The FNOL call is the most time-critical touchpoint in the entire claims lifecycle. Delays at first notice correlate directly with fraud risk, evidence loss, and policyholder dissatisfaction. Traditional call centre FNOL handling involves 15–25 minute average handle times, queue waits of 8–20 minutes during peak periods (monsoon accidents, post-festive hospitalisations), and data capture errors that cause downstream processing delays.

An AI voice agent eliminates these friction points. The moment a policyholder calls — at 2 AM after a road accident or on a Sunday afternoon when a pipe bursts — the agent picks up immediately, authenticates via policy number + registered mobile, and initiates a structured intake flow. Claim type determines the branching path: motor accident triggers FIR advisory + spot survey scheduling; health hospitalisation triggers TPA pre-authorisation; property damage triggers surveyor appointment; life triggers nominee verification and death certificate checklist.

The agent integrates with the insurer's Policy Administration System (PAS) to verify coverage, confirm IDV for motor, confirm sum insured for health/property, and check policy lapse status before registering the claim. A unique Claim Reference Number (CRN) is generated in real time and sent to the policyholder via SMS and WhatsApp. The document checklist — tailored to claim type — is dispatched simultaneously.

Production data from insurers using AI FNOL shows 78–85% of first-notice calls fully resolved without human handoff, claim registration TAT reduced from 4–6 hours to under 8 minutes, and policyholder NPS improvement of 22–30 points attributed directly to first-call resolution.

Motor insurance is the highest-volume FNOL segment in India, with IRDAI recording 2.3 crore motor claims annually. Peak periods — monsoon (July–September), Diwali weekend, year-end — flood call centres with simultaneous FNOL registrations, creating 30–90 minute wait times precisely when policyholders need immediate guidance.

The AI handles motor FNOL across three sub-types: Own Damage (OD), Third-Party Liability (TP), and Comprehensive. For OD claims, the agent captures: accident date, time, and location (with pin-code for nearest cashless garage look-up), registration number, odometer reading, description of damage, and whether the vehicle is driveable. For third-party claims, FIR is mandatory regardless of claim amount — the agent advises this immediately and provides the nearest police station number if required.

For own damage claims above Rs 2 lakh, FIR is required per the insurer's standard terms. The agent confirms this threshold, advises the policyholder not to move the vehicle until the spot surveyor arrives (scheduled within 24 hours of FNOL for total loss/major damage), and connects to the insurer's 24×7 roadside assistance if applicable.

Cashless claim eligibility is confirmed in real time by querying the PAS: if the vehicle is in a cashless garage network city, the agent provides the 3 nearest network garages by proximity. For reimbursement cases (non-network repair), the agent advises on the advance deposit process and original invoice retention. Drunk driving, overloading, and racing endorsement exclusions are checked automatically — if flagged, the agent notes the condition without denying the claim (denial authority rests with the human surveyor).

Health insurance FNOL has two distinct tracks: planned (elective) procedures requiring advance pre-authorisation, and emergency admissions requiring rapid post-admission intimation. Failure to intimate within the insurer-specified window (typically 24 hours for emergency, 48–72 hours pre-admission for planned) is the single most common cause of cashless claim rejection — affecting an estimated 12–18% of health claims per IRDAI data.

The AI voice agent eliminates this lapse by proactively prompting intimation. For policyholders admitted to hospital, the agent walks through: policy number verification, patient name + relationship to insured, hospital name and city (checked against TPA network in real time), diagnosis/procedure, admission date and expected discharge, and treating doctor's name. The pre-auth request is electronically transmitted to the relevant TPA (Medi Assist, MDIndia, Paramount, Vipul Medcorp, Health India TPA, or insurer in-house TPA) via API.

For cashless claims at network hospitals, the agent advises the policyholder to present the e-health card at the hospital TPA desk, confirms the pre-auth reference number, and sets a follow-up call for 4 hours after request submission to confirm TPA approval status. For non-network hospitals (reimbursement track), the agent advises on bill collection requirements: all original bills, discharge summary, investigation reports, and doctor prescriptions — none can be submitted as copies.

Room rent sub-limits (1% of sum insured per IRDAI standard) and co-payment clauses (common in senior citizen and zone-based policies) are read from the PAS and disclosed to the policyholder upfront, preventing bill shock at settlement. OPD claims (where covered) follow a separate simplified intimation flow.

Property and home insurance FNOL requires immediate damage preservation advice — policyholders who repair or clean up before surveyor inspection risk partial or full claim rejection. The AI agent's first priority after confirming coverage is to advise the policyholder to: photograph all damage from multiple angles, retain damaged items where safe to do so, and not commence repairs without surveyor sign-off except for emergency safety measures (e.g., covering a broken window to prevent further rain damage).

Claim type determines the FIR requirement. Theft and burglary claims require an FIR from the local police station — the agent provides this guidance and the nearest station number if available. Fire claims may require a fire brigade certificate (obtained from the local fire station), especially for total loss claims. Flood/earthquake claims trigger the IRDAI catastrophe claims processing protocol, which may activate bulk surveyor deployment by the insurer.

Surveyor appointment is scheduled based on insurer tie-ups with licensed surveyors under the Insurance Surveyors and Loss Assessors (Licensing, Professional Requirements and Code of Conduct) Regulations 2015. For claims above Rs 50,000, an independent licensed surveyor is mandatory. The AI books a 2-hour appointment window within 48 hours of FNOL and sends confirmation to both the policyholder and the surveyor via WhatsApp.

Document checklist dispatched includes: photographs of damage (minimum 6 angles), FIR copy (for theft), fire brigade report (for fire), property purchase deed or lease agreement, previous claim history, and repair cost estimates from 2 authorised vendors. For rental property claims, the tenant-landlord insurance arrangement must be confirmed before proceeding.

Life insurance death claim intimation is handled with heightened sensitivity protocols. The AI agent is configured to recognise the emotional context of a death notification call and uses a subdued, empathetic call flow — slower pace, shorter sentences, more explicit confirmation of each step, and an immediate offer to transfer to a human relationship manager if the nominee prefers personal handling.

The agent captures: full name of the deceased, policy number(s) — many families discover multiple policies during estate settlement, date of death, cause of death (natural/accidental/suicide — relevant to claim processing tracks), hospital or location of death, and name and mobile number of the primary claimant/nominee. Policy bond number and nominee details are verified against the PAS.

For early claims (death within 3 years of issuance), the insurer is required to conduct an investigation. The AI advises this upfront — the investigation is standard, not an accusation — and the timeline is typically 30–90 days from last document submission. Medical records, hospital discharge summary (if hospitalised), and attending physician's certificate are added to the checklist.

For accidental death claims, the AI adds: FIR/police report, post-mortem report, newspaper clipping (for public accident), and viscera report if available. Nominee who is a minor triggers an additional advisory about appointing a natural guardian or legal guardian with court order for claim receipt.

IRDAI mandates life claim settlement within 30 days from receipt of all documents for non-investigated claims, and within 45 days for investigated claims. The agent quotes these timelines explicitly to manage expectations and reduce follow-up inquiry volume.

IRDAI's Protection of Policyholders' Interests Regulations 2017 (amended 2024) defines strict settlement timelines. Breach of these timelines triggers mandatory interest payment to the policyholder at 2% above bank rate from the date of intimation — a financial penalty that insurers actively monitor and avoid.

The AI voice agent enforces timeline compliance through three mechanisms. First, at FNOL it quotes the expected settlement timeline explicitly — this sets policyholder expectations and reduces follow-up calls. Second, it sends proactive status updates at key milestones: document checklist dispatched (Day 0), document submission confirmed (Day X), survey completed (Day Y), and settlement processed (Day Z). Third, it runs an internal SLA monitoring loop — when a claim is approaching its deadline with outstanding documents or survey completion, the agent places a proactive call to the policyholder (for pending docs) or alerts the claims team (for internal bottlenecks).

For health cashless claims, the timeline differs: pre-auth approval must be given within 1 hour (emergency) or 2 hours (planned) per the network hospital agreement. Delay in pre-auth triggers an automated escalation to the TPA medical team. Final cashless settlement with the hospital must occur within 7 working days of discharge document receipt.

IRDAI's Integrated Grievance Management System (IGMS) is the formal escalation channel. Policyholders who do not receive resolution within IRDAI timelines can file on IGMS. The AI proactively advises IGMS rights at the 30-day mark if a claim is still pending — this transparency reduces hostile escalations because policyholders feel informed rather than abandoned.

Insurance fraud costs the Indian insurance industry an estimated Rs 30,000–45,000 crore annually per GI Council estimates, with motor and health segments accounting for the majority of fraudulent claims. FNOL is the first and most cost-effective intervention point — a claim flagged at intake costs a fraction of a fraudulent claim that passes through survey and settlement.

The AI applies a multi-layer fraud detection protocol at FNOL. Layer 1 is policy-level checks: claim within 30 days of issuance (early claim flag), multiple claims in 12 months, policy purchased with intent to claim (short policy tenure + high sum insured + immediate claim). Layer 2 is incident-pattern analysis: incident date/time/location cross-referenced against known fraud-prone corridors (e.g., specific highway stretches with high staged-accident frequency), multiple vehicles involved with same insurer, or incident reported to police 48+ hours after occurrence.

Layer 3 is identity verification: caller's mobile number against policy-registered mobile, nominee name exact match, cross-check with UIDAI Aadhaar verification if high-value claim. Layer 4 is claim history: the AI queries the Insurance Information Bureau (IIB) database — a shared claims history registry — for prior claims across insurers.

Flagged claims receive a normal FNOL confirmation (CRN issued, checklist dispatched) to avoid alerting the fraudster. Internally, a SIU (Special Investigation Unit) flag is set in the Claims Management System (CMS) with the specific fraud indicators, and the assigned surveyor receives an advisory before conducting the spot inspection. This covert routing reduces fraudulent claim payments without affecting legitimate claim processing speed.

Document incompleteness is the leading cause of claim settlement delay in India — IRDAI data shows 38–45% of claims experience delay due to incomplete or incorrect documentation at first submission. The AI's document automation directly attacks this bottleneck at two points: immediate dispatch at FNOL and systematic follow-up until all documents are received.

Document checklists are dynamically generated based on claim type, sum insured, cause of event, and policy-specific endorsements. Motor OD checklist: RC copy, driving licence, insurance policy copy, FIR (if applicable), repair estimate from cashless garage or authorised repairer, photographs of damage. Motor TP checklist: FIR, court summons copy (if litigation), Claim Tribunal notice. Health reimbursement checklist: hospital discharge summary, all original bills (OPD + IPD), investigation reports, doctor's prescriptions, pharmacy bills, NEFT mandate. Life death claim checklist: death certificate (original + 2 attested copies), claimant's statement Form A, physician's certificate Form B, original policy bond, NEFT details.

The WhatsApp message includes a document-by-document checklist with individual upload links. Each link routes to the document type in the insurer's portal with the CRN pre-filled — removing the friction of re-entering claim details for each document. Upload confirmation triggers an automated acknowledgement and updates the CMS in real time.

Day 3, Day 7, and Day 14 reminders are placed via voice call (not just SMS/WhatsApp) — voice follow-up achieves 3.2x higher document submission rate than digital-only reminders per production data. Each reminder specifies exactly which documents remain outstanding, not a generic 'documents pending' message.

India's health insurance ecosystem operates through a three-party structure: insurer (underwriter), TPA (Third Party Administrator — licensed claims processor), and network hospital. The AI FNOL agent must navigate this structure fluidly — identifying which TPA manages the specific policy, what the TPA's network hospital list covers, and whether the admission is at a network or non-network facility.

TPA integration works via HL7/API connectivity or, where TPAs lack modern APIs, via structured email automation with standardised pre-auth request formats. The AI captures the minimum required data fields for pre-auth: patient name, age, diagnosis/procedure code (ICD-10 or plain language), hospital name and address, admission date, expected duration, and treating doctor's registration number (MCI/NMC). This data is transmitted within the call, not after — eliminating the 2–4 hour gap between FNOL call and pre-auth initiation in manual processes.

For group health policies (employer-employee schemes), the AI handles additional complexity: verifying employee status (active vs terminated affects claim eligibility), confirming the sub-limit structure (maternity, OPD, room rent), and identifying whether the policy has a family floater or individual coverage structure. Corporate HR portals may be queried via API to confirm current employment status.

Post-admission, the AI places a Day 2 follow-up call to confirm TPA approval has been received at the hospital, troubleshoot any pre-auth rejection (most common reasons: non-covered procedure, sub-limit exceeded, exclusion period), and initiate an appeal process if needed. TPA rejection appeal success rate improves 28–34% when the appeal includes specific clinical documentation guided by the AI checklist.

Surveyor coordination is a critical post-FNOL workflow that directly affects claim settlement TAT and policyholder satisfaction. Delays in surveyor appointment are the second most common cause of settlement delay (after document incompleteness), accounting for 22–28% of delayed motor claims per IIB data.

The AI's surveyor scheduling module queries the insurer's empanelled surveyor database by: pin-code of incident (for motor/property), claim value (determines surveyor tier — licensed surveyor for claims above Rs 50,000, senior surveyor for above Rs 5 lakh), and surveyor workload (real-time slot availability). For motor total loss cases, the AI adds the requirement for a motor engineer surveyor with IDV valuation credentials.

Surveyor appointment details are pushed to the Claims Management System (CMS) simultaneously with policyholder notification. The surveyor receives the assignment via the insurer's surveyor app with: claim details, policyholder address and contact, photos submitted at FNOL, and any fraud flags (anonymised as 'additional verification required' to the surveyor, with full details visible to the SIU).

A Day-prior reminder call is placed to the policyholder to confirm the appointment time, ensure the vehicle/property is accessible, and remind them to have all relevant documents ready for the surveyor (RC, DL, repair estimate for motor; property purchase documents and photographs for property). If the surveyor reschedules, the AI immediately notifies the policyholder and offers the next available slot within the IRDAI timeline window.

For catastrophe events (cyclone, flood, earthquake), the AI activates bulk surveyor deployment mode — prioritising claims by damage severity (total loss first) and coordinating 20–30 surveyor appointments simultaneously per day.

Two-wheelers account for approximately 65% of motor insurance policies in India by volume, making two-wheeler FNOL the highest-frequency motor claim type. The AI handles two-wheeler FNOL with specific adaptations.

IDV for two-wheelers depreciates rapidly — a 3-year-old scooter may have an IDV of Rs 35,000–55,000. Total-loss claims (repair cost exceeds 75% of IDV) are therefore common at relatively modest damage levels. The AI confirms the IDV from the PAS and advises whether the claim is likely to be settled as total loss vs repair-and-return — this sets policyholder expectations immediately.

For theft claims (extremely common for two-wheelers), the AI adds: police FIR (mandatory), RTO 'Non-Traceable' certificate (issued after 90 days of FIR if vehicle not recovered), Form 35 (hypothecation cancellation if under finance), NOC from financer (if hypothecated), and original RC book. The claim is provisionally registered but settlement requires the Non-Traceable certificate — the AI advises the 90-day timeline upfront to prevent repeated follow-up calls.

Cashless garage networks for two-wheelers vary by brand — Honda, TVS, Hero, Bajaj, and Suzuki authorised service centres are typically on-network for respective insurer tie-ups. The AI identifies whether the damaged vehicle's brand has a nearby authorised service centre on the cashless network, and if not, routes to reimbursement with repair estimate guidance.

Travel insurance is a high-urgency, time-compressed claim category — medical emergencies abroad require immediate response, often with the policyholder in a hospital or airport in a foreign country. The AI is configured for travel FNOL with multi-language capability and international-number call acceptance.

Medical emergency abroad: the AI immediately transfers to the insurer's 24×7 international assistance company (Allianz Assistance, Europ Assistance, or insurer in-house) after capturing the policyholder's location, nature of emergency, and policy number. The assistance company handles direct hospital guarantee of payment (GOP) — eliminating the need for the policyholder to pay out-of-pocket in most cases. The AI captures this handoff reference number and confirms it with the policyholder.

Trip cancellation claims require specific covered peril documentation: medical certificate (for cancellation due to illness), death certificate of family member, visa rejection letter, or airline/operator cancellation notification. The AI captures the specific reason and dispatches the relevant document checklist. The 30-day filing window from the event date is confirmed explicitly.

Baggage loss or delay claims require the Property Irregularity Report (PIR) from the airline — the AI advises this must be filed before leaving the airport baggage claim area. For baggage delay, the insurer covers essential purchases above a threshold (typically 12–24 hour delay, Rs 2,000–5,000 per day coverage) — the AI advises keeping original purchase receipts.

Passport loss abroad is handled as a travel assistance case: the AI provides the nearest Indian embassy/consulate contact and initiates the emergency passport process advisory alongside the insurance claim.

Insurance technology stacks in India range from modern cloud-based platforms (Majesco SaaS, DuckCreek OnDemand) to legacy on-premise systems (older versions of TCS BaNCS Insurance, FINPLEX, or proprietary in-house platforms built in the 2000s). The AI FNOL agent is built to integrate with all of these via a middleware abstraction layer.

For modern API-first platforms (Majesco, DuckCreek, Sapiens): RESTful API integration with OAuth 2.0 authentication pushes FNOL data in real time. The PAS returns policy status, coverage details, IDV/sum insured, TPA assignment, and network hospital list to the agent within 2–4 seconds — fast enough to use during the live call. CRN is returned from the CMS API and read back to the policyholder.

For legacy platforms without RESTful APIs: the integration layer uses structured data extraction via the insurer's existing web application interface (screen scraping with consent under controlled conditions), or a nightly SFTP batch for policy data sync combined with a lightweight claims registration webhook for real-time CRN generation. This hybrid approach is deployed in 40–50% of production integrations in India's insurance market.

TPA integration follows a similar pattern: Medi Assist and MDIndia have modern APIs; smaller TPAs use email automation with structured pre-auth request templates. For health claims, HL7 FHIR messaging standards are used where TPAs support them for clinical data exchange.

Kallix FNOL integrations completed in insurance: HDFC ERGO (motor + health), Star Health, SBI Life, and three regional general insurers. Typical go-live: 5–7 weeks from kickoff. Post-integration, the AI handles 78–85% of FNOL calls end-to-end without human intervention.

IRDAI's Integrated Grievance Management System (IGMS) — now rebranded as Bima Bharosa under IRDAI's 2024 reforms — is the central grievance portal where policyholders file complaints against insurers. An IGMS complaint triggers a mandatory response within 15 days from the insurer, and unresolved IGMS complaints escalate to the Insurance Ombudsman, which can award up to Rs 30 lakh in compensation without court proceedings.

For insurers, each IGMS complaint has three costs: direct time cost of response (4–8 hours per complaint), regulatory scrutiny cost (high-frequency IGMS filers attract IRDAI inspection risk), and reputational cost (IGMS data is publicly visible in aggregated form). Preventing complaints is worth 10–20x more than resolving them.

The AI's complaint prevention protocol works in three phases. Phase 1 (at FNOL): set accurate expectations for settlement timeline, document collection milestones, and next-action owner. Phase 2 (at Day 20 for 30-day timelines): proactive status update call confirming current claim status and remaining steps — most IGMS complaints are filed because policyholders feel uninformed, not because of actual delay. Phase 3 (at Day 28 if claim is unsettled): explicit advisory that the IRDAI timeline is approaching, offer of internal escalation to a Claim Relationship Manager, and IGMS/Bima Bharosa advisory as the policyholder's right — framed neutrally, not defensively.

For genuine grievances (claim repudiation, settlement shortfall), the AI captures the complaint reason, escalates to the senior claims team, and initiates an internal review workflow — giving the insurer a 15-day window to resolve before IGMS filing.

The cashless vs reimbursement distinction is poorly understood by a large proportion of health insurance policyholders — IRDAI consumer surveys show 42% of first-time claimants are unaware that cashless is only available at network hospitals, leading to bill shock at discharge or botched reimbursement claims due to missing original documents.

The AI addresses this gap at the FNOL call. When a policyholder calls to intimate a hospitalisation, the agent's first action (after policy verification) is to ask for the hospital name and location, then query the TPA network database in real time. The response is definitive: 'Yes, this hospital is on the cashless network — present your e-health card at the TPA desk' or 'This hospital is not on the network — you will need to pay and claim reimbursement.'

For reimbursement track, the agent is explicit about the document retention requirements — particularly that all bills must be originals (pharmacy cash memo, consultant fees receipt, diagnostic reports), and that insurers routinely reject claims where originals have been discarded. The agent advises retaining a photocopy before submitting originals.

For cashless track, the agent advises: present e-health card (or policy card) at the hospital's TPA desk on admission; the TPA desk will submit the pre-auth request on the patient's behalf; do not pay the hospital deposit if the TPA desk confirms pre-auth received; and post-discharge, the insurer settles directly with the hospital within 7 working days.

Co-payment clauses (common in senior citizen policies, zone-based policies, and policies with deductible) are disclosed upfront — the policyholder's share is confirmed from the PAS so there are no surprises at discharge.

Marine and goods-in-transit (GIT) insurance FNOL is predominantly B2B — the policyholder is typically a manufacturing company, trading firm, or logistics provider rather than an individual. The AI is configured for business-to-business FNOL intake with a professional register: structured data capture, reference to bill of lading/LR numbers, and an understanding of commercial terms (CIF, FOB, warehouse-to-warehouse).

Critical at marine FNOL: the carrier liability clause. Under the Carriage of Goods by Sea Act 1925 and Road Carriers Act, the transporter has limited liability (Rs 2,000 per package for road transport under standard conditions). The insurer covers the balance. The AI advises the policyholder to file a subrogation-ready claim: note damage on the delivery receipt, retain the damaged packing, and photograph before unpacking further. This documentation is critical for the insurer to pursue subrogation (recovery from the carrier) — failure to note exceptions on delivery receipt can invalidate subrogation rights.

Survey waiver amount varies by policy. Most marine open covers have a survey waiver for individual consignment claims below Rs 5,000–10,000, settled on self-declaration with photographs. Above this threshold, an independent marine surveyor is appointed (Capt. inspection for vessel damage, cargo surveyor for goods). The AI confirms the waiver threshold from the policy and routes accordingly.

For international cargo (import/export), the AI advises on the International Underwriting Association (IUA) Institute Cargo Clauses (A, B, or C) applicable to the policy — Clause A covers 'all risks', B and C are named peril. This determines whether the loss is covered before proceeding with the FNOL.

PA FNOL calls arrive in two modes: the injured person calling directly (post-discharge for minor injuries) or a family member calling on behalf of someone hospitalised or incapacitated. Kallix AI handles both, adjusting call pacing and vocabulary based on the caller's relationship to the injured.

For road accident PA claims: the AI flags at FNOL that both the motor insurance claim and the PA claim may be relevant — initiating both claim numbers simultaneously. It also captures whether the accident was a third-party fault scenario (another driver at fault), which may enable the PA claimant to also file a compensation claim via MACT.

Injury documentation guidance: the AI provides specific document collection guidance matched to the injury type. For disability claims: hospital discharge summary, disability certificate from a registered medical practitioner, and the treating physician's certificate confirming the nature of permanent disability. For death claims: post-mortem report (mandatory for accidental death PA claims), FIR copy, death certificate, and nominee's identity proof.

Occupational accident PA claims: for workplace accidents (falls, machinery, chemical exposure), the AI additionally collects the employer's accident report, the factory inspector's report (for scheduled industries), and the ESIC status (if the worker is ESIC-covered, PA claim and ESIC benefits are coordinated separately — ESIC benefit is not deducted from the PA payout under most policy terms).

Ambulance and emergency treatment coverage: many PA policies include an ambulance reimbursement clause (Rs 2,000–5,000 per accident) and emergency treatment benefit. The AI confirms these benefits at FNOL and advises the claimant to retain ambulance receipts — most policyholders discard these as minor costs, unaware they are reimbursable.

CI FNOL is an emotionally high-stakes call — the policyholder has just received a serious diagnosis and may be reaching out in a state of distress. Kallix AI's CI FNOL protocol is calibrated for empathy: slower pacing, extended listening, and explicit acknowledgment before data collection.

CI category confirmation is the most important technical step: CI policies cover a defined list of conditions (IRDAI standard critical illness list includes 20 conditions; individual insurer lists range from 7 to 37 conditions). The AI confirms whether the diagnosed condition matches the policy's list. Common mismatches: the policyholder is diagnosed with Stage 1 cancer but the policy covers major-stage cancer only; the policyholder has a TIA (transient ischaemic attack) but the policy requires a permanent neurological deficit for stroke coverage.

Survival period clause: most CI policies require the claimant to survive 30 days post-diagnosis to claim the payout. The AI collects the diagnosis date and treatment start date to confirm whether the survival period requirement is on track. For claims submitted within 30 days of diagnosis, the AI explains: 'Your claim has been registered. The settlement will be processed [N] days after your confirmed diagnosis date — you do not need to do anything further during this period.'

First symptom vs first diagnosis: some CI policies have a date-of-manifestation clause — the CI condition must have first manifested after the policy's waiting period end date. The AI collects first symptom date from the treating physician's notes and cross-checks against the policy issue date and waiting period. This pre-screening prevents claims that will be denied on temporal grounds — allowing the policyholder to build a reconsideration argument before formal submission.

OPD claims are the most frequently abandoned insurance benefit in India — most policyholders with OPD cover do not know the submission process, the applicable sub-limit, or the document requirements. Kallix AI activates this dormant benefit at the point of use.

OPD reimbursement call triggered by WhatsApp: Kallix monitors the insurer's WhatsApp service line for OPD claim submission requests. When a policyholder sends OPD bills via WhatsApp, the AI initiates an FNOL conversation: confirming the OPD benefit available (annual sub-limit, YTD utilisation), the document completeness, and the bank account for reimbursement.

Document completeness check: the most common OPD claim rejection is missing original prescriptions (lab tests without a doctor's prescription are typically not covered) or illegible bills (pharmacy receipt without medicine names and quantities). The AI reviews document quality in the WhatsApp submission and requests re-uploads for non-compliant documents before formal submission.

Dental and vision OPD: many comprehensive health plans include a dental and vision sub-limit (Rs 3,000–8,000/year each) that is almost never utilised because policyholders are unaware. Kallix identifies policyholders with unclaimed dental/vision sub-limits approaching the policy year-end and proactively sends a reminder — 'You have Rs [X] in unclaimed dental OPD benefit that expires on [date]. If you have any dental expenses from this year, you can still submit a claim.'

Health and wellness benefits: some plans include wellness benefit credits (gym membership, health app subscription) reimbursable under OPD. The AI confirms these benefits at the first OPD claim interaction and ensures the policyholder knows all available OPD categories.

After-hours claim intake is one of the most compelling value propositions of AI FNOL — 34–42% of motor accident FNOL calls arrive outside business hours, as do 28–34% of health emergency admission calls. A human-only FNOL operation either misses these calls entirely or provides a degraded experience via understaffed night shifts.

Kallix's after-hours FNOL is functionally identical to business-hours FNOL: the same qualification steps, the same CMS integration, the same cashless authorisation capability for health claims. For motor accidents at night: the surveyor appointment is booked for the following morning, but the claim registration is immediate — preventing the 6–8 AM pile-up when multiple overnight accident victims all call in simultaneously at business opening.

For health emergency admissions at night: the AI initiates the cashless pre-authorisation request with the TPA's after-hours desk (TPAs are required to operate 24/7 under IRDAI regulations). If the pre-auth is not received within 2 hours, the AI advises the hospital on the reimbursement pathway and triggers an internal escalation to the TPA.

Natural disaster FNOL surge management: during flood events, cyclones, or earthquake aftermaths, FNOL call volume can spike 10–15x normal volume. Kallix AI handles this surge without degradation — all calls are answered within 2 rings, claims registered in real time, and the insurer's catastrophe management team receives an aggregated claim registration feed. For the 2023 Tamil Nadu floods, insurers with AI FNOL registered 8,000+ claims within 48 hours of the event — vs 1,200 in the same window for comparable manual-only operations.

Public holiday claims: motor claims on Diwali night (peak road accident volume), New Year's Eve, and long weekends account for a disproportionate share of annual motor FNOL. Kallix's proactive holiday-period capacity confirmation ensures no call queue buildup on these high-volume days.

Subrogation rights are extinguished or weakened when the insured: (1) settles directly with the negligent third party and signs a release; (2) fails to preserve evidence (damaged property disposed, accident vehicle repaired before inspection); or (3) does not file an FIR for a motor accident. Kallix AI addresses each risk at the moment of FNOL.

Motor accident subrogation: if the accident was caused by another driver, the AI advises at FNOL: 'Do not accept any cash settlement or sign any document from the other driver or their insurer before speaking to us. Your insurer has the right to recover your claim cost from the responsible party — accepting a settlement without the insurer's consent may waive this right.' This advisory prevents the most common subrogation right loss scenario.

Product liability subrogation: for property damage caused by a defective product (fire caused by a faulty appliance, for example), the AI instructs the policyholder to preserve the product: 'Do not discard or repair the [appliance] that you believe caused the fire — your insurer needs to examine it to pursue the manufacturer for recovery. Please store it safely and photograph it from multiple angles before the surveyor visits.'

Marine/cargo subrogation: as highlighted in the marine cargo FNOL section, the surveyor must inspect cargo damage before the consignment is moved or unpacked further. For imported goods, the Bill of Lading and shipping documents establish the carrier's liability — the AI confirms these are collected at FNOL and the carrier has been put on notice in writing within the notice period specified in the B/L (typically 3 days for visible damage, 15 days for hidden damage).

Subrogation data in FNOL record: all subrogation-relevant information collected (third party details, FIR number, evidence preservation status) is flagged in the CMS claim record for the subrogation recovery team. This structured capture enables the recovery team to act immediately post-settlement without a time-consuming post-settlement data reconstruction exercise.

Group life death claims are high-sensitivity, time-critical events. The nominee family is grieving and needs financial support — delays in claim settlement amplify distress. Kallix AI's group life FNOL protocol prioritises speed and clarity above all else.

Nominee intake call: the AI begins with empathy: 'I am calling to register a claim for [Employee Name] and to make sure the family receives the full benefit they are entitled to as quickly as possible. I will need to collect a few details — please take your time.' The data collected: employee name, date of birth, employee ID (or policy number), employer name, date and cause of death, nominee's name and relationship, and current bank account for settlement.

Document checklist dispatched via WhatsApp within 5 minutes of the call: death certificate (original or notarised copy), nominee's Aadhaar and PAN, proof of relationship to the deceased, and — for accidental death — FIR, post-mortem report, and doctor's certificate. For natural death: medical records or doctor's cause-of-death certificate.

HR bulk intake: for larger employers with Kallix integrated into the HRMS, the death claim can be initiated by the HR team via an API call or a structured WhatsApp submission. The AI sends a pre-filled claim form to the HR team for each notified death, reducing HR administrative burden while maintaining speed.

Parallel benefit identification: Kallix AI identifies other entitlements simultaneously at FNOL: (1) EDLI (Employee Deposit Linked Insurance under PF Act): provides an additional death benefit of Rs 2.5–7 lakh from EPFO — nominee should file simultaneously with the EPFO claim; (2) ESIC death benefit if the employee was covered under ESIC (Rs 3 lakh dependent benefit); (3) Gratuity if applicable. The AI confirms all entitlements so the nominee family receives the full package, not just the group life payout.

Regional language FNOL is not a courtesy feature — it is a claims quality lever. A claimant who struggles to describe their accident or medical situation in English may omit critical details that affect claim eligibility. The same claimant narrating in Telugu provides a more accurate account, which reduces post-submission document requests and speeds settlement.

Kallix's language detection uses the first spoken phrase to identify the language. If the caller mixes languages (common in South India — Tamil with English technical terms; Hindi with Punjabi vocabulary), the AI adapts to the dominant language and defaults to simpler vocabulary for insurance concepts.

Terminology localisation examples: 'cashless claim' is explained in Tamil as 'naanee paNam kaттavillai — neeankkal hospital kaттuvaaнkal' (you don't pay — they pay the hospital); 'surveyor appointment' in Kannada uses the local government term for inspector ('parishIlikara nemakaatikege appointment'); 'FNOL' is never used in regional language calls — it is always 'accident vivarana nonthipu' (accident detail registration) in Malayalam.

Regional language document instruction: the WhatsApp document checklist sent after FNOL is also in the caller's language. For Hindi and Gujarati, the checklist includes the Hindi/Gujarati name of each document alongside the formal English name (e.g., 'FIR — Prathama Soochana Prativedan'). This reduces the confusion where claimants go to the police station and ask for the wrong document.

Dialect sensitivity: Kallix's AI handles common regional dialect variations — Bhojpuri-inflected Hindi from Eastern UP, Konkan Marathi from coastal Maharashtra, and coastal Andhra Telugu — with vocabulary flexibility rather than strict standard-language compliance.

Multi-product FNOL is one of the highest-value AI use cases in insurance: a single integrated call replaces multiple independent claims processes, reduces the claimant's burden significantly, and ensures no eligible claim is left unfiled.

Most common multi-product scenarios: (1) Road accident: motor OD + owner-driver PA + health (if injuries require hospitalisation) — 3 claims from one event; (2) House fire: property/building claim + contents claim + business interruption (if home-based business) + personal accident (if anyone was injured in the fire) — up to 4 claims; (3) Overseas medical emergency: travel insurance medical claim + regular health insurance (if travel policy limit is exhausted) + PA (if injury was accidental).

CRM policy linking: the AI checks whether the caller has multiple policies that could be triggered by the described event. If a policyholder calls for a road accident motor claim and their CRM profile shows an active health policy and a PA cover, the AI proactively asks: 'Were you injured in the accident? If so, your PA cover and health policy may also provide benefits — shall I register all three claims now?' This proactive linkage recovers 18–26% of PA claims that would have otherwise been filed separately or not at all.

Consolidated document checklist: for a 3-policy claim, the AI generates a single unified document list that avoids asking for the same document (e.g., FIR, hospital admission record) three times. Documents required by all three policies are listed once; policy-specific documents are listed separately with the applicable policy number.

Surveyor coordination: for property damage claims with motor damage in the same event (vehicle in the garage also damaged during fire), the AI books a single surveyor with expertise in both property and motor assessment — reducing the number of surveyor visits from two to one.

Commercial liability FNOL is structurally different from retail claims: the insured is a business, the claimant is a third party, and the insurer's role is to defend the insured and settle legitimate third-party claims. The AI's FNOL intake focuses on building a complete incident record that the legal defence team can act on immediately.

Public Liability FNOL: triggered when a business's premises or operations cause injury or property damage to a third party — a customer injured at a retail store, a contractor's equipment causing property damage at a client site. The AI collects: incident date, time, and location; third-party details (name, contact, injury description); witness details; CCTV footage preservation advisory ('please preserve CCTV recordings for the past [N] hours — they may not be recoverable after 48 hours'); and whether any demand letter or legal notice has been received.

Product Liability FNOL: when a product manufactured or sold by the insured causes injury or damage. The AI collects the product batch number, purchase date, and the nature of the defect or failure alleged. The insurer's recall and quality team is notified alongside the claims team if the incident pattern suggests a batch-wide issue.

Directors and Officers (D&O) FNOL: triggered by regulatory investigations, shareholder suits, or employment practices claims. D&O claims are highly sensitive — the AI handles intake with confidentiality protocols, collects the nature of the claim, the regulatory body involved (SEBI, MCA, CBI), and the insured's external legal counsel's contact. D&O FNOL is always routed to the senior commercial claims team for immediate review.

Equipment breakdown FNOL: for industrial machinery, HVAC systems, and electronic equipment failures. The AI collects the equipment details (make, model, age, maintenance records), the failure description, and confirms the business interruption exposure (is production stopped? What is the daily revenue impact?). Business interruption co-ordinate with the machinery claim is initiated simultaneously.

Home loan-linked insurance claims are complex because the bank is both a policy beneficiary (for the loan exposure) and the entity that facilitated the insurance purchase. Kallix AI coordinates the three-party communication — insured, insurer, and bank — from the FNOL call.

For property damage claims: the AI confirms which insurer the home loan policy was placed with (some banks bundle the insurance with the loan at origination; others allow the borrower to arrange separately). If the policy was bank-assigned, the settlement cheque is issued jointly to the borrower and the bank. The AI explains this upfront and collects both parties' required documentation in a single coordinated process.

For fire total loss claims on mortgaged properties: the bank's outstanding loan balance at the date of loss determines the settlement split. Kallix AI triggers a loan balance confirmation request to the bank's loan servicing team simultaneously with the claim registration — so both figures (insurance settlement amount and outstanding loan) are available to the claims assessor within 48 hours.

Credit life FNOL for death claims: the nominee calls the bank about the outstanding home loan. Kallix AI, integrated with the bank's contact centre, identifies the credit life policy attached to the loan and initiates both the claim with the life insurer and a temporary loan freeze (interest accrual suspension during claim processing) — protecting the nominee from accumulating interest while the claim is being processed.

For disability credit life claims: the AI collects the borrower's disability certificate, confirms the policy's definition of qualifying disability (typically permanent total disability), and coordinates the premium waiver or loan repayment on the borrower's behalf.

FNOL quality is the foundation of claims performance: every missing data field at registration creates a downstream task — a follow-up call, a document request, an assessor query. Kallix's automated quality audit catches these gaps before the claim is assigned to an assessor.

The completeness audit checks 14 mandatory fields for motor FNOL: date of loss, time of loss, location (GPS or address), type of incident (own damage, third-party, theft, fire, flood), third-party details (if applicable), FIR reference (if required), vehicle registration number, witness details, immediate repair status, surveyor preference, and document checklist confirmation. For health FNOL: 11 mandatory fields. For life FNOL: 9 mandatory fields.

Caller distress and call drop recovery: when a caller is in visible distress (crying, shouting, rapid speech pattern) during the accident FNOL, the AI prioritises empathy over data completeness — accepting whatever the caller can provide and flagging the incomplete fields for a follow-up call within 2 hours. The follow-up call begins: 'I spoke with you earlier today about your accident. I wanted to check how you are doing and collect a few additional details to complete your claim registration.' This two-call approach achieves higher data quality than forcing distressed callers to complete a full data set in one sitting.

Claim duplication detection: the quality audit also checks whether the same claim has been registered twice (common when policyholders call multiple times in distress). Duplicate CRNs are merged automatically, and the policyholder is notified: 'We found that your claim has been registered once — there is no need to register again.'

Audit-to-insight: aggregate quality audit data identifies the most frequently missed data fields across the FNOL agent's calls. This data drives script improvement — if 30% of motor FNOL calls miss the witness details field, the script is updated to ask more directly for this information.

Insurance FNOL is a high-volume, high-stakes operation where every minute of delay has measurable financial and reputational consequences. The ROI case for AI FNOL is built on five quantified levers, all validated across production deployments at general and life insurers.

Lever 1 — Cost per FNOL: human agent FNOL costs Rs 450–750 per call (including training, attrition, management overhead, and after-hours shift premium). AI handles the same call at Rs 120–200 all-in. At 50,000 FNOL calls per year (typical for a mid-size general insurer), the annual saving is Rs 1.6–2.8 crore.

Lever 2 — Registration TAT: 4–6 hour average in manual processes (call centre queue + agent availability + data entry) reduced to under 8 minutes. Faster registration enables faster surveyor dispatch and faster settlement — each day of settlement TAT improvement has actuarial value in fraud prevention and claimant satisfaction.

Lever 3 — Document collection TAT: 18–22 days reduced to 8–11 days via Day 3/7/14 automated reminders. Each 10-day reduction in document TAT translates to earlier settlement and reduced working capital tied up in outstanding claims reserves.

Lever 4 — IGMS complaint reduction: 35–50% fewer settlement-delay grievances. Each IGMS complaint costs 4–8 hours of internal response time plus regulatory exposure. At average 2,000 delay complaints per year reduced by 40%, saving Rs 80–120 per complaint resolved = Rs 64–96 lakh annually.

Lever 5 — Fraud prevention at FNOL: real-time IIB check and fraud pattern detection. Each fraudulent claim intercepted at FNOL (vs discovered post-settlement) saves the average claim value of Rs 85,000–2,50,000 for motor OD, with incremental AI cost near zero per check.

Deployment timeline: 5–7 weeks — Week 1–2 PAS/CMS integration scoping, Week 3–4 API development and testing, Week 5 pilot (100 live calls with monitoring), Week 6–7 full rollout. Multi-insurer group (holding company with life + general + health subsidiaries): 8–10 weeks for unified FNOL across all three lines.

PMFBY (Pradhan Mantri Fasal Bima Yojana) is the world's largest crop insurance scheme by enrolled area — covering 5.5 crore farmers and 50 million hectares annually. FNOL under PMFBY is structured differently from private insurance: the 72-hour intimation window is strict, the settlement trigger is a yield-based formula (not individual crop inspection), and the premium is subsidised (2% for Kharif, 1.5% for Rabi, 5% for commercial crops).

The AI handles PMFBY FNOL via voice in the local language (Hindi, Telugu, Tamil, Marathi, Gujarati, Kannada, Odia, Bengali, Punjabi, and Malyalam are priority languages for agricultural states). The agent captures: farmer's name and Aadhaar, district and tehsil, village code (for NCIP), bank account for subsidy (Kisan Credit Card or Jan Dhan account), crop type and sown area (in hectares), date and nature of damage, and nearest CSC for document submission if the farmer lacks smartphone access.

The AI transmits this data to NCIP via API integration. The farmer receives an SMS acknowledgement with a Claim Reference Number. Post-intimation, the state government conducts Crop Cutting Experiments (CCE) — random crop sampling to determine average yield loss in the notified unit (typically taluk or block level). The AI advises the farmer that individual field inspection is not standard under PMFBY — the CCE-based yield estimate applies to all farmers in the notified unit.

For localised calamity claims (hailstorm, landslide, inundation — where damage is plot-specific, not area-wide), individual crop inspection is conducted — the AI captures GPS coordinates of the damaged plot for the surveyor.

Resheduled sowing compensation (under PMFBY's Prevented Sowing component) is advised when drought prevents sowing altogether — 25% of sum insured is paid if sowing is prevented. The AI captures sowing failure details for prevented sowing claims specifically.

Cyber insurance is the fastest-growing line in Indian commercial insurance, driven by the Digital Personal Data Protection Act 2023 (DPDP) and increasing ransomware incidents targeting Indian SMEs. FNOL for cyber claims is uniquely time-critical: the policy breach notification window (24–72 hours), active incident containment requirements, and regulatory notification obligations all converge simultaneously.

The AI handles cyber FNOL with a triage approach. For active ransomware incidents (system encrypted, operations frozen), the agent skips extended data capture and immediately connects to the insurer's cyber incident response (CIR) hotline — typically a tie-up with a cybersecurity firm (IBM X-Force, Arete, Mandiant, or insurer in-house). The agent captures only: policy number, company name, nature of incident, and callback number — transfer happens in under 60 seconds.

For post-incident claims (BEC/financial fraud discovered after funds transferred, data breach discovered by third party), the AI conducts full FNOL intake: incident timeline, amount lost (for BEC/wire fraud), data categories affected (personal data, financial data, health data — each triggers different DPDP regulatory obligations), affected customer count, and whether a Cyber FIR has been filed with the local Cyber Crime Police Station or National Cyber Crime Reporting Portal (cybercrime.gov.in).

DPDP Act 2023 notification obligations: data fiduciaries must notify DPDP Board of India within 72 hours of a personal data breach. The AI advises this obligation and its timeline — insurers increasingly offer DPDP breach notification management as a covered service under cyber policies. GDPR notification (72 hours to supervisory authority) is advised for Indian companies with EU customer data.

Cover components advised at FNOL: forensic investigation costs, legal and regulatory defence costs, notification costs (informing affected individuals), ransomware extortion payment coverage (subject to regulatory guidelines), and business interruption losses during system recovery.