G·AI·G Home Report Brief Decision Tool Code

G·AI·G — the tool

One engine, ours, that measures whether GIG is named when a real buyer asks an AI. This page is about capability — not cost. GIG pays nothing; the stack is Articulate-run. What matters is what each tool lets us see.

Who. Built by Articulate, owned by the Performance function · What. An AI share-of-answer tracker with AXA→GIG attribution no off-the-shelf tool does · When. Weekly — the 4–6 week citation-decay window makes it a standing metric · How. ~250 lines of Python, no platform (see Code).

The capability stack — what each tool unlocks

Five engines plus two data layers. Each is a capability we do not have without it.

Engines we run today

Have
OpenAI · ChatGPT
The largest answer engine — and 96% of GIG's measured AI traffic. Unlocks:
  • Visibility on the surface where most UAE buyers now ask
  • Brand + 11-competitor citation detection per prompt
  • The spine of share-of-answer scoring
Have
Anthropic · Claude
The enterprise-leaning engine — GIG's strongest surface (100% in the baseline). Unlocks:
  • The high-intent professional / broker channel
  • A second independent reading to cut single-engine noise
  • Cross-engine comparison (where GIG wins vs loses by model)
Have
Serper · Google
Google SERP + AI Overview surface, UAE-localised — and the demand proxy. Unlocks:
  • The query universe — what buyers actually ask (autocomplete + People-Also-Ask)
  • Prompt-set sizing by real intent, not guesswork
  • The Google AI-Overview answer alongside the chat engines

Engines to add — the capability gap

Add
Perplexity · Sonar
The citation-native answer engine. Unlocks a capability we do not have today:
  • Source-level visibility — which URLs the AI cites, not just whether GIG is named
  • The cited-link leaderboard — whose pages win the answer (GIG vs aggregators vs rivals)
  • A direct read on the fastest-growing pure answer engine
Add
Google · Gemini
Covers Google AI Overviews — ~2.5B reach, the biggest and fastest-growing answer surface. Unlocks:
  • The single largest AI surface we are currently blind to
  • AI Mode coverage as Google rolls it out
  • Closing the gap our own GA4 already shows (Gemini = 1.5% of captured AI traffic, hugely under-measured)

Data layers — beyond the engines

Have · X read-stack
x-ingest · Apify · X API
Our own X read-stack feeds the panel-of-experts pipeline. Unlocks the voices:
  • Identify the named experts AI quotes (Avinash Babur, Stephen MacLaren, Dr Shady Ashraf…)
  • Track sentiment + themes per voice, weekly
  • See whose framing the engines adopt — and where GIG is absent
Add · GIG CDN logs
Crawler truth
GIG's own server/CDN logs. Unlocks the supply side of citation:
  • Whether GPTBot / PerplexityBot / OAI-SearchBot actually fetch GIG's pages
  • Which pages get crawled, how often — the "why" behind being cited or not
  • Proof a new page is even seen by the engines before we wait for it to rank

What it can & cannot do

Can
Measure share-of-answer across engines, by phase & line of business · attribute AXA→GIG (no tool does this) · name the rivals taking the answer · capture cited URLs + sentiment · run weekly, fully owned.
Cannot
Report true per-prompt query volume (no engine exposes it) · guarantee a citation — influence, not control · see inside a closed model's reasoning · replace a licensed consumer prompt-panel · track zero-click impact without the CDN logs above.

The ladder

R0 · now · 5/10
Baseline
3 engines, single-shot, mention-only, AXA-aware, weekly. 83% baseline live.
R1 · reliable · 6.5/10
+ Perplexity + Gemini
n≥3 sampling (confidence bands), cited-URL parsing. Defensible for a board read.
R2 · target · 7.5/10
Complete
+ demand proxy, + CDN crawler logs, + sentiment & the voices. The full operational picture.

R3 productised (multi-client, live dashboard) and R4 frontier (licensed demand panel + agentic content fixes) follow — not needed for the GIG engagement. Recommended stop: R1 → R2.