Cobl Raises €6M to Automate B2B Sales Proposals with Multi-Agent AI

Paris-based Cobl has secured €6 million in fresh funding to accelerate the commercial rollout of its multi-agent AI platform, which turns a sales rep’s messy deal context into polished, on-brand commercial documents. The round, announced on 7 April 2026, comprises €4 million in equity led by Eurazeo and €2 million in debt financing from Bpifrance, BNP Paribas, Banque Populaire and CIC. SuperCapital-Side Angels, Apok Invest and Paris startup campus Station F also participated in the equity tranche.

The raise positions Cobl within an increasingly crowded category of AI sales proposal software, where generative tools promise to shorten the grind of producing RFPs, pricing pages, security questionnaires and case studies. Cobl’s pitch is that a horizontal AI assistant is the wrong tool for the job — and that a specialised, multi-agent architecture produces materially better commercial documents.

From Thinkeo to a specialised sales AI

Cobl was founded in 2020 by Damien Hontang and Quentin Marquet, originally under the name Thinkeo with a broader mission to help enterprises build AI applications. After closing an initial €1 million round in September 2023, the pair narrowed focus sharply, relaunching as Cobl.ai with a single question in mind: what if every asset a salesperson has to produce — proposals, RFP responses, ROI calculations, security documentation — could be generated from a short brief and the company’s existing knowledge base?

The company operated out of Station F and has now grown to a team of 12. Early customers span French enterprises including Orano, Vinci Energies, Econocom and Free Pro, with Cobl primarily targeting companies of 50 to 1,000 employees in consulting, software publishing and telecoms.

A multi-agent architecture, not a single model

Cobl’s differentiation rests on how it assembles a document. Rather than prompting a single large language model, the platform deploys an orchestrator agent that interprets the sales rep’s brief and then dispatches specialised agents for distinct tasks: layout, brand compliance, visual search, chart and diagram generation, proofreading and final quality control. Each agent can request information from the others and adapt iteratively to user feedback.

The platform plugs into the systems B2B sales teams already use — Salesforce, Notion, SharePoint, Microsoft Teams and Gmail — pulling in past proposals, call notes and collateral to personalise each output. Documents can be exported as PDF, Word or PowerPoint files.

Pricing starts with a free tier limited to three documents per month, moving to a paid subscription at €29 per user per month.

Why investors are paying attention

The sales proposal automation market has become a natural early target for generative AI: the work is high-volume, structured, repetitive and directly tied to revenue. Analysts increasingly view time-to-proposal as a measurable sales efficiency metric, and finance-backed sales enablement teams are beginning to budget for it accordingly.

For Eurazeo, which last year raised a €650 million fund specifically targeting AI startups with more than €10 million in annual revenue, Cobl fits a broader thesis that vertical, workflow-specific AI products will outperform general-purpose assistants in the enterprise. The debt line from France’s core banking syndicate suggests Cobl already has the commercial traction to support non-dilutive financing alongside the equity cheque.

What’s next

Chief executive Damien Hontang has been explicit that Cobl is building for international markets from day one, with initial commercial campaigns planned for the United States and the United Kingdom. There are no immediate plans for a hiring spree; instead, the company intends to rely heavily on its own AI tooling internally to stay lean as it scales.

The competitive field is not empty. Cobl will have to carve out space against incumbents in sales enablement, horizontal copilots such as ChatGPT Enterprise and a growing crop of AI-native proposal tools. Its best argument — that a specialised, agentic architecture trained on the specific shape of B2B sales documents beats a generalist — is one the market will now test at scale.