For decades, underwriting portfolio management has been guided largely by hindsight. Insurers reviewed performance through monthly management reports or quarterly portfolio reviews, identifying trends only after risks had already been written and exposures had begun to develop.
That model is now beginning to shift. Advances in data infrastructure and artificial intelligence are allowing insurers to embed portfolio insight directly into underwriting decisions, bringing real-time intelligence into what has historically been a periodic process.
The scale of the shift is reflected in growing investment priorities across the industry. According to a 2025 study from Accenture, 71% of underwriting executives believe investment in AI and automation is critical or very critical for improving underwriting performance.
Rather than analysing results weeks or months later, insurers are increasingly looking to give underwriters visibility of portfolio impacts at the moment a decision is made.
“Traditionally, portfolio management in insurance has been done retrospectively through monthly MI packs, quarterly reviews, and corrective actions based on this information,” said Pankaj Patil, Vice President of Insurance Products at IntellectAI.
“AI is shifting this toward being a continuous process. In modern underwriting workbench environments, portfolio signals such as concentration build up, pricing deviation and risk clustering are visible during the underwriting decision itself in real time, rather than weeks later in reports.”
From hindsight to foresight
According to the aforementioned Accenture study, 75% of underwriting organisations will have been impacted by the implementation of AI/GenAI in a meaningful way in the next three years. This compares to just 17% as of September 2025.
This expected shift is dramatic. And akin to moving away from paper and pen to the advent of the internet in its seismic nature for underwriters, with portfolio intelligence moving into the daily underwriting workflow.
“It moves portfolio control from a back-office analytics function into the day-to-day underwriting workflow,” Patil explained. “Decisions are no longer just case-based. They become portfolio aware at the moment they are made, while remaining fully controlled by the underwriter equipped with the right insights.”
In practice, this allows insurers to identify emerging patterns across the portfolio much earlier than before.
“Traditionally, portfolio management was a rear-view mirror exercise,” explained Dan Simmons, Managing Director and Founder at Quensus. “We are shifting this to live asset intelligence.”
This is already beginning to surface in the water risk management sector that Simmons primarily operates in. He explained that through platforms such as FlowReporter, insurers are already able to monitor water-related risk data across property portfolios in real-time.
“AI monitors water data across a portfolio continuously. If a building’s risk profile changes, whether due to occupancy shifts or ageing infrastructure, the underwriter is informed immediately rather than waiting until the next annual renewal.”
The result is a more responsive form of underwriting oversight, where risk signals can be acted upon as they emerge rather than being discovered after losses have already occurred.
Strengthening underwriting discipline
Beyond improving visibility, AI is also helping insurers bring greater discipline and consistency to underwriting decisions. Patil explained that modern underwriting workbenches are designed to provide underwriters with contextual insight at the moment a case is reviewed.
“AI improves discipline by making guidelines, peer benchmarks and appetite signals visible at the point of decision,” he said. “When underwriters see comparable risks, historical outcomes and exposure impact alongside the case, exceptions become more deliberate, fact-driven and better documented.”
He added that AI can also reduce the selective bias that often shapes human judgement.
“The human memory recalls recent or unusual cases, while AI references the full book. This broader context strengthens risk selection and keeps underwriting behaviour aligned with portfolio strategy.”
For Simmons, the same principle applies in property risk, where insurers are increasingly able to link underwriting decisions with measurable risk mitigation.
“AI allows for evidence-led underwriting,” he said. “Instead of relying on blunt pricing, underwriters can reward active mitigation.”
For example, technologies that monitor water systems within buildings can change how insurers assess exposure to one of the industry’s most common causes of claims.
“A building equipped with Quensus LeakNet is fundamentally a lower risk, and AI-enabled portfolio steering allows insurers to recognise and capture that higher quality business.”
From analysis to portfolio steering
As these capabilities mature, the benefits are becoming increasingly tangible for insurers operating in high-volume environments.
Patil noted that AI-assisted underwriting environments are already improving operational performance by helping underwriters process submissions more efficiently while maintaining consistency across decisions.
“AI-based ingestion and case summarisation allows faster submission triage, while deviation alerts and peer case comparisons improve pricing consistency,” he said.
These tools also help insurers detect changes in portfolio composition earlier, identifying segment overexposure or shifts in underwriting appetite before they become structural issues.
In parallel, Simmons said the ability to analyse operational risk data continuously is already influencing portfolio outcomes.
“Insurers are seeing a stabilisation of loss ratios,” he said. “By identifying patterns in leak data before they develop into catastrophic escape of water claims, insurers can reduce the frequency of some of the industry’s most expensive attritional losses.”
Taken together, these developments suggest that the modern underwriting strategy is moving toward a more dynamic model of portfolio management.
Rather than relying on retrospective analysis to guide the next move, insurers are beginning to steer their portfolios in real time, combining human decision-making with data-driven insight to make more informed decisions as risks are written.
As AI becomes more deeply embedded in underwriting workflows, the defining feature of the modern underwriting strategy may not be the technology itself, but the ability to connect individual underwriting decisions with the performance of the wider portfolio at the exact moment those decisions are made.
