) ### What is the catch? The catch is that this category will only hold if the agents do real operational work. Mobile studios have seen plenty of dashboards and automation promises already. So Kohort now has to prove that its agents can move spend decisions, reporting speed, or forecasting accuracy enough to justify trust. “Agentic” branding opens the door — but retention, not novelty, decides whether this becomes software teams rely on every day. (eu-startups.com) ### Bottom line? Kohort’s raise is a small but telling bet on specialized AI, not general-purpose magic. In mobile game UA, the problem is expensive, quantifiable, and painful enough that a focused agent platform might actually earn its keep. (eu-startups.com) (gamigion.com)