A broker I spoke to recently put it plainly: “I get three cold calls a week from AI vendors, and I can’t tell which ones are real.” That is a fair description of where many UK broking teams find themselves right now. The tools are real, the claims vary wildly in credibility, and the FCA is paying close attention to how firms use them.
What AI tools are insurance brokers actually using?
Several categories of tool have moved from demo to deployment in UK broking. Autonomous quoting platforms, AI features embedded inside broker management systems, call transcription tools, and client-facing chatbots are all in active use in 2025. The most widely deployed are workflow support tools: renewal reminders, document generation, email triage and data validation, which appear as built-in features in current commercial broking platforms.
The most visible of the newer entrants is autonomous quoting. Jointly AI, a UK-based insurtech, launched a brokerage platform for personal-lines brokers in 2024 that uses five coordinated AI agents to collect client needs, call insurers, handle IVR menus, capture quotes and return a ranked recommendation. Processes that typically took hours or multiple days complete in approximately 35 to 45 minutes, according to the company.
Further down the stack, broker management software providers such as Applied Systems are embedding AI into existing systems rather than asking firms to adopt entirely new tools. Email triage, task prioritisation and next-best-action prompts are available as workflow enhancements. Broker Central, serving the UK commercial market, reports AI supporting renewal reminders, document generation, data validation and workflow tracking as standard features on current platforms.
For client contact, tools such as CloudTalk offer AI-driven call transcription, sentiment analysis and conversation intelligence. Chatbot platforms with pre-built insurance templates handle FAQs, initial policy queries and early claims questions across web and WhatsApp channels, allowing a small team to respond outside office hours without additional headcount.
Why does this matter for your broking team right now?
The commercial case is straightforward. Broking is a relationship business where speed, accuracy and availability drive retention. A team that returns a quoted shortlist the same afternoon, answers a client query at 9pm via chatbot, and never misses a renewal date is easier for clients to stay with. The margin argument comes from handling the same book of business with less manual time spent on administrative work.
Industry surveys report productivity gains of around 30% and cost reductions of 40 to 60% in targeted processes where insurers have deployed AI. These figures are self-reported and should be read as directional rather than guaranteed. The actual result depends on which process you automate, how well it is configured, and whether your team adopts it consistently. A carefully configured renewal automation in a system your team already uses is likely to deliver. A sprawling chatbot deployment with no clear owner probably will not.
The competitive pressure is real regardless. Independent broker management platforms are adding AI features as defaults. Larger competitors are investing in autonomous quoting tools. Clients who encounter AI-powered responsiveness in other areas of their financial lives are developing expectations about broker response times too.
For a small or mid-sized broking firm, the question is less whether to act and more where to start. That question has a clearer answer than many vendors will admit: begin with the administrative layer, where the risk of a bad deployment is contained and the benefit is immediate.
Where will you actually meet AI in your day-to-day workflows?
The most practical touchpoints for a broking team are in the work that repeats daily but rarely requires expert broker judgement: reading long emails and flagging what needs action, comparing policy versions, generating routine client communications, and transcribing calls for review. These are the tasks where AI reduces time spent without affecting the quality of the professional decisions that follow.
Document comparison is worth picking out specifically. Several platforms now let a broker upload two policy PDFs and receive an automated summary of coverage differences and wording changes. An experienced broker can do this carefully in 45 minutes. A junior team member may take longer. An AI tool that surfaces the key differences in two minutes frees the broker to focus on advising the client, removing the reading time so that the judgement can happen faster.
For data-intensive risk decisions, platforms such as Planck aggregate information from multiple sources to generate automated risk insights for underwriting and pricing, while Gradient AI offers machine-learning tools for claims optimisation and fraud detection. These are more relevant to MGAs and carriers than to independent brokers at present, but they illustrate where the industry is heading over the next three to five years.
General-purpose AI tools are also in wide use. A 2024 guide for independent insurance agencies counted more than 100 AI tools in active use across the sector, with ChatGPT, Microsoft Copilot and Claude appearing alongside sector-specific vendors for the everyday writing and summarisation work that takes up broker time between client calls.
When does AI adoption make sense, and when does it create risk?
The safe zone for a UK broking team is AI on administrative and research tasks, with regulated advice and suitability decisions staying with the broker. Applied Systems recommends starting with clearly scoped, low-risk use cases: email summarisation, data extraction from documents, or renewal reminders. Measure what one use case delivers before expanding. Running manual and AI processes in parallel indefinitely doubles workload rather than reducing it.
Where to hold back: any AI tool that makes a personalised coverage recommendation without broker review sits in contested territory. Under the FCA’s Consumer Duty and the UK GDPR’s rules on automated decision-making, if a system affects pricing, eligibility, or what cover a client is offered, the client generally has the right to understand the basis of that decision and to request human intervention. Delegating that step to an AI without oversight creates compliance exposure.
The governance risk extends to data. Sending client records, claims histories or sensitive personal information to an AI vendor without a clear data processing agreement and understood data-location arrangements conflicts with ICO guidance on international data transfers and data minimisation. The 2020 Blackbaud ransomware incident, which affected UK insurance intermediaries including specialist broker Tysers, illustrated how third-party system access can cascade into a data breach across multiple organisations.
A useful rule of thumb: if an AI tool handles a task a well-trained new team member could handle under supervision, it is likely in safe territory. If the task requires the professional judgement of a licensed broker, a licensed broker should be reviewing the output before it reaches a client.
What does the UK regulatory picture mean for a broking firm?
UK brokers adopting AI have three regulators paying close attention: the FCA on conduct and Consumer Duty, the ICO on data protection and automated decision-making, and the NCSC on cyber security for AI systems. None are obstacles to adoption. Each has published clear guidance, and a firm that works through that guidance before deploying any tool is well ahead of peers who adopt first and ask the compliance questions later.
The FCA’s Consumer Duty requires that any technology used in client interactions supports good client outcomes. If an AI-driven process produces worse outcomes for certain customer groups, that risks breaching the Duty. The FCA’s 2022 machine learning survey found many financial firms struggling with explainability and bias management, and it remains a live supervisory concern.
The ICO’s guidance on AI and data protection sets out expectations for data protection impact assessments before deployment, lawfulness of processing, and transparency around automated decisions. Where AI affects what cover a client can access or the price they pay, clients have rights to understand the basis of that decision and to challenge it.
The NCSC’s guidelines for secure AI system development apply directly when adopting cloud AI tools that handle sensitive data. The core questions are where client data is being sent, who can access it, and what happens in the event of a breach. The FCA’s outsourcing guidance FG16/5 also applies where brokers rely on external AI providers for material business functions.
Carrying out a data protection impact assessment and confirming you have a data processing agreement with any vendor handling client data is the foundation of a compliant deployment. Those steps also tend to surface the right questions about a vendor before you have committed to anything.
