Underwriters are at the heart of any insurance business. Their ability to assess risk accurately, based on a wide variety of information, is critical to product innovation, effective pricing and ultimately whether a business can make a profit in an increasingly competitive market. Optimising underwriting is a constant pursuit and today that increasingly means artificial intelligence (AI).
Underwriting has already advanced enormously in recent years. Underwriters now sift through rafts of proprietary and third-party data and the challenge has shifted from “do we have enough data” to “how do we process and analyse it fast enough to keep up with customer expectations.” That’s where AI comes in and where the conversation has moved fastest: from single-model automation to agentic AI, where orchestrated AI agents handle whole segments of the underwriting workflow rather than just answering one question at a time.
How is artificial intelligence being used in insurance underwriting?
AI technologies including natural language processing, large language models, generative AI and machine learning can ingest, process and analyse huge volumes of unstructured data far faster than humans. That’s no longer a promise, it’s in live production at major carriers.
Hiscox has proven this in production: a 99.4% cycle time reduction for London Market specialty lines, cutting quote turnaround from three days to three minutes, while keeping underwriters in control of final pricing.
AIG has moved beyond single-model generative AI to a multi-agentic solution, an orchestration layer coordinating specialised AI agents across underwriting and claims. Its internal tool, AIG Assist, is now live across most of its commercial lines and CEO Peter Zaffino has said AI has advanced faster than the company expected over the past nine months. Separately, QBE’s Cyber Underwriting AI Assistant has driven a 65% reduction in review times by handling initial submission reviews.
Adoption is broadening too: nearly half of underwriting and insurance executives already have AI built into daily workflows, according to Pacific Life’s 2026 Underwriting Outlook Survey and 70% of insurance operations leaders now have AI deployed in live operations, up from 58% a year earlier, per Covenir’s 2026 Insurance Operations Leaders Trends Report.
What are the benefits of artificial intelligence in insurance underwriting?
AI is genuinely changing underwriting economics. A McKinsey survey of more than 50 insurance leaders found gen AI could drive productivity gains of 10 to 20 percent, premium growth of 1.5 to 3.0 percent and a similar improvement in technical results. Here’s why:
- Increased efficiency and accuracy: automating data collection, form-filling and document review speeds up underwriting significantly, boosting customer satisfaction and cutting admin costs.
- Risk assessment and management: AI can spot patterns in unstructured data that human underwriters might miss and feed claims data back into underwriting decisions, sharpening pricing and reducing loss ratios.
- Fraud detection and prevention: fraud isn’t going away. UK insurers detected £1.16bn in fraudulent claims in 2024, up from £1.14bn in 2023, across more than 98,400 claims, a 12% rise on the year before. AI’s ability to spot suspicious patterns across large datasets is a direct response to that trend.
- Freeing up underwriters: automating the repetitive work frees experienced underwriters for the complex risk assessment, strategic thinking and product innovation that still needs a human.
Is artificial intelligence a panacea for insurance underwriting?
No. Treating it as one is the risk. AI tools are only as good as the data they’re trained on, poorly trained or incomplete data risks biased or discriminatory outcomes, the opposite of what the system is meant to deliver.
Trust matters too. Many customers still want to know a human is involved, particularly on data privacy, security and regulatory compliance. Human underwriters remain essential as the human-in-the-loop, overseeing both the inputs and outputs of an AI system and keeping decisions fair, transparent and genuinely responsive to customer needs. As Generali’s Farjzaneh Traheritabar put it: “AI doesn’t replace jobs, it replaces tasks. Underwriters who adopt AI are poised to outpace those who don’t.”
The future of artificial intelligence in insurance underwriting
AI is changing the face of underwriting and the pace of change has, if anything, accelerated since agentic AI moved from pilot to production. But adopting it is a genuine transition, not a plugin. Insurers and MGAs need the right data foundations, an upskilled workforce and a plan for the cultural change involved. The sooner companies start experimenting, the less ground they have to make up as the technology keeps moving.
How to implement AI in insurance underwriting
- Start small: apply AI to one clearly defined underwriting problem first, rather than trying to transform everything at once.
- Set your objectives: decide what you’re trying to achieve, faster decisions, a lower loss ratio on a specific product, before you pick a tool.
- Define clear KPIs: set targets and benchmarks up front so stakeholders agree on what “good” looks like.
- Manage the risk: think through what could go wrong early and communicate clearly with staff about how the change affects their work.
- Digitise the underlying data and processes: AI is only as good as the data behind it. This is where software like Insly and Nora, Insly’s AI layer, comes in, digitising underwriting and the wider workflow to support a proper data strategy.
- Invest in the right skills: if AI is going to be core to your underwriting strategy, budget for people who understand it, whether that’s upskilling your team or bringing in outside expertise.
- Use what already exists: there’s no need to build from scratch. Tools like Nora are built specifically to remove manual work from insurance operations without adding headcount, which can save significant development time.
- Test and test again: AI implementation isn’t an exact science. Trial and error is part of getting it right.
If you’re taking the first steps on your AI journey, Insly makes it simple. Nora, Insly’s AI layer, removes manual work across the underwriting lifecycle, from submission intake to document extraction and sits on top of a platform that already covers product design, distribution, accounting and claims. The Claims Portal takes that further downstream, giving every claim an AI recommendation with a confidence score and referenced policy terms.
We work with 70+ MGAs and insurers and on average our customers handle double the gross written premium they managed on their previous systems.
