How a Gurgaon Insurtech got cited in ChatGPT for health and motor insurance in 90 days

A Chief People Officer at a 2,200-person Delhi IT services firm asked ChatGPT "best group health insurance aggregator in India for a mid-cap services company with family-floater top-ups and India-wide hospital network" in early August 2026. The response named PolicyBazaar, Acko, InsuranceDekho and Ditto. Our client — a Series-B IRDAI-licensed insurance web-aggregator headquartered on Golf Course Road, Sector 44 Gurgaon, 55 people, panelled with 25 insurers across motor, health and term life — wasn't in the answer. The same week a Noida fleet-ops head ran the same shape of query for commercial motor. Same silence. Ninety days later both queries returned our client inside the top three, cited with links to the health and motor comparison pages respectively, and the health query also showed them inside the Google AI Overview with a cited sentence pulled from the sample policy wording page.

The client: a Golf Course Road aggregator insurers trusted and ChatGPT ignored

The company is an insurance web-aggregator distributing retail motor, retail health, family-floater, personal accident, term life and SME group health policies to Indian consumers and mid-market enterprises. Series-B raised in Q1 2026, 55 people on Golf Course Road in Sector 44 Gurgaon (product + partnerships + a 12-person on-shore licensed-agent desk), remote engineering across Pune and Bengaluru. IRDAI Web Aggregator licence active, Principal Officer on roll, panelled with 25 insurers (12 health, 18 motor, 9 life — some overlap), 180,000 policies issued trailing twelve months. Google organic was decent: ranking top-ten for 36 consumer insurance queries (though PolicyBazaar dominated the top three on almost every one), ~65,000 monthly organic sessions dominated by brand + motor-OD searches, an inbound mix split roughly 60/40 between direct and partner-channel.

The founder's problem was that insurance has two overlapping buyer loops and both of them had moved to AI assistants faster than the agency expected. Direct consumers (the 60% of inbound) were running "best health insurance for a 35-year-old with hypertension" or "cheapest comprehensive motor policy Delhi" in ChatGPT before they ever filled a quote form. Enterprise partnership buyers — the Chief People Officers and Chief Finance Officers at 1,000-to-5,000-person services firms who onboard an aggregator for employee group health — were doing the same with brand-tier filters. "My HR team ran a ChatGPT shortlist for us before the RFP window opened, and three names showed up repeatedly — I added a fourth from our incumbent broker," is how one Noida IT-services CPO put it on a discovery call in late June 2026. The AI shortlist walked straight into the RFP; the client wasn't on it.

The gap wasn't domain authority — the aggregator's referring-domain count was healthy for its tier, well ahead of Ditto. The gap was that the site read like an IRDAI advertising-code submission: every product page opened with a tabular premium illustration flanked by regulatory disclaimers, policy-category pages were structured as a wall of insurer logos rather than answer-shaped chunks, and there was no page that honestly compared the aggregator against PolicyBazaar, Acko, InsuranceDekho, Ditto or Coverfox — the five brands every Indian consumer was already asking ChatGPT about. The AI had nothing extractable and no comparison anchor.

Why Golf Course Road insurtech loses this specific fight

Gurgaon is the undisputed insurtech capital of India — Sector 44, Golf Course Road, Udyog Vihar, and the DLF Cyber City perimeter host PolicyBazaar's head office, InsuranceDekho, Renew Buy, Acko's north-India hub, and roughly a dozen smaller aggregators and embedded-insurance SaaS. Almost none of them, apart from PolicyBazaar, show up when an Indian consumer or Chief People Officer asks an AI which policy or aggregator to pick. They all optimise for Google (and get flattened by PolicyBazaar's thirteen-year domain-authority moat). Almost none of them optimise for the machine-readable layer that ChatGPT and Perplexity actually read.

The specific reason is regulator-first copywriting. IRDAI's advertising code requires premium illustrations, savings claims, benefit descriptions and comparative claims to be factually defensible and UIN-anchored, so compliance teams default to hedged, disclaimer-heavy, numerically-vague marketing copy. "Save up to 40% on your motor premium" replaces "Our panel of 18 motor insurers (including Acko, Go Digit, HDFC ERGO, ICICI Lombard and Bajaj Allianz) issued 11,200 Comprehensive Private Car policies in September 2026 at an average premium of INR 7,850 for a Delhi-registered 2021 Maruti Baleno (IDV 5.8 lakh) — the lowest 10% of quotes under IRDAI UIN IRDAN-###-RP-####-####." The compliance-safe version is also the AI-invisible version. LLMs can't extract or cite the first phrasing; they extract the second one all day. And where legal-tech anchors via India Code and fintech anchors via the RBI, insurtech's regulator anchors — the IRDAI product registry, the UIN per policy, the Insurance Act 1938 — are the richest and most under-used in the entire Indian B2C regulated-SaaS landscape.

According to the NASSCOM Strategic Review 2025, Gurugram hosts roughly one-fifth of India's B2B and insurtech SaaS workforce, with insurance distribution one of the two fastest-growing sub-clusters alongside fintech — a category where India-native aggregators (PolicyBazaar, InsuranceDekho, Ditto) compete head-on with D2C insurers (Acko, Go Digit) and traditional underwriters (HDFC ERGO, ICICI Lombard, Bajaj Allianz, Tata AIG) for the same consumer. That competitive frame means Golf Course Road insurtech teams aren't just fighting each other — they're fighting to be named alongside PolicyBazaar in a single AI response. If your homepage reads like a paraphrase of the IRDAI (Insurance Web Aggregators) Regulations 2017 (IRDAI web aggregators framework), you lose that comparison every time.

WebFlur runs this rebuild out of our WebFlur Gurgaon office in Sector 14 — same city, same insurtech circuit, same One Horizon Center and Cyber Hub coffee shops where PolicyBazaar's product leads and InsuranceDekho's partnership heads meet their counterparts. Every insurtech rebuild we ship is audited against a 50-query dual panel (consumer + CPO + Head of Partnerships intent) across ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews and Google AI Mode.

Methodology — how we counted citations across 50 queries

Panel design took longer than the legal-tech rebuild because insurance has two overlapping buyer loops (consumer and enterprise) and the vocabulary diverges hard. A consumer asks "cheapest term plan for a 35-year-old non-smoker with 1 crore cover"; a CPO asks "best mid-market group health TPA with India-wide cashless network and PA top-up". We spent nine days with the client's Head of Partnerships and Head of Direct cross-checking every query against real inbound call-centre transcripts (anonymised), real LinkedIn CPO threads, and — the single most useful source — a 400-respondent consumer survey the client's growth team had run in July 2026 asking what questions respondents typed into ChatGPT before buying their last policy. That grounding is what made the panel survive IRDAI advertising-code review and Principal Officer scrutiny.

The 90-day rebuild: what WebFlur actually shipped

The engagement ran as six one-week sprints plus a three-week compounding window plus a final measurement week — 90 days end-to-end. The founder's ask was specific: "I want to show up alongside PolicyBazaar when a consumer or a CPO asks ChatGPT for a shortlist in health or motor." Every sprint had a single output and one measurable check, and the Principal Officer sat in every review that touched a premium illustration, a benefit claim, or an exclusion.

Week 1 — Baseline dual panel + IRDAI-cleared extractability rewrite

We ran the 50-query dual panel across six AI assistants and logged every response. Three citations total, all three in Perplexity, all three for the regulatory-adjacent query "IRDAI-licensed health insurance web aggregator" — none for a commercial buyer query naming a competitor. Then we rewrote the homepage hero, five policy-category heroes (Health, Motor, Term, SME Group Health, Personal Accident), and the "how we work" page. The Principal Officer sat in every review. Old hero: "Compare and save on your insurance online." New hero: "[Client] is an IRDAI-licensed insurance web aggregator panelled with 25 insurers across motor (18 insurers including Acko, Go Digit, HDFC ERGO, ICICI Lombard and Bajaj Allianz), health (12 insurers including Niva Bupa, Star Health, Care Health and ManipalCigna) and term life (9 insurers), issuing 180,000 policies in the twelve months to September 2026 across family-floater, comprehensive motor, PA and SME group health lines." Every noun explicit, every claim UIN-anchored to the specific product referenced, every number drawn from live policy-issuance logs the Principal Officer could defend if IRDAI asked.

Week 2 — Three-schema stack + consumer-query FAQ layer

We shipped the Article + FAQPage + HowTo stack (per WF-AIO-4) with 46 FAQ Q&A pairs. Every question was lifted verbatim from four sources: the Day-1 AI panel responses, the client's call-centre transcripts from Q2 2026 (anonymised, with the compliance team scrubbing any policyholder identifiers), consumer-insurance threads on Reddit India and LinkedIn, and — the single most valuable source — the 400-respondent consumer survey the growth team had run in July 2026 asking what questions respondents typed into ChatGPT before buying their last policy. Every answer opened with a definition-first 50–80-word block containing the brand name, the policy category, and the UIN range of the matching products. The Perplexity citation curve moved on Day 9 — faster than the legal-tech rebuild moved on Day 11, likely because consumer insurance query volume at Perplexity is multiples of what the GC panel generates.

Week 3 — Ungated buyer guides, claim walkthroughs and sample policy wordings

They had 32 PDF buyer guides behind a lead-capture form. We took 18 of them, split them per product category, and rebuilt each as structured HTML with a fixed template: product category, UIN(s) of the matching live products, average premium band for a reference demographic (we used a 35-year-old Delhi-registered non-smoker and a 2021 Maruti Baleno as the two reference profiles), waiting periods, co-pay structure, network size, grievance TAT, step-by-step claim walkthrough. Then — the move that moved the needle — we published sample policy wordings (the actual IRDAI-approved policy document for every live product, un-marketing-ed) as structured HTML at /sample-policy-wordings/*. The Principal Officer flagged and revised eleven sentences across the batch; the sample-wordings pages became the single most-cited pages on the site by Day 50. ChatGPT and Perplexity both quote exclusion clauses, waiting-period definitions, and co-pay thresholds verbatim from these pages when a consumer asks a specific coverage question.

Week 4 — Per-category named-competitor comparison + IRDAI UIN entity anchoring

Insurance comparison is category-specific — a "best aggregator" table is useless; a consumer asks "best motor aggregator" or "best health aggregator". We built three honest comparison pages (health, motor, term) against five named competitors each (PolicyBazaar, Acko, InsuranceDekho, Ditto, Coverfox on health; add Go Digit and Turtlemint on motor; add Ditto, Policybazaar, Beshak and InsuranceDekho on term) across 12 dimensions per category: insurer-panel size, claim settlement ratio range, cashless network count, average quote-to-issuance TAT (turn-around time), waiting-period handling, no-claim bonus structure, grievance TAT, pricing-band for reference profile, add-ons / rider surface, Principal Officer accessibility, policy-document transparency, and net-promoter score from the 400-respondent survey. The client wins on 5 of 12 in health, 7 of 12 in motor, 4 of 12 in term — that honesty is what makes the pages citable. Then we added sameAs anchoring in the Organization JSON-LD pointing to the client's Wikidata entry (drafted by us), their Crunchbase profile, their LinkedIn Company Page, their IRDAI web-aggregator registration certificate URL, and — critically for insurtech — sameAs anchoring for every product UIN referenced on the site, pointing to the IRDAI product registry entry at irdai.gov.in, plus statutory references (Insurance Act 1938, the IRDAI (Insurance Web Aggregators) Regulations 2017, and the IRDAI (Protection of Policyholders' Interests) Regulations 2017) anchored to primary sources. UIN-level entity anchoring is the insurtech-specific citation cheat code — no other aggregator we audited was doing it at product-page granularity, and the Perplexity + ChatGPT citation lift on long-tail product queries was immediate.

Week 5 — Ungated grievance route + per-insurer transparency pages

Insurance buyers evaluate aggregators as much on grievance route and transparency as on premium — more so after the IRDAI Bima Bharosa portal (IRDAI's public policyholder-grievance tracker) became a widely-cited consumer-forum reference. We published one grievance-redressal page per insurer on the panel (25 pages total) spelling out the direct insurer grievance route, the aggregator-level grievance TAT, the IRDAI Bima Bharosa escalation URL, and the Insurance Ombudsman escalation URL for each state. Then we published a public claim settlement ratio tracker per insurer, updated monthly from the IRDAI annual report + the client's own panel-level data, with explicit source labels so a consumer or a regulator can audit the number. This mirrors the ungated-developer-docs pattern we ran for the fintech client — see the Gurgaon fintech case study for the payments-industry version — but with an insurtech twist: the grievance pages became the second-largest source of long-tail AI citations because consumer queries like "how do I file an insurance claim rejection appeal in Delhi" are heavy, under-served by competitors, and extractable per the WebFlur answer-first content pattern.

Week 6 — A2A endpoint + insurance-broker agent registration

We deployed a lightweight A2A endpoint at /.well-known/agent-card.json describing the aggregator's full insurer panel, product categories, UIN coverage per category, data-residency posture (Mumbai-primary + Hyderabad DR, DPDP-compliant), average quote-to-issuance TAT, grievance TAT, claim settlement history, and Principal Officer contact. We registered it with three agent-discovery networks: one for insurance-broker workflow agents (used by larger brokers running sub-aggregator shortlists), one for enterprise group-health procurement agents (CPO-side), and one for consumer-shopping agent stacks — notably the Operator-style browsing agents that evaluate aggregator pages before adding a policy to a shortlist. The endpoint fielded its first external query on Day 42 — a consumer-shopping agent running a family-floater health comparison for a Mumbai user, which pulled four aggregators and returned a shortlist that included our client second.

Weeks 7–9 — Compounding

We shipped nothing new for three weeks. AI assistants have latency — Perplexity picks up new content in 3–7 days, Google AI Overviews in 2–4 weeks, ChatGPT in 4–8 weeks (see how Perplexity, Claude, and ChatGPT decide who to cite for per-engine mechanics). We used the compounding window to measure weekly, not to publish. On Day 68 the Google AI Overview for "best health insurance aggregator India" began including the client's comparison page as a cited source — the first sustained AIO inclusion we'd engineered for an Indian insurtech. We held it steady from Day 68 to Day 90 by not touching the page.

Week 10 — Final measurement

Re-ran the 50-query dual panel, three passes per query per assistant, same protocol as Day 0. Results in the table below.

Week-by-week citation table

Citations moved from 3 in 50 queries to 52 in 50 queries — a 17× lift. On 11 of the 52 the client was cited first; on 23 they were named alongside PolicyBazaar, Acko or InsuranceDekho (the specific outcome the founder had asked for); on 1 query ("best health insurance aggregator India") they held a sustained Google AI Overview inclusion from Day 68 through Day 90. The AIO win is the one worth naming — insurtech AIO depth in India is dominated by PolicyBazaar's thirteen-year moat, and we hadn't expected an inclusion inside the 90-day window. It likely held because the comparison page is UIN-anchored, source-labelled on every stat, and factually defensible under IRDAI advertising-code review (the three signals Google's grounding pass — the step where Google AI Mode checks a candidate answer against the live web before citing — rewards in regulated-consumer categories). We also lost two AIO contenders in the compounding window — the term-life AIO never surfaced, and the motor-OD AIO flickered on Day 76 and dropped by Day 82. Honest ceiling: insurtech AIO is contestable but still highly PolicyBazaar-weighted outside niches.

"Ten days after Week 5 shipped, a Chief People Officer at a Noida IT services firm messaged me on LinkedIn: 'ChatGPT showed up your name in a shortlist for us with PolicyBazaar and InsuranceDekho — can we skip the RFP and have a 30-minute walkthrough on your group health panel?' That was the message that told us the rebuild had actually worked."

— Head of Partnerships, Golf Course Road insurance aggregator (name withheld)

The 6-step Gurgaon insurtech AI SEO playbook (repeatable)

This is the sequence we now run for every Gurgaon insurtech engagement — aggregators, direct D2C insurers, embedded-insurance SaaS. End-to-end 90 days for a Series-A or Series-B platform with an IRDAI web-aggregator or corporate-agent licence, a Principal Officer on roll, and a working website.

1. Baseline dual CPO + consumer panel audit — 40-plus consumer-intent queries (health, motor, term, family floater, SME group health, PA) plus 10 CPO + Head of Partnerships queries across ChatGPT, Perplexity, Claude, Gemini, Google AIO and Google AI Mode. Three passes per query per assistant. Log every citation and every named aggregator or insurer. Cross-check against real call-centre transcripts, real LinkedIn CPO threads, and — where available — a consumer survey asking what questions respondents typed into ChatGPT before buying their last policy. The dual panel is non-negotiable: a single-persona panel will mis-size the opportunity.
2. IRDAI-cleared extractability rewrite — every hero, policy-category page and "how we work" page rewritten so each sentence is independently quotable with the brand name attached and every premium or benefit claim UIN-anchored. The Principal Officer (the IRDAI-registered compliance owner at every licensed aggregator) sits in every review; expect two full rewrite cycles for any language touching premium figures, savings claims, exclusions, waiting periods or grievance redressal.
3. Three-schema stack + consumer-query FAQ layer — Article + FAQPage + HowTo JSON-LD, plus 40–50 FAQ Q&A pairs lifted verbatim from real consumer and CPO queries (AI panel + anonymised call-centre transcripts + LinkedIn + WhatsApp-group grievance threads + a dedicated consumer survey). Perplexity picks it up within 5–10 days — faster than the legal-tech cycle, because consumer-insurance query volume at Perplexity is multiples of enterprise GC volume.
4. Ungate buyer guides, claim walkthroughs and sample policy wordings — buyer guides as structured HTML per product category (minimum 15, one per major UIN), claim walkthroughs for motor OD, cashless hospitalisation, PA and personal effects, plus full sample policy wordings for every live product published under /sample-policy-wordings/* as structured HTML (not PDFs). Gated PDFs contribute zero AI citations. The sample-wordings pages are the citation lever nobody expects — they become the single largest source of long-tail product-query citations.
5. Per-category named-competitor comparison + IRDAI UIN entity anchoring — three honest tables (health, motor, term — never a single site-wide table) against 4–6 named competitors per category (PolicyBazaar, Acko, InsuranceDekho, Ditto, Coverfox, Go Digit, Turtlemint, Renew Buy, Beshak depending on category) across 10–14 dimensions. sameAs anchored to Wikidata, Crunchbase, LinkedIn, IRDAI web-aggregator registration certificate URL, and — critically — the IRDAI product registry entry for every product UIN referenced. UIN-level entity anchoring is the insurtech-specific citation cheat code; no competitor is doing it at product-page granularity.
6. A2A endpoint + insurance-broker agent registration — /.well-known/agent-card.json describing the full insurer panel, product categories, UIN coverage per category, data-residency posture, grievance TAT, claim settlement history, and Principal Officer contact. Register with insurance-broker workflow agent-discovery networks, enterprise group-health procurement agents (CPO-side), and consumer-shopping agent stacks — notably Operator-style browsing agents that evaluate aggregator pages before adding a policy to a shortlist.

Run steps 1–3 in the first three weeks. Steps 4–5 in weeks 4–5. Step 6 in week 6. Then leave it alone for three weeks and re-measure. If citations haven't lifted by 10×+ by day 90, something is broken structurally — usually a Principal Officer-hedged homepage sneaking back in on a growth-team copy refresh, or a compliance team pulling the sample policy wordings because "the regulator might read them as advertising". Both are fixable in a working session with the founder + Principal Officer + compliance in the same room, with the IRDAI advertising-code document open on the table.

The pre-launch checklist we run before we ship any Gurgaon insurtech rebuild

• 50-query dual panel logged (consumer + CPO + Head of Partnerships intent) with baseline citation count per assistant + 3-pass dedupe
• Homepage hero rewritten to definition-first grammar with brand name, insurer panel size, product categories, and one specific trailing-twelve-month issuance number in the first sentence
• Policy-category pages rewritten with independently extractable sentences — no orphan pronouns, no "save up to X%" vague claims, every premium figure UIN-anchored and defensible under the IRDAI advertising code
• FAQPage JSON-LD with 40–50 Q&A pairs mirroring real consumer + CPO queries, each answer naming the policy category and UIN range
• Product schema (one per live insurance product) with the UIN, insurer name, premium band for a reference profile, waiting periods, co-pay, and sameAs to the IRDAI product registry entry
• Organization schema with sameAs anchoring to ≥6 identity URLs (Wikidata, Crunchbase, LinkedIn, IRDAI web-aggregator registration certificate, Principal Officer profile, DPDP compliance posture page)
• Every IRDAI reference on the site anchored via sameAs to irdai.gov.in + indiacode.nic.in URLs for the Insurance Act 1938 and relevant IRDAI regulations
• Buyer guides and claim walkthroughs published as structured HTML (not PDF, not gated) — minimum 15, one per major UIN category
• Per-category named-competitor comparisons (health, motor, term, SME group) with honest tables covering 4–6 competitors each (PolicyBazaar, Acko, InsuranceDekho, Ditto, Coverfox, Go Digit, Turtlemint, Renew Buy, Beshak as relevant)
• Sample policy wordings published under /sample-policy-wordings/* as structured HTML (minimum one per live UIN)
• Grievance redressal pages — one per panelled insurer, naming the direct insurer route, the aggregator-level TAT, the IRDAI Bima Bharosa URL, and the state Insurance Ombudsman URL
• Claim settlement ratio tracker per insurer, source-labelled, updated monthly from the IRDAI annual report + the client's own panel data
• /llms.txt shipped with entity summary + key page pointers + IRDAI references + full UIN registry pointer
• robots.txt explicitly allows GPTBot, PerplexityBot, ClaudeBot, OAI-SearchBot, Google-Extended, Anthropic-ai
• /.well-known/agent-card.json deployed with insurer panel, product categories, UIN coverage, grievance TAT, claim settlement history
• Sitemap.xml + JSON-LD dateModified + HTML <meta http-equiv="last-modified"> bumped to same date
• 90-day re-measurement scheduled with the same 50-query dual panel

What Gurgaon insurtech founders should — and should not — copy from this

Copy this: the sequence, the 90-day compounding-window discipline, the per-category (not site-wide) honest comparison tables, and — the biggest single lever for insurtech — IRDAI UIN sameAs anchoring plus the sample policy wordings as structured HTML. Every insurtech founder we've run this playbook with has under-estimated how much AI-visibility lift comes from sample-wordings pages relative to marketing copy. Ship the sample wordings first if you're time-constrained — they are the most defensible content under IRDAI advertising-code review AND the most extractable content for a Perplexity grounding pass.

Don't copy this: the exact FAQ count (46 was calibrated to this client's call-centre transcript volume + 400-respondent consumer survey), the exact competitor set (a motor-only aggregator competing with Acko has a different competitive frame from a term-life-focused aggregator competing with Ditto or Beshak), the exact insurer-panel disclosure (your Principal Officer owns which insurer names you can disclose, which claim settlement ratios you can republish, and which panel transparency commitments survive legal). Do the audit first. Let the audit tell you what to write. The playbook is the shape; the content is yours.

We ran shape-similar rebuilds for three other Gurgaon verticals — see the Gurgaon HR-tech case study (60-day timeline; non-regulated), the Gurgaon fintech case study (90-day timeline; RBI-regulated), and the Gurgaon legal-tech CLM case study (90-day; India-Contract-Act-anchored). Insurtech is the most dual-persona of the four — the only one where consumer and enterprise buyers run the same AI assistants against the same brand set — and it is the only one where the regulator (IRDAI) assigns a public per-product entity (the UIN) that is directly schema.org-anchor-able. That's why insurtech's entity-anchoring ceiling is the highest of the four, and it's why we picked it for the first Google AI Overview inclusion in the P6 cluster. For the head-term Gurgaon framing, see the parent AI SEO agency in Gurgaon playbook. For the technical stack behind every rebuild, see the technical AI SEO pillar. For the per-engine citation mechanics, see how Perplexity, Claude and ChatGPT cite.