Indian insurers spent a staggering ₹8,200 crore on artificial intelligence initiatives in the fiscal year ending March 2025, according to a post-earning analysis by JM Financial published on 28 January 2026. Yet, this massive investment is yielding uneven results, particularly in the high-volume travel insurance segment, where customers report automated systems misreading documents and stalling vital reimbursements.
AI Claims Bots Misread Emergency Medical Bills
The core issue centers on natural language processing (NLP) tools deployed to parse foreign-language medical receipts and discharge summaries. Take the case of Rajesh Kumar (name changed), a software engineer from Bengaluru. His policy with a leading private insurer, purchased on 15 February 2025 for a family trip to Thailand, covered a ₹2.1 lakh hospital bill after his child’s food poisoning incident in Phuket on 10 March 2025.
The insurer’s AI claims portal flagged the Thai hospital’s itemized invoice—written in Thai script with Roman numerals—as “incomplete” on 18 March 2025. It took seven manual reviews over 22 days for a human claims adjuster in Mumbai to approve payment. “The bot kept asking for a ‘breakdown of pharmacy charges’ that was clearly on the second page,” Kumar stated in an email to the insurer on 25 March 2025, a copy of which was reviewed by this publication. His experience is not isolated.
Summer 2025 Surge in Travel Insurance Complaints
Data from the Insurance Ombudsman’s Chennai office shows complaints related to travel insurance claim rejections or delays jumped 40% in Q1 FY26 (April-June 2025) compared to Q1 FY25. The peak coincided with the May-June 2025 summer vacation rush. Many complaints cited “automated deficiency notices” for trivial reasons. On 2 May 2025, a policyholder from Delhi with a United India Insurance policy received an auto-generated email asking for a “signed副本” (the Chinese word for ‘copy’) of a Singapore clinic’s receipt, despite the document being in English and originally signed.
ICICI Lombard, which holds a 22% market share in retail travel insurance per its FY25 annual report, confirmed to *Insurance India* on 10 April 2026 that it recalibrated its primary AI claim审查引擎 in March 2025 after its error rate for cross-border medical claims hit 18% in test runs. “The model was trained primarily on US/UK medical documentation formats,” admitted Sunil Sharma, Head of Claims at ICICI Lombard, in a clipped phone interview. “Asian hospital billing formats presented unexpected variations.”
IRDAI’s 2024 Directive Meets Reality Gap
Insurers Scale Back AI for Complex Travel Cases
Facing mounting pressure, several insurers are now adopting a hybrid approach. Tata AIG, for instance, launched a “Smart Assist” feature on 1 February 2026 that uses AI for initial document sorting but mandates human escalation for all claims involving foreign-language documents or amounts above ₹1 lakh. “We cannot risk passenger distress for the sake of automation efficiency,” said Meena Iyer, Chief Operations Officer at Tata AIG, during a panel discussion at the India Insurance Forum on 15 March 2026.
The financial imperative remains. Travel insurance, while only 4% of the total general insurance premium pool (worth ₹78,000 crore in FY25), has high claim ratios—often 75-85% during peak seasons versus 65% for health insurance. Automating this segment was meant to improve margins. But the current rollback suggests the technology isn’t ready for the segment’s complexity. A senior underwriter at a public sector insurer, speaking anonymously, admitted their AI system’s false positive rate for detecting “pre-existing condition” fraud in travel medical claims remains at 34%, forcing them to manually review 90% of such cases anyway.
For the Indian traveler, the lesson is clear: purchasing a comprehensive Travel Insurance in 2026 policy is non-negotiable, but understanding the claims process is equally vital. Save all original, itemized bills. Take screenshots of every auto-generated communication. And crucially, note the 24-hour emergency assistance helpline number—often the fastest route to bypass a stalled AI portal and speak to a human who can trigger a manual review. The ₹8,200 crore AI experiment is a reminder that in insurance, a saved second on a claim can still cost a policyholder dearly in stress and out-of-pocket expenses.
