Sonant says insurers should automate highest-volume workflows first

The fastest automation wins in P&C are hiding in plain sight

Sonant’s central advice lands with unusual clarity: do not try to automate insurance operations all at once. In a P&C environment, the highest-return targets are the workflows that repeat constantly, follow predictable rules, and create the most friction when they slow down, especially quote intake, servicing, renewal outreach, FNOL, post-bind work, and outbound communications.

That sequencing matters because insurance operations are rarely broken in the same way everywhere. A rigid tool can struggle when a workflow is packed with edge cases, handoffs, approvals, and exceptions. Sonant’s framing turns automation into an operating-model decision, not a software shopping exercise, and that is the right lens for carriers and agencies that want speed without adding operational risk.

Why volume and standardization come first

The easiest mistake in automation is to start with the hardest process. Sonant’s guide pushes the opposite logic: begin with the workflows that are both high-volume and standardized, then build toward more exception-heavy work only after the underlying process is clean enough to support automation.

In practice, that means looking for tasks where the same forms, fields, decisions, and notifications recur every day. When a process is repetitive enough, automation can improve speed, consistency, and customer experience at the same time, which is why the business case in P&C is bigger than simple labor savings. It is also why modernization programs should treat automation as a way to redesign work, not merely accelerate the old way of doing it.

FNOL is powerful, but only when intake is disciplined

First notice of loss is one of the most obvious places to pursue automation, but it is not automatically ready for deep automation just because it is important. FNOL works best when intake is structured, data is reliable, and routing rules are clear. If incoming information is messy or the process depends on too many manual clarifications, a new workflow engine will just move the chaos faster.

That is where process maturity matters. Some FNOL flows are ready now for visibility, exception routing, auditability, and human-in-the-loop decisioning. Others need cleaner data, fewer intake variations, and simpler handoffs before they can be trusted in production. The difference is not subtle: a well-designed FNOL workflow can shorten response time and reduce friction, while a poorly designed one can magnify errors.

Document intake and outbound communications are low-risk places to start

Document intake is one of the strongest early candidates for automation because it usually combines high volume with repeatable patterns. The same goes for outbound communications, where reminders, status updates, and renewal notices often follow structured triggers. These are the kinds of workflows where workflow automation can coordinate multi-step processes with conditional logic instead of relying on staff to move items from one queue to the next.

That is also where the distinction between workflow automation and robotic process automation becomes useful. Sonant describes RPA as brittle, rule-following task execution. Workflow automation, by contrast, manages a process end to end, including handoff controls and exception routing. For insurance leaders, that distinction matters because a document intake stream that simply copies data from one place to another is not enough if the business still needs auditability and a clear path for exceptions.

Servicing and renewals reward consistency more than cleverness

Policy servicing and renewal outreach may not sound glamorous, but they are exactly the kinds of workflows where automation can pay back quickly. Servicing work often involves repetitive status changes, routine customer requests, and information updates that only become expensive when they are handled inconsistently. Renewal outreach is similar: it is a recurring process with predictable timing, making it a natural candidate for automated reminders, routing, and follow-up.

These flows also show why workflow maturity matters. If servicing data lives in too many systems or if renewal logic is full of manual overrides, automation will not fix the underlying disorder. The best results come when teams first clarify the process, then automate the repeatable parts, and keep humans in the loop where judgment still matters.

RPA is not the same as an operating model

Insurers often talk about automation as if every tool solves the same problem. Sonant’s guide draws a sharper line. RPA is best understood as task execution at the edge of a process, while workflow automation coordinates the whole process with conditional logic, visibility, and control.

That distinction is more than technical nuance. It changes how leaders should design governance from day one. If the goal is to improve claims intake, quote processing, or renewal handling, the question is not whether a bot can click through a screen. The question is whether the process can be observed, routed, audited, and adapted when the exception arrives.

The regulatory backdrop is getting more active

This is not happening in a vacuum. The National Association of Insurance Commissioners established its Big Data and Artificial Intelligence (H) Working Group in 2019 to study AI’s use in insurance and its implications for consumer protection, privacy, market dynamics, and the state-based regulatory framework. The NAIC also says AI is already used in underwriting, pricing, customer service, claims handling, marketing, and fraud detection, which makes workflow automation an operations and governance issue, not just an IT initiative.

The regulatory agenda is still moving. In 2026, the working group’s charges include monitoring state, federal, and international AI activity and supporting adoption of the Model Bulletin on the Use of AI Systems by Insurers. For automation leaders, that means auditability, model oversight, and human review are not optional extras. They are part of the operating model.

The market is already moving from pilots to production

A lot of insurers have already tasted automation, but many are still living in pilot territory. UiPath says that by the end of 2025, most insurers had already deployed generative AI somewhere in the business, yet many were still stuck in isolated pilots and demos. That gap is exactly where Sonant’s advice becomes valuable: production-grade automation belongs in the most repeatable workflows first, not in the most complicated ones.

Accenture’s claims research points in the same direction. It says AI and generative AI can reduce claims cycle time from days to minutes, and that many claims meet the conditions for straight-through processing. That makes claims one of the clearest examples of where standardization can unlock speed, but only if the process has been designed to support it.

The financial case is bigger than efficiency

The business case for automation has moved well beyond cost cutting. NAIC’s 2025 P&C annual report says U.S. P&C underwriting income rose by more than $40 billion versus the prior year, and policyholders’ surplus reached a new high of $1.27 trillion at December 31, 2025. That combination of stronger balance-sheet capacity and continuing industry pressure makes operational discipline even more important in 2026.

The broader industry data backs that up. NAIC’s 2025 Health AI/ML Survey received responses from 93 companies and found that 84% of health insurers were using, planning, or exploring AI/ML. Accenture’s June 2025 research adds another layer, saying outperformers can achieve improved premium revenues of 8.1 percentage points and reduced expense ratios of 2.6 percentage points more than peers. The message is clear: automation is shaping operating performance, not just back-office efficiency.

A practical roadmap for P&C leaders

The most durable automation programs in P&C start with a simple test: is this workflow high-volume, standardized, and ready for clean handoffs? If the answer is yes, it belongs near the front of the queue. If the answer is no, the work may need redesign, better intake discipline, cleaner data, or simpler routing before automation can succeed.

That is the real lesson in Sonant’s guide. The winning strategy is not to automate everything that can be automated. It is to automate the workflows that can absorb automation well, preserve human judgment where exceptions still dominate, and build an operating model that improves speed without creating new fragility.