Artificial intelligence
Zelros is a French company specializing in AI applied to the insurance and financial services sector. It was founded in 2016 in Paris by Christophe Bourguignat, Fabien Vauchelles and Damien Philippon. The founders, who come from insurance and data science backgrounds, identified early on that the insurance professions were facing two major challenges: increasingly dematerialized customer relations and increasingly complex products. Zelros’ mission is to transform the experience of advisors and agents by providing them with an intelligent co-pilot that enhances the personalization of offers, anticipates customer needs and automates certain time-consuming tasks.
The company’s B2B SaaS status has enabled it to raise $16.5 million from leading European funds such as BGV, ISAI Cap Ventures, Plug and Play, Hi Inov, 42CAP and astorya.vc. Zelros now has offices in Paris, Montreal and New York, a sign of its international ambitions. The company is positioned as one of the European leaders in generative AI applied to insurance, and has organized emblematic events such as the Data Science Olympics, a machine learning challenge bringing together hundreds of data scientists every year.
Zelros offers a SaaS platform hosted in sovereign environments and designed for players in the insurance, mutual and banking sectors. The solution comprises four main capabilities:
Magic Questions – Needs discovery: the tool identifies a customer’s needs by analyzing their profile, history and previous interactions. Zelros models are trained from thousands of anonymized dialogues and business knowledge to ask the relevant questions that will enrich the advisor’s understanding.
Magic Recommendations – Personalized advice: the recommendation algorithms use reinforcement learning techniques to suggest products or services tailored to each customer. As explained in a Forrester report featured on the Zelros website, the start-up employs similar techniques to those used by OpenAI to refine ChatGPT: the AI receives human feedback and continuously adjusts its recommendations.
Magic Answers – Instant answers: the agent answers the customer’s questions in natural language, based on the insurer’s data and state-of-the-art linguistic models. These answers cover regulatory topics (claims conditions, contractual obligations) as well as advisory topics, and can be used by human agents or directly in self-service via a chatbot.
Magic Automations – Process automation: workflows automate data entry, contract document preparation, notification and claims management. The aim is to increase productivity, reduce processing times and make processes more secure.
Under the hood, Zelros combinesgenerative AI models (GPT, LLaMA, Mistral, Granite, etc.) with its own recommendation algorithms. The company offers a training studio for integrating new models, including open source models. In February 2025, Zelros announced a partnership with IBM to integrate watsonx, IBM’s suite of AI tools. This partnership aims to bring additional predictive and customization capabilities, and allow customers to choose to host the model in their own infrastructures (sovereignty option). IBM emphasizes watsonx’s compliance with financial sector regulations, and Zelros sees this as a guarantee of confidence for its customers.
Zelros has been recognized by several analysts and research firms:
Gartner named it Cool Vendor in its 2023 report on Insurtechs Adding GenAI. This distinction underlines the innovation and relevance of the solution in a traditionally conservative sector.
CB Insights cited Zelros as an exemplary GenAI solution for insurance distribution in its report How GenAI Is Reshaping the Insurance Value Chain.
Forrester highlighted Zelros’ use of reinforcement learning to deliver personalized advice.
The company regularly organizes webinars, training courses and popularization content to support insurers in adopting AI. It collaborates with universities, startups and major groups (Microsoft, Capgemini) to promote European sovereignty in AI.
Agents and advisors: Zelros provides a co-pilot interface used by banking agents and advisors. During an interaction, the agent has access to a 360° view of the customer, dynamic scripts, product recommendations and ready-to-use answers. This increases conversion rates by 65%, generates a 30% increase in cross-selling and improves productivity by 15%, according to figures published on the site.
Contact centers: for customer service and back-office departments, Zelros automates ticket analysis, request routing, file preparation and the formulation of standardized responses. The solution integrates with existing call center platforms.
Marketing and digital: marketing teams use templates to generate personalized content (landing pages, emails, banners), segment customer bases and test different proposals to improve the customer experience.
Data & IT: the platform offers low-code, API-first access for data scientists and IT managers. It enables rapid integration of new data sources, regulatory compliance (RGPD, DORA) and model deployment in sovereign environments.
In April 2025, Zelros was in the news whenEarnix, an Israeli player specializing in predictive underwriting, announced the acquisition of the company. The aim is to combine Earnix’s predictive capabilities (premium calculation, risk management) with Zelros’s generative and agentic functions (recommendation and interaction) to offer an end-to-end solution: from rate decision-making to real-time recommendation. This synergy aims to increase conversion and customer satisfaction, and meet regulatory requirements. Earnix’s existing customers (Aviva, AXA, BPCE, IAG, Generali, Matmut, Santander, Tokio Marine) will benefit from this innovation, and Zelros will gain the resources needed to accelerate its expansion into new markets.
Following this acquisition, Zelros retains its brand identity, but joins the Earnix teams. This change of ownership should strengthen the company’s ability to invest in R&D and expand into North American and Asian markets.
Zelros has a culture ofambition (long-term vision, technological leadership),humility (collective progress) and trust (transparency towards customers, partners and employees).zelros.com. The company puts forward values of ethics and responsibility: its solution is designed to respect data protection rules (RGPD) and provide transparent recommendations, explaining the logic behind each suggestion.
Zelros anticipated the growing importance of generative AI in insurance, a sector characterized by strict regulations and high distribution costs. By bringing intelligence to customer relations and automating administrative tasks, the company offers its customers a significant competitive advantage. Its partnership with IBM strengthens the performance of its models and its regulatory compliance, while its acquisition by Earnix gives it the financial and commercial clout of a recognized group.
Looking ahead, Zelros faces a number of challenges:
Post-acquisition integration: maintain the culture of innovation while integrating into the Earnix structure and aligning product roadmaps.
Increased competition: the insurtech ecosystem is seeing the emergence of similar platforms (Allphins, Shift Technology). Zelros will need to maintain its differentiation through specialization and strategic partnerships.
Regulatory developments: AI in insurance is under scrutiny (draft IA Act). Zelros will have to prove the compliance of its algorithms (explainability, absence of bias) and anticipate new standards.
In conclusion, Zelros embodies a European success story: in less than ten years, the start-up has become a leader in AI for insurance and banking, recognized by Gartner and CB Insights, with prestigious customers and robust technology. Its future, now linked to Earnix, will depend on its ability to maintain innovation and conquer new markets while respecting ethical and regulatory standards.