One adjuster described it on Glassdoor like this: AI misclassifies the claim, hallucinates facts in the summary, and you’re the one who passes it along before catching the error. The customer doesn’t know a bot was involved. They just know you got it wrong. That adjuster — Ahmad Jackson, profiled by WIRED — eventually left his employer for a different carrier. According to Glassdoor data covering June 2025 through May 2026, 98% of claims adjuster reviews mentioning AI are negative. The cross-industry average is 53%. This isn’t abstract AI anxiety. It’s a specific workforce absorbing AI’s failures while watching their profession disappear around them — much like the humanoid robot displacement already hitting other service roles.
The Numbers Don’t Lie
The structural threat underneath the sentiment story is already showing up in hiring data and automation forecasts.
The Bureau of Labor Statistics projects roughly 18,200 claims adjuster jobs gone over 2024–2034, with AI productivity gains cited as a central driver. Entry-level postings dropped 50% in a single year, according to Glassdoor and Indeed — a stark contrast to efforts like Meta’s job guarantee aimed at cushioning workforce disruption. Independent models combining BLS data with automation-risk research put claims adjusting at approximately 72% probability of high automation — particularly for document review, settlement calculation, and accident report triage. Here’s what that looks like on the ground:
- Chatbots handle first notice of loss, misclassify complex claims, and reroute the wreckage back to humans — without flagging that anything went wrong.
- LLMs summarizing hundreds of pages of medical records can hallucinate from a single smudged PDF, cascading into incorrect payouts.
- Automated pipelines send payouts in seconds, with accountability notably absent when the number is wrong.
- Management tracks individual AI adoption rates, layering surveillance pressure onto existing workload pressure.
“AI is just a tool — it should never be given the keys.” — Geoffrey Conrad, claims executive
Lemonade’s Vision vs. the View from the Trenches
Two versions of the same story, depending entirely on where you’re standing.
Lemonade’s chatbot AI Jim now handles 96% of first notices of loss. About 55% of claims are fully automated. Spokesperson Paul Staats acknowledges AI will “impact many jobs and threaten many incumbents,” but argues automation frees humans to focus “empathy, care, and expertise on the most complex claims” — a framing examined more broadly in discussions of AI productivity. State Farm, meanwhile, positions its “human plus digital” approach as a deliberate differentiator. Both framings assume the handoff between machine and human works cleanly.
Former adjuster and consultant Sandy Avina, speaking to WIRED, says it doesn’t. Customers rarely know AI was involved, so when a hallucinated medical summary produces a wrong payout, the adjuster takes the hit. Glassdoor economist Chris Martin, quoted in Forbes, calls the 98% figure “shockingly” high and ties it directly to workers being handed subpar tools with no recourse. Workplace AI mentions have surged 240% since 2019 — while overall positivity toward AI at work collapsed from 81% to 43%. Claims adjusters didn’t start that trend. They just pushed it to its logical extreme.
“People need a human who checks on their safety and well-being after a total loss — not a photo upload followed by an automated estimate.”
After a house fire or a totaled car, someone is having one of the worst days of their life. The question was never whether AI could process the claim. It’s who absorbs the cost when it gets it wrong. Right now, that’s the adjuster.
