Commercial P&C Underwriting: AI Tool Adoption Surges but Strategy Confidence Lags

The Survey Behind the Signal

Convr’s 2026 Insurance Talent and Tech Trends Survey draws on responses from 211 commercial insurance professionals, making it a credible benchmark for where the sector actually stands rather than where it aspires to be.

The headline finding is not that AI adoption is slow. It is that adoption has outpaced the organizational infrastructure needed to make it durable. Tools are being introduced faster than the frameworks, training programs, and operating models required to extract consistent value from them.

This is a pattern the broader AI tools ecosystem has seen repeatedly across industries. The difference in insurance is that the stakes — risk assessment, pricing accuracy, regulatory exposure — are structurally higher than in most other sectors.

What Is Actually Slowing Underwriting Down

Before examining the confidence gap, it is worth understanding what underwriters themselves identify as their core friction points. The survey is precise here.

Manual data entry leads at 35.1%, followed by legacy and dated technology at 27.5%, and too many submission data sources at 24.6%. These are not abstract technology complaints. They describe a daily operational reality in which underwriters spend significant time on tasks that do not require underwriting judgment.

The structural context matters: 63% of respondents operate in hybrid technology environments — legacy core systems overlaid with cloud-based tools. This architecture creates integration complexity that no single AI tool resolves on its own. Adding a new AI layer to a fragmented stack does not eliminate the fragmentation; it adds another surface to manage.

The Confidence Gap in Detail

More than 40% of leaders placed themselves in the bottom half of the confidence scale when asked about their AI strategy. A further 56.9% described their organization’s attitude toward AI as cautiously open — interested, but still waiting for proof before committing more deeply.

This is not a sign of technophobia. It is a rational response to deploying tools without a clear measurement framework, integration roadmap, or change management plan. When organizations cannot answer basic questions — which workflows are producing ROI, how AI decisions are audited, what happens when a model underperforms — confidence appropriately remains low.

John Stammen, CEO of Convr, frames the issue directly: the industry has moved past debating whether AI belongs in underwriting. The question now is how carriers implement it. Deployment without strategic architecture produces activity, not advantage.

What Underwriters Actually Want

The survey’s findings on what would most help underwriting teams are revealing precisely because they do not simply ask for more AI.

The top three responses were:

• AI tool training — cited by 47.4%
• Pre-screened and enriched submissions — cited by 46.9%
• Simpler access to data — cited by 45.5%

Read together, these responses describe a demand for usable AI rather than more AI. Practitioners want tools they understand, data they can trust, and workflows that reduce rather than multiply complexity. The appetite is not for additional capability in isolation — it is for capability embedded in a coherent operating model.

This distinction matters for anyone evaluating or building AI tools for the insurance sector. Feature richness is not the differentiator. Workflow fit, data quality, and training support are.

The Hybrid Stack Problem

The 63% hybrid environment figure deserves more attention than it typically receives in AI adoption discussions.

A hybrid stack — legacy core systems running alongside cloud-based AI tools — is not a temporary state for most insurers. It is the medium-term reality. Legacy systems carry decades of policy data, regulatory compliance logic, and institutional process. They cannot be replaced quickly or cheaply.

This means AI tools in commercial P&C underwriting must be designed for integration complexity, not clean-slate deployment. Tools that assume modern data infrastructure will consistently underperform in environments where submission data arrives from multiple sources in inconsistent formats. The integration challenge is not a bug in the adoption story — it is a central feature of it.

Where Competitive Advantage Will Emerge

Convr’s interpretation of the survey data points toward a clear thesis: the next phase of AI adoption in commercial P&C will not be won by the carriers that move fastest. It will be won by those that combine automation with strategic clarity, deliberate training, and a more disciplined operating model.

This is a meaningful reframe. Speed was the competitive variable in 2024 and 2025 — who could get tools into production first. The variable shifting now is execution quality: who can demonstrate that their AI investments are producing measurable, repeatable improvements in underwriting accuracy, submission throughput, and loss ratio performance.

For carriers still in the cautiously open majority, the practical implication is that the window for strategic differentiation is open but not indefinitely so. Early movers have tools deployed. The firms that build the strategy layer on top of those tools next will be the ones that convert deployment into durable advantage.

What This Means for the AI Tools Ecosystem

For tool builders and evaluators watching this space, the survey data offers a clear signal about where product and go-to-market gaps exist.

Training and onboarding are not afterthoughts — they are primary purchase criteria. Data enrichment and submission pre-screening represent high-value integration opportunities. And tools that can operate effectively within hybrid legacy-cloud environments will find a larger addressable market than those requiring clean infrastructure.

The demand is real and growing. 65.9% of insurers plan to introduce additional AI tools in 2026. But the buyers are increasingly sophisticated about what they need beyond the tool itself.

Closing Reflection

The commercial P&C underwriting sector has crossed the threshold from AI experimentation to AI deployment. That is a meaningful milestone. But deployment without strategy is a cost center, not a competitive asset.

The 20.4% confidence figure is not a failure metric — it is an honest diagnostic. It tells the industry where the real work begins. Closing the gap between tool adoption and strategic confidence is the underwriting challenge of the next two years, and the carriers that treat it with the same rigor they apply to risk will be the ones that define what AI-enabled underwriting actually looks like at scale.