ProYarn

Deals · Enterprise AI

Duvo raises €28.7M Series A to take enterprise AI execution from Prague to the US

A Prague-based enterprise AI company whose platform maps how employees actually do their jobs — capturing the undocumented steps that live only in people's heads — then runs those processes automatically via AI agents across ERP systems, spreadsheets, and inboxes.

ProYarn Desk · Read this in French
in𝕏✆
“Getting AI agents to work reliably across the enterprise remains incredibly difficult. Duvo's unique ability is in how it uses AI to capture how work actually happens in the enterprise and what the causes of loss are.”
Rob Moffat — Partner, Balderton Capital

Duvo (Prague) has raised a €28.7M ($32M) Series A led by Balderton Capital. Existing investors Index Ventures, Northzone, and Credo Ventures participate, bringing total funding to €43M ($48M). The company was co-founded by Tomáš Čupr, Marek Paris, and Martin Pecha — all former Rohlik Group colleagues.

What Duvo builds

Duvo describes its product as an Enterprise Execution OS with two distinct layers. The first, called Clarity, is an AI-powered process-mapping tool: it conducts structured interviews with employees, reviews screen recordings, and reads existing documentation to capture how work actually happens — including the undocumented exceptions that standard process diagrams miss. The output is a map of the real workflow, not the idealised one.

The second layer executes that map via AI agents. The agents operate across ERP systems, supplier portals, spreadsheets, inboxes, and legacy applications, connecting through APIs where available and through the regular user interface where they don't exist. Decisions requiring judgment or relationship management are routed to the responsible human; routine steps run automatically.

Customers include Notino, Morrisons, Holland & Barrett, Boden, and Allegro, as well as Rohlik itself. At Notino, the quarterly reconciliation of approximately $45 million in annual supplier bonuses — previously a 70-day, 120-specialist-hour process — now takes 15 days and under 10 hours of specialist time. The company claims more than $100 million in value unlocked across its client base, though that figure is Duvo's own calculation and has not been independently verified.

The US expansion

The Series A is primarily a US bet. Duvo will start with retailers and consumer packaged goods companies, where Čupr's Rohlik experience gives the company credibility in domain-specific process complexity. Michael Green, who joins as Chief Commercial Officer from Databricks (previously Salesforce), and Spencer Insolia, former President of Kmart and now Managing Director for the US, are already in post.

Čupr has been explicit about the logic: Rohlik, the Czech online grocer he founded in 2014 and grew to a ~$2 billion valuation, runs on thin margins and tight operational deadlines. Duvo has been a live Rohlik tool throughout its development. That doubles as a reference installation and as a forcing function: if the product can reduce costs at Rohlik, it can reduce them anywhere in retail.

The platform is model-agnostic and can run open-weight models on dedicated GPUs or customer-hosted models. It carries SOC 2 Type II, ISO 27001, and ISO 42001 certifications and is available through Microsoft's Azure Marketplace — three procurement checkboxes that US enterprise sales teams will want to see ticked.

Sources

  1. 01Rohlik Founder Tomáš Čupr's Duvo raises €28.7M Series A to Take Enterprise AI Agents to the US — The Recursive
  2. 02From Rohlik to enterprise AI: Tomáš Čupr's Duvo raises $32M Series A — Tech.eu
  3. 03Duvo Raises $32 Million Series A Led By Balderton Capital For Enterprise Execution OS — Pulse 2.0
  4. 04Rohlik founder Duvo raises $32M Series A led by Balderton — Tech Funding News

Threaded to this story

Every European round, in your inbox by 8am.

The day's seed and Series A rounds across France and Europe — threaded, sourced, and read in two minutes. Free.

Double opt-in. We'll send one confirmation email. Unsubscribe anytime.