Insly

Nora is the engine of two core jobs: turning broker submissions in any format into complete, structured files, and recommending an outcome on every claim. Insly also sells Nora on its own, to sit on top of another vendor's policy system. Product building, rating, binding, accounting and bordereaux run on rules and work without the models, so the AI speeds up the intake and triage ends of a platform that would function without it.

Every claim gets an approve, reject or needs review recommendation with a confidence score, written reasoning and the policy terms and laws it relied on, and handlers accept, reject or change it. Clients set routing rules with priorities, amount thresholds and keywords, so routine claims move on and the rest reach a handler, and the intake chat hands over to a live agent with full context on escalation triggers the client defines. Nora handles submissions in two modes that differ only in how much the underwriter reviews, and its guide reader answers only from the client's own underwriting guide, citing the section.

Each claim recommendation carries a confidence score and written reasoning that cites the policy terms, conditions, laws and limits behind it, and the underwriting guide reader cites the section of the guide it answered from. Teams can review and rate AI conversations and flag edge cases for improvement. Accuracy rates, a validation method and a mapping to a supervisory framework are not published.

Ridge Canada, a cyber MGA with eight underwriters, binds its specialist cover in ten minutes on three products built on Insly. Accelerate Underwriting's founder Scott Brown says his UK MGA runs about three traders for nearly £18 million of premium, where a typical MGA would need around 12 underwriters, and its broker portal has sold more than 500,000 policies. Encompass set up a broker portal on Insly for BIBA in two weeks. Insly's own averages, such as clients handling double the premium of their previous systems and 60 to 80 percent faster submission processing with Nora, come without a method.

Each client's data is kept strictly separate, and Insly says it is never shared and never used for model training. Clients shape the claims AI with their own terms, local laws and claims guidelines and rate its conversations to tune their own deployment, though Nora is also marketed as learning with every interaction. Where the models run, which provider supplies them and how long inputs are kept go unstated.

Insly's privacy policy reaches the software: policyholder data that clients load is treated as client data, notice to policyholders stays with each client, and personal data is processed within the EU and EEA, with transfers outside only under standard contractual clauses or equivalent safeguards. The policy still cites the Privacy Shield for transfers, and a retention schedule and subprocessor list are not public.

Insly first achieved ISO 27001 in 2020, recertified to the 2022 version of the standard in April 2025 and added ISO 27017 for cloud services at the same time. Nora protects data with AES 256 encryption and TLS 1.3, granular permissions and a full audit trail. The certification body goes unnamed, and no certificate or trust portal is posted.

Insly sells software, states no regulatory status of its own and has not mapped its AI to the EU AI Act. It rewrote more than 600 pages of internal policies, procedures and templates to align with DORA, the EU rule on digital operational resilience, which matters to the insurers and MGAs that depend on it as a technology provider, and its claims module reports in real time to support their compliance work.

A recommendation to approve or reject a policyholder's claim is a decision about a person, and Insly's safeguard is that the handler makes the final call. Beyond that commitment to human judgment, Insly publishes no fairness testing, responsible AI policy or governance process for its models.

Claims recommendations and Nora's actions sit in a full audit trail, and the accounting ledger logs every action a user or an API takes without overwriting earlier entries, so a client can trace what happened and when. The published license terms cap Insly's aggregate liability at 12 months of license fees received and exclude several kinds of business loss, and they say nothing about AI output or about recourse for a policyholder affected by it.

The models and model provider behind Nora and the claims AI go unnamed, as does the cloud provider that hosts the platform. Insly describes the AI only in general terms, as language models and document extraction.

As a policy administration platform, Insly is the system of record, and its claims module and Nora also connect to other vendors' policy systems through APIs. The API based rating service lets broker partners pull rates directly at high volume, a live database replica gives business intelligence tools the full dataset, and claims move in and out as bordereaux.

Client work names Lloyd's Canada's MGA platform and accounting software at Ridge Canada, and sanctions checks and flood mapping inside the Accelerate Underwriting portal. The API reference itself is not public.

The platform runs as a cloud service, and personal data is processed within the EU and EEA, with transfers outside only under contractual safeguards. Socrates Systems now shares the same infrastructure and security standards. The hosting provider and regions go unnamed, and an on premises option is not described.

Insly prices on system scope, implementation and the volume transacted through the platform. MGAs pay a monthly fee for a base package plus any modules they add, with implementation as a fixed fee or pay as you go depending on size, and insurers negotiate full stack projects. Users are unlimited, brokers included, and every client gets a key account manager. The figures themselves are quoted.

MGAs are the main focus, with insurers, brokers and third party claims administrators also served, and each buyer type has its own package. Clients write aviation, fleet and motor, cyber, professional lines, directors and officers cover for financial institutions and personal lines across Europe, North America and Australia, in direct, delegated and embedded models. In Poland, Insly runs a separate CRM platform for insurance agents and brokers, and Socrates Systems adds a UK platform for specialist insurers with complex underwriting.