Warsaw-based retail artificial intelligence startup Replenit has closed a $2.5 million pre-seed round, aiming to shift commerce platforms from predicting customer behaviour to reasoning about it in real time.
The round was co-led by Polish venture firm Movens Capital and Vastpoint, with additional participation from Logo Ventures, DigitalOcean Ventures, Finberg and Caucasus Ventures. Angel capital came from Mati Staniszewski, co-founder and chief executive of the London-based voice artificial intelligence company ElevenLabs.
Founded roughly a year ago by a team of six Turkish entrepreneurs — Ilyas Kurklu, Alp Karacaev, Omer Ozden, Caner Demir, Egemen Akdan and Cenk Karacaev — Replenit positions itself as a reasoning layer that sits on top of a retailer’s existing data and orchestration stack. Rather than issuing a forecast and leaving humans to translate it into action, the system interprets behavioural signals as indicators of intent and decides what to offer each shopper at a given moment.
Ilyas Kurklu, co-founder and chief executive, framed the problem plainly when the round was announced: “Retailers can no longer rely on prediction alone. They need to understand intent, reason in context, and decide what to do next for each individual customer.”
From prediction to decision
The language matters. Much of the first wave of retail AI has focused on recommendation engines and propensity scores — outputs that still require a marketer or merchandiser to act on them. Replenit argues that the next step for commerce teams is an automated layer that moves from insight to action, triggering the right offer, message or replenishment prompt at the point of intent.
The platform ingests signals such as browsing patterns, purchase timing, replenishment cycles and engagement history, and then draws on large-scale behavioural data to infer lifecycle needs. According to the company, early customers include L’Occitane en Provence and the flash-deal retailer iBOOD. L’Occitane reported a 235 per cent increase in post-purchase revenue after deploying Replenit’s engine, while iBOOD attributes 6.3 per cent of total company revenue to Replenit-driven decisions.
Those figures are self-reported, and Replenit has yet to be tested at the scale of larger incumbents. They nonetheless point to the commercial logic behind the bet: decision automation is harder to replicate than dashboards, and retailers are under pressure to extract more margin from existing traffic as acquisition costs remain stubbornly high.
A martech team with scale experience
Replenit’s founding team brings more than 40 years of combined experience in business-to-business software and martech, having previously helped build and scale companies to unicorn status. The group is based across Warsaw, with technical operations in the Netherlands, and plans to open a presence in the United States by the end of 2026.
Movens Capital and Vastpoint are familiar backers of early-stage Central and Eastern European software plays. The presence of Mati Staniszewski as an angel is notable: ElevenLabs has become one of the highest-valued artificial intelligence companies to emerge from Europe, and his participation signals continued interest from operators in backing reasoning-layer infrastructure rather than purely generative consumer tools.
What the money will do
The capital will be used to expand Replenit’s product and AI research teams in Poland and the Netherlands, deepen integrations with commerce platforms, and establish an initial commercial footprint in the United States. The company has indicated that hiring will focus on senior engineering roles and applied research, reflecting the computational demands of running decisioning in real time against live customer data.
European retail technology has seen a resurgence of interest from venture investors over the past year, with several pre-seed and seed rounds closing for companies operating at the intersection of commerce data and generative artificial intelligence. Replenit joins a cluster of startups arguing that the value in this stack is moving from prediction accuracy to decision quality — a distinction that will matter more as retailers embed large language models deeper into customer-facing workflows.
For a one-year-old company, a $2.5 million pre-seed is a measured cheque: enough to buy twelve to eighteen months of runway to prove that the reasoning-layer thesis translates into repeatable commercial outcomes. The next test will be whether Replenit can replicate its early customer results across a broader set of categories and geographies without the hands-on founder attention that early deployments typically attract.
For more coverage of European funding rounds, visit our fundraising hub.