Articulate · GIG Gulf · Competitive Intelligence

Motor Insurance Price-Checker V1.7 · 19 Jun 2026

An automated bot that pulls live UAE motor-insurance quotes across multiple insurers and car profiles, then normalises them so GIG can see where it really sits — not just on headline price.

Goal

Give GIG Gulf a repeatable, like-for-like read on how its motor premiums compare to the market. Headline aggregator prices mislead — every insurer quotes its own agreed value and excess — so the goal is a normalised comparison (rate per unit of cover, same risk profile) plus the structural intel: who competes at each price point, and what they all include.

Process

Funnel audit — mapped every UAE aggregator + direct-insurer quote funnel; classified which give an instant online price vs gate behind OTP/captcha/callback.
Synthetic personas — three fictional applicants, identical risk profile (35, British, Dubai, 5+ yrs licence, no claims), differing only by car, mapped 1:1 to three car tiers.
Automated run — headless browser drives the InsuranceMarket.ae funnel for each persona, from a UAE residential IP, and captures the full quote set.
Normalise & analyse — convert each premium to a rate per AED 1,000 of agreed value, flag agreed-value gaps, and extract what the comparators share.

Tools

Execution

Mac Studio (always-on, Dubai/Etisalat UAE residential IP) → headless Chromium via Playwright. Native local exit beats any cloud proxy for looking like a real UAE customer.

Harness

Deterministic Playwright form-filler against mapped funnel selectors. (Browserbase verified-browser + UAE residential proxies staged for rotation/anti-bot in later versions.)

Source

InsuranceMarket.ae aggregator (instant online prices, no OTP). Policybazaar staged as second spine. GIG & Sukoon direct funnels are OTP-gated — need a SIM.

Output

Structured price table per persona → normalised rate analysis → this versioned page.

V1 Run — results

InsuranceMarket.ae · comprehensive (and TPL) · pulled live 19 Jun 2026 · as-quoted (each insurer's own agreed value, not pinned).

Persona A — Toyota Corolla 2021

Insurer	Plan	Premium AED	Agreed value	Excess	Rate / 1,000
Fidelity United	Fidelity Plus	1,600	50,000	350	32.0
Sukoon	Gold	1,700	42,500	250	40.0
GIG Gulf (AXA)	Motor Perfect	1,800	29,085	250	61.9
RAK	Full Option	1,805	30,673	350	58.8
Al Wathba	Comp. Non-Agency	2,000	43,616	350	45.9
Watania Takaful	Mumtaz	2,500	49,400	300	50.6

Persona B — Nissan Patrol 2022

Insurer	Plan	Premium AED	Agreed value	Excess	Rate / 1,000
Fidelity United	Fidelity Plus	1,957	130,500	1,000	15.0
GIG Gulf (AXA)	Motor Perfect	1,827	116,294	750	15.7
Sukoon	Gold	2,867	120,275	1,000	23.8
GIG Gulf (AXA)	Motor Perfect Plus (agency)	6,790	116,294	750	58.4
Al Wathba	Comp. (agency)	9,900	141,422	750	70.0

Persona C — Range Rover Sport 2023

Insurer	Plan	Premium AED	Agreed value	Excess	Rate / 1,000
GIG Gulf (AXA)	Motor Perfect	3,329	277,440	1,200	12.0
Sukoon	Gold	7,588	398,905	1,200	19.0
GIG Gulf (AXA)	Motor Perfect Plus (agency)	7,742	277,440	1,200	27.9

TPL across all personas clustered tightly (AED 830–1,050) — see "what comparators share". Add-ons (GCC cover, MEDEX, car hire, excess cashback) excluded from headline premiums.

The headline finding (and how V1 corrected V0)

From Persona A alone we nearly shipped a wrong conclusion. On the Corolla, GIG looked 3rd-cheapest on headline price but most expensive on rate-per-cover (61.9) — suggesting GIG over-charges. Running B and C corrected it: on the Patrol (15.7) and Range Rover (12.0) GIG is rate-competitive, even cheapest. The Corolla rate was a small-car / minimum-premium artefact, not a market pattern.

The real, consistent pattern is agreed value. GIG quotes the lowest or near-lowest valuation on all three cars — Corolla −42%, Patrol −18%, Range Rover −30% vs the highest market valuation. On the Range Rover that's ~AED 121,000 of cover a GIG customer wouldn't get on a write-off. GIG isn't simply "cheaper" — it's quoting less cover, which lowers the premium and creates an under-insurance exposure (and a complaint/conversion risk).

What the comparators share

A shrinking panel as value rises. Mainstream cars (Corolla, Patrol) draw 5–6 insurers; the Range Rover drew only GIG + Sukoon. Fewer underwriters compete at the top end — GIG's competitive set narrows to one or two on premium vehicles.
TPL is commoditised. Third-party premiums cluster in a tight AED 830–1,050 band across every insurer and every car — effectively rate-capped, no differentiation. The whole contest is in comprehensive.
Shared excess bands. Excess scales with value on a common convention (≈250–350 small, ≈750–1,000 mid, ≈1,200 premium) — broadly aligned across insurers, so excess is rarely the swing factor.
Shared "table-stakes" benefits. Comprehensive plans uniformly include roadside assistance, driver & passenger cover, Oman cover and fast-track claims. Differentiation lives in agency-vs-non-agency repair, car-hire, MEDEX and excess cashback — and in the agreed value.
Two-tier structure everywhere. Each insurer fields a non-agency comprehensive and a pricier agency ("Plus") variant; agency premiums run 3–5× the non-agency rate.

Rate surface — nationality & age V1.2

Nissan Patrol 2022, age 35, licence 5+ yrs, one axis varied at a time. GIG "Motor Perfect" (basic comprehensive) vs the two consistent comparators. 8 nationalities pulled live 19 Jun 2026.

Nationality (35)	GIG basic	GIG agency	Fidelity	Sukoon	Cheapest
British	1,827	—	1,957	2,867	GIG
Filipino	2,001	10,244	2,972	2,490	GIG
American	2,143	7,614	2,972	2,835	GIG
Indian	2,718	—	2,218	3,313	Fidelity
Lebanese	2,850	11,669	3,373	2,995	GIG
Pakistani	3,754	11,669	3,935	4,682	GIG
Emirati	4,235	11,669	4,656	4,745	GIG
Egyptian	4,615	11,669	3,053	5,000	Fidelity

Age (Indian, Patrol): 25 → 4,434 · 35 → 2,718 · 45 → 2,447. Young-driver loading is steep (~+63% at 25 vs 35). Agency tier (Motor Perfect Plus) runs 3–5× basic and plateaus near AED 11,670 for higher-rated nationalities.

GIG is cheapest almost everywhere — except two blind spots that matter enormously. Across eight nationalities GIG wins on price for British, Filipino, American, Lebanese, Pakistani and Emirati drivers. It loses to Fidelity for exactly two: Indian (2,718 vs 2,218) and Egyptian (4,615 vs 3,053, a 51% gap). Those are two of the largest expat populations in the UAE — so GIG isn't broadly uncompetitive, it has a targeted pricing gap against Fidelity in two of the highest-volume segments in the market. That's a precise, addressable finding: match Fidelity's Indian/Egyptian rate, or accept ceding that volume. (n=1 per cell; Camry & Land Cruiser routed to a callback page rather than instant quotes — logged for V2.)

Demographic penalty heatmap V1.3

GIG basic comprehensive vs the 3-insurer market average (GIG · Fidelity · Sukoon), Nissan Patrol 2022, age 35. Green = GIG below market (competitive); red = GIG above market (penalising the segment).

Nationality (35)	GIG	Market avg	GIG vs market
Filipino	2,001	2,488	−20%
American	2,143	2,650	−19%
British	1,827	2,217	−18%
Pakistani	3,754	4,124	−9%
Lebanese	2,850	3,073	−7%
Emirati	4,235	4,545	−7%
Indian	2,718	2,750	−1%
Egyptian	4,615	4,223	+9%

By car value (British 35, vs market avg): Corolla +6% · Patrol −18% · Range Rover −39% — GIG gets more competitive as value rises.

Versus market average, GIG penalises exactly one segment: Egyptian (+9%). For every other nationality it sits at or below market, and its position improves with car value — the opposite of a wealth penalty (for British). The "Indian penalty" only exists against Fidelity specifically (Fidelity is cheap for Indians); market-wide GIG is at par. Benchmark matters: vs the single cheapest competitor the Indian (+23%) and Egyptian (+51%) gaps look far larger. n=1 per cell; Indian wealth-gradient (across car values) pending a proxy re-run.

Indian × car value — vs full market V1.4

Indian driver, 35, licence 5+. Benchmark = average of all insurers that quoted each car (excl GIG): Fidelity, RAK, Sukoon, Al Wathba. Both metrics shown — raw premium (sticker) and AED per 1,000 of cover (like-for-like).

Car	GIG raw	Market raw	raw rank	GIG /1k	Market /1k	/1k rank
Corolla (budget)	1,900	1,826	4 / 5	65.3	47.6	dearest 5/5
Patrol (mid 4×4)	2,718	2,766	mid / 3	23.4	19.9	dearest 3/3
Range Rover (premium)	3,329	8,802	cheapest	12.0	22.1	cheapest

For Indian drivers, GIG looks mid-pack on the sticker but is the most expensive in the market per unit of cover on budget & mid cars (+37% Corolla, +18% Patrol vs the field), reversing to cheapest only on premium. The divergence between GIG's raw rank and its per-1,000 rank is the mechanism: GIG quotes a low agreed value (Corolla 29k vs the field's 30–50k), which flatters the headline premium while the underlying rate is the market's highest. So it isn't wealthy Indians being penalised — it's budget Indians on cheap cars, and the penalty is partly hidden by under-valuation. Premium cars draw a thin panel (only Sukoon quoted the Range Rover); GIG under-values it ~30%, so part of the −46% is less cover. Named majors Watania/Orient/ADNIC didn't quote these specific cells — full-panel nationality re-runs queued.

What this pricing does to conversion V1.5

Model: imagine GIG draws organic, market-representative prospects. Conversion on a comparison funnel tracks raw sticker rank; motor cover is highly price-elastic. GIG's raw gap-to-cheapest per segment → relative CVR index → weighted by each segment's share of representative UAE demand → blended vs ceiling (GIG cheapest everywhere = 1.00). Nationality dimension, Patrol.

Segment	Est. demand weight	GIG vs cheapest (raw)	Modelled rel. CVR
British	6%	cheapest	1.00
American	3%	cheapest	1.00
Filipino	9%	cheapest	1.00
Lebanese	4%	cheapest	1.00
Pakistani	16%	cheapest	1.00
Emirati	13%	cheapest	1.00
Indian	40%	+22%	0.44
Egyptian	9%	+51%	0.10

Blended market-neutral CVR ≈ 70% of GIG's ceiling (60–80% across elasticity assumptions). GIG wins the sticker on 6 of 8 nationalities, but loses the single largest segment — Indian (~40% of representative demand) — and that one segment is ~75% of the entire conversion shortfall. GIG's pricing is tuned for a profitable niche (low-risk Western, premium); on representative organic traffic that structure leaves roughly a third of winnable conversions unclaimed, almost all of it Indian. Reprice Indian to market and blended CVR recovers most of the gap in one move.

Assumptions / health warning: illustrative model, not measured CVR. Demand weights are UAE-expat-mix estimates; price→conversion elasticity is assumed (low/base/high tested); conversion is driven by raw sticker, so GIG's low agreed value helps sticker conversion while reducing cover; n = 8 nationalities on one car. Pressure-test weights against GIG's real funnel data before acting.

Visibility × conversion × demand — effective yield V1.6

Visibility = GIG's price rank in the cheapest-first results (rank 1 = seen first). Position-attention curve (1st 1.00 · 2nd 0.55 · 3rd 0.35…) × conversion index × demand weight = effective per-segment yield vs ceiling. Nationality, Patrol.

Segment	Demand wt	GIG rank	Visibility	Conversion	Combined yield
British	6%	1st	1.00	1.00	1.00
American	3%	1st	1.00	1.00	1.00
Filipino	9%	1st	1.00	1.00	1.00
Lebanese	4%	1st	1.00	1.00	1.00
Pakistani	16%	1st	1.00	1.00	1.00
Emirati	13%	1st	1.00	1.00	1.00
Indian	40%	2nd	0.55	0.44	0.24
Egyptian	9%	2nd	0.55	0.10	0.06

Blended effective yield ≈ 61% of ceiling. Visibility strips another ~9 points off the conversion-only 70%, almost all of it Indian. For Indian the effects compound: GIG is shown second (behind Fidelity) AND is less likely to be picked once seen — so it captures only ~24% of the Indian potential while Indian is ~40% of the market. It's the largest funnel leak, and it's a visibility problem and a price problem stacked. One fix — reprice Indian to lead the list — recovers both at once.

What's real vs modelled: the scrape gives GIG's price rank (visibility input) for real. The position-attention curve, conversion-vs-price, and demand weights are modelled — replace with GIG's own funnel analytics (CTR by position, CVR by price gap, real traffic mix) for true numbers. Ranks here use thin 3-insurer panels; full-panel re-runs in progress will sharpen them.

Traffic-mix scenarios — why GIG can't scale on representative traffic V1.7

Anchored on GIG's real observed CVRs (English 22% · Indian 6% · Other ~12% assumed). Blended CVR = Σ(segment traffic share × segment CVR). Each scenario shifts the mix the way a real growth move would.

Scenario	English	Indian	Other	Blended CVR
Today — English-search-led	45%	35%	20%	14.4%
Push PMax (≈2× Indian reach)	25%	55%	20%	11.2%
Exit / lose English search	20%	50%	30%	11.0%
Scale + wealthy leave Dubai	12%	68%	20%	9.1%

GIG's blended CVR is propped up by a high-intent, English-heavy mix; every realistic growth move drags it toward the Indian 6% floor. Pushing PMax (broad → representative → Indian-heavy), losing English search, or wealthy expats leaving all cut blended CVR from ~14% to ~9% — a ~37% relative collapse, ≈ +60% CAC, with pricing unchanged. GIG cannot scale volume on representative traffic without its unit economics breaking, until it fixes Indian conversion — a pricing problem. It is currently dependent on a shrinking, expensive English-search niche where it converts at 22%. Growth and CVR are in direct tension.

Real vs modelled: CVRs (22% / 6%) are GIG's actual figures. Mix percentages are illustrative scenarios — plug GIG's real channel mix + PMax projections for exact blended numbers. Note: scrape shows GIG's Indian price rank is volatile (1st–2nd, swinging on competitors' quotes), consistent with a weak 6% CVR vs a consistently rank-1 English position at 22%.

Caveats (V1)

Single run, one persona per car (n=1 per cell). Aggregator/broker pricing — may differ from insurer-direct. Agreed values set by each insurer, not pinned. A couple of A-tier premiums carried a second card figure (low-confidence, flagged for re-run). GIG & Sukoon direct funnels OTP-gated, not yet captured.

Next steps

V1.1 ~~Confirm the two Persona-A cells~~ — done (RAK 1,805 / Al Wathba 2,000). ~~Policybazaar as 2nd spine~~ — blocked: Policybazaar serves Akamai "Access Denied" to plain headless; moved to V2 behind the verified browser.
V1.2 Pinned-value mode — fix one agreed value across insurers for a true rate comparison alongside as-quoted.
V2 Browserbase verified-browser + UAE residential proxies (rotation/anti-bot); 3 personas re-run on a schedule, deltas tracked over time.
V2 GIG & Sukoon direct funnels via SIM-backed OTP capture — broker-vs-direct price delta.
V3 Widen personas (age, nationality, claims history) to map GIG's rate surface, not just three points.

GIG Motor Price-Checker · V1 · 19 Jun 2026 · Articulate (articulate-ai.work). Synthetic personas, fictional identities. Aggregator pricing; agreed values vary by insurer. Internal competitive-intelligence — not for external distribution without Belinda/Kendall sign-off.