A Glassdoor analysis found that 98% of claims-adjuster reviews mentioning AI were negative, the highest criticism rate among all occupations studied. The findings, reported September 1, 2026, come as insurers push automation deeper into the claims process, from first notice of loss to damage estimation and document review.
Adjusters say the tools are creating errors that workers then have to fix. Company leaders keep rolling out AI systems that don't work well, according to the reviews, and the failures are making it harder to serve policyholders.
Where automation goes wrong
Ahmad Jackson saw the problem while working in the claims department at a major insurance company about a year ago. The company began using AI for initial loss reporting, the first stage of a claim when an insurer gathers details about an accident, fire, or other loss. The goal was to move routine cases faster and route difficult claims to people.
Instead, Jackson said the AI often classified claims incorrectly. Adjusters had to transfer files to the right department and deal with faulty claim summaries. "AI is getting things wrong," Jackson told Wired. "And it's implementing more work onto the adjusters."
Jackson left and joined another carrier. He still sees uses for AI in basic administrative tasks, such as extending a rental-car reservation for a claimant. But he does not think it should be trusted with work that requires judgment or detailed knowledge of a claim.
Insurers double down on automation
Insurance companies are expanding AI across claims operations. Customers may file a claim through a chatbot instead of speaking with a representative. Insurers use software to review photos or videos of damaged homes and vehicles, estimate repair costs, and process uploaded receipts and records. AI tools can also summarize medical files that run hundreds of pages.
Lemonade has built its claims systems around automation. The company said its AI chatbot Jim handled initial reports 96% of the time by the end of last year, and automation handled about 55% of all claims. Paul Staats, a Lemonade spokesperson, said automation can allow employees to focus their "empathy, care, and expertise on the most complex claims." He also said the Glassdoor findings should be taken seriously across the industry.
For professionals looking to understand how these systems are being deployed, AI for Insurance training covers claims processing and underwriting applications. Coursework on AI Agents & Automation also addresses the oversight questions raised by adjusters in the field.
Job pressure compounds the frustration
The rollout comes as claims-adjuster jobs are already under pressure. The Bureau of Labor Statistics said in 2024 that it expected the number of claims adjusters in the United States to fall by 18,900 over the next decade, a 5% decline. Employment in the sector dropped 21% between May 2025 and May 2026, according to BLS data. Glassdoor also found that entry-level job postings had fallen 50% since 2025.
Chris Martin, Glassdoor's senior economist, said AI criticism often rises when workers worry that their jobs are disappearing or when they believe a company is forcing unreliable technology onto workers and customers. "There's no jobs, there's no good job prospects," Martin said.
The concern is not limited to job losses. Sandy Avina, a former claims adjuster and insurance consultant, said AI can make costly mistakes when it reads incomplete or poor-quality documents. A smudge on a document from an attorney, for example, can lead to a hallucination. If an AI-generated summary leaves out an important detail from a medical report, it can result in an inaccurate payout. In many cases, customers do not know that AI caused the problem. They may assume the adjuster made the mistake.
Geoffrey Conrad, a claims executive in Mobile, Alabama, said adjusters are tired of the steady introduction of new AI systems. "There's an AI fatigue," he said. "We're pretty much exhausted as far as the amount of AI being shoved down our throats."
Conrad said AI can be useful, but it should not be allowed to make final decisions on its own. "AI is just a tool," he said. "It should never be given the keys." His view is shaped by his own experience as a homeowner. Before he entered the insurance business, Conrad lost his house in a fire. He said that during a crisis, policyholders need more than a computer-generated estimate based on photos. "I needed somebody to make sure that I'm OK, my family's OK, that we're safe," Conrad said.
Why this matters for insurance professionals
The Glassdoor data signals a workforce that is skeptical of AI rollouts, not resistant to all technology. Adjusters quoted in the report distinguish between administrative automation they can live with and judgment-based tasks where errors carry real financial and human costs. If you work in claims, the practical takeaway is to document AI-related errors and rework in your own workflow. That record matters for two reasons: it protects you when a customer assumes you made the mistake, and it gives your leadership concrete data on where the tools are failing rather than anecdotal frustration.
Your membership also unlocks:
