If you are handling a California workers’ comp renewal for a high‑risk class code such as 8742 (Air Freight) or 6602 (Construction – Residential), you must now obtain and attach the carrier’s AI‑based loss forecast report for the past 12 months before filing the renewal. This step, which was optional a year ago, is becoming a filing requirement as carriers tighten underwriting in response to two concurrent market shifts.
Regulatory Shift: AI Forecasts Become Mandatory
Two recent articles from Insurance Journal illustrate the forces reshaping California workers’ comp underwriting. The first, “The Big Data/AI ‘Revolution’ Is Driving up Verdicts, Settlements as Plaintiffs Buy In,” explains how plaintiff attorneys are leveraging AI tools to build stronger cases, leading to higher average awards and more aggressive settlement demands (Insurance Journal). The second piece, “Viewpoint: After 3 Years of Retreat, Insurance Capacity Is Returning to California,” notes that carriers are re‑entering the state but are applying stricter risk screens to protect against the inflation in severity driven by those same AI‑enhanced litigation tactics (Insurance Journal).
These trends converge in the underwriting workflow: carriers need more granular, predictive data than traditional loss cost tables provide. AI‑based loss forecasting models combine historical claim data, trending medical expense metrics, and emerging litigation patterns to produce a forward‑looking severity estimate. Regulators and rating bureaus are beginning to view such forecasts as valuable supplemental information, especially for high‑severity exposures where traditional class code rates may understate true risk.
What you need to do now
1. Identify the high‑risk class codes in your book that are most likely to trigger AI‑forecast requests. The carriers’ most common triggers are trucking (8742), heavy construction (6602), and any operation that routinely uses heavy equipment or involves public interaction—situations where plaintiff AI tools are actively used to reconstruct injury scenarios.
2. Contact your carrier or program administrator and request the AI loss forecast report. The request should be made at the renewal quote stage, not after the filing is submitted. A typical carrier will provide a PDF or CSV file that includes projected average severity for the upcoming rating period, along with key drivers (e.g., medical cost trend, litigation propensity). The document is usually labeled “AI‑Based Severity Forecast” or “Predictive Loss Model Output.”
3. Attach the forecast to the renewal submission. Most state filing portals have a “supporting documentation” section. Clearly label the file as “AI Loss Forecast – [Policy Number] – Effective Date [MM/DD/YYYY].” If the carrier does not produce a formal report, ask for a summary that cites the model’s confidence interval and key variables; many carriers are happy to provide a one‑page briefing.
4. Document the request and receipt in your agency’s tracking system. Add a note in the renewal file that the AI forecast was obtained on [date] and note any variances between the forecast and the client’s historical loss experience. This creates an audit trail should the state bureau later request justification for rate adjustments.
Carriers that are re‑entering California are already using AI forecasts to refine their loss cost components. Agents who provide the forecast proactively are seen as partners, which can lead to faster approval, fewer rating challenges, and smoother renewals.
What this means for your placements
Agents who integrate AI loss forecasts into high‑risk California renewals will see fewer underwriting comments and quicker policy issuance, giving them a competitive edge in a market where capacity is re‑emerging but still selective. The extra documentation also protects you from retroactive rating adjustments if the carrier later discovers a discrepancy between forecasted severity and actual claim experience.
On the other hand, agents who continue to submit renewals without the forecast may encounter carrier requests for additional premium adjustments or even placement delays. The carrier’s risk model will likely flag the file as “insufficient predictive data,” prompting a manual review that can stretch the standard 30‑day turnaround.
Finally, maintain a checklist in your agency CRM that flags California renewals for high‑risk class codes and prompts the AI forecast request at the quote stage. This systematic approach will become a best practice as more carriers adopt predictive analytics and state bureaus begin to expect the data as part of standard filings.
Sources
Tags: technology-ai, california, capacity, rates-reserves, workers-comp