Automation in regulated sectors isn't new, but the $5 million seed round bagged by Panora signals a sharp acceleration: AI entering insurance processes, where data sensitivity and strict rules have long held cloud platforms at bay. Founders Diane du Paty and Fabian Langlet spent more than 200 hours watching brokers in Brest, Nice, Antwerp, and London re-type the same information into fifteen different insurer portals. The verdict was brutal: “their job has become data entry instead of advice.” Panora flips that with an AI platform that aims to give brokers back their time and accountability, built on a model-agnostic architecture and control mechanisms that echo the requirements of anyone evaluating on-premise deployment.

At its core, the platform acts as an execution layer that automates document collection, quoting, commission reconciliation, and compliance checks. It’s not a smart interface alone: the underlying infrastructure supports multiple LLMs without locking into a single provider, applies firm-level data encryption, and logs every step in traceable, auditable workflows. In a European insurance market where intermediaries juggle dozens of carrier systems that lack modern integrations, that technical choice matters. Most SaaS solutions trade internal transparency for speed; Panora seems to have grasped that for large brokers and international insurers, the ability to audit, prove compliance, and keep control over data is a prerequisite, not a nice-to-have.

Beyond cloud: what “model-agnostic” means in a sector that can’t afford mistakes

The platform’s model-agnostic design reveals a broader awareness. It’s not just about dodging vendor lock-in: it means the broker (or its IT provider) can pick the most suitable model for each task, potentially running it locally if data sensitivity demands it. Panora doesn’t tout explicit on-premise deployment, but the mix of firm-level encryption, auditable flows, and carrier-specific logic suggests an architecture ready for hybrid or private environments. In an industry where a single compliance slip can be extremely expensive, merely having the ability to trace inference and separate AI accountability from the human operator becomes a strategic asset.

The funding round, led by Isai with participation from Kima Ventures, 100in, 199 Ventures, and the founders of Pennylane, comes three months after the commercial launch, during which Panora has already signed 40 clients, including some of the industry’s largest players. The capital will expand the engineering team, deepen carrier portal integrations, and improve the reliability of AI agents. Geographic expansion, kicked off in France, Belgium, and the UK, now targets other European markets. It’s a trajectory familiar from other vertical software companies that digitized craft professions—Pennylane in accounting comes to mind: they catch a deep operational pain, save hours of manual work, and in doing so, reshape incentives. Here the broker doesn’t lose control but regains it, and the end customer gets more advice, less paper.

The structural message for anyone building or investing in AI for financial services is clear: data sovereignty and process auditability are no longer barriers to innovation, but its entry ticket. After years when the cloud seemed the only horizon, the insurance sector is showing that the real enabler isn't blind scalability, but the ability to give precise guarantees to regulators and clients. In this light, platforms like Panora, which embed security and governance best practices without tying themselves to a single stack, could speed up demand for local inference infrastructure, reducing reliance on external APIs and lowering total cost of ownership over the long haul. It’s not just about technology: it’s the answer of an industry that has understood that its clients’ data is worth more than any LLM.