What is source-set bias?

Standard deviation of per-engine presence scores. High bias = brand cited heavily on one engine (e.g., Perplexity) but absent on others. Low bias = distributed across engines. Distribution is more durable because each engine has different ranking priors; concentration leaves the brand exposed to one engine's algorithm change.

How does this relate to the AI SEO Audit tool?

AI SEO Audit (/tools/ai-seo-audit-free) is per-page across 6 surface dimensions (Schema, snippet, citations, linking, answers, freshness). GEO Audit is per-brand across 5 generative engines. AI SEO is the page-level fix-list; GEO is the engine-level surface measurement. Run AI SEO on your top 20 commercial pages; run GEO on the brand-level query set.

How accurate is the per-engine presence score?

Median accuracy +/- 6 points against the FORKOFF managed-engagement ledger (n=120 audits, 18-month rolling window). Generative engines re-weight on 4–8-week cycles, which means scores can shift +/- 10 points between audits. Treat the score as a 30-day directional estimate.

What does the lift plan look like in practice?

Per-engine, ordered by composite-lift-per-effort. Example: '+9 Google AI: Add FAQPage + HowTo schema, earn 3 inbound from .edu adjacency. +5 Perplexity: Strengthen primary-source citations on top 5 pages. +3 SearchGPT: Add source diversity via X + LinkedIn references.' Each move names the engine, the change, and the projected composite lift.

How quickly can a Tier-C brand reach Tier-B?

30–60 days for surface moves (schema, citations, freshness). 60–120 days for the engine to re-weight and surface the brand. Brands that work the top 3 lift-plan moves typically see +10 to +18 composite points by T+60 and another +5 to +10 by T+120.

Does the audit need a domain claim or analytics access?

No. The audit runs against public engine responses on the target query set. No GSC, no analytics, no signup. The managed engagement (FORKOFF AI Search Optimization) plugs into your CRM + GSC for closed-loop measurement.

Can I audit a competitor?

Yes. Same model, swap the domain. Useful for: identifying which engines a competitor is winning + which queries are uncontested + what schema or citation pattern correlates with their composite.

How does GEO connect to the FORKOFF AI Search Optimization service?

The audit is the kickoff artifact. Every FORKOFF AI Search Optimization engagement starts with a baseline GEO audit + a per-page AI SEO audit on the top 20 commercial pages. The 90-day delivery program targets +20 composite points on GEO and Tier-A on the top 5 pages. The free version exposes the model; the managed version applies it.

Yes. No signup, no email gate. Run as many brand audits as you want. The managed FORKOFF AI Search Optimization engagement applies the model to your portfolio with weekly reporting + closed-loop citation tracking.

FORKOFF · Tools

Are you in the Google AI Overview?

Q: Why isn't Claude or Gemini in the GEO list?

Because Claude + Gemini are chatbot products, not search engines. They surface in the AEO Checker tool (/tools/aeo-checker) which scores chatbot citation share. GEO covers the search-engine-generative surface specifically — Google + Bing + SearchGPT (ChatGPT browse mode) + Perplexity + You.com. Different products, different priors, different audit.

See what a 5-engine generative-SERP audit looks like — per-engine presence across Google AI Overview · Bing Copilot · SearchGPT · Perplexity · You.com, a weighted composite tier, and the lift-plan shape. The free tool returns the format; the managed engagement runs the actual per-query SerpAPI + Perplexity + SearchGPT calls.

Run the audit See the AI Search Optimization service →

How it works

Five generative engines decide whether the buyer ever sees the brand.

Each engine has different ranking priors. Google AI Overview rewards Schema + freshness. Perplexity rewards authority. SearchGPT rewards source diversity. The audit scores presence per engine, weights by search-volume share, and surfaces source-set bias.

Engine weights

Google AI Overview is the 32% engine. The other four split the remainder.

Weights are calibrated against 2026 search-volume share + click-flow attribution. Brands that index hard on Perplexity look strong on the per-engine bar but underweight in the composite. Google AI is the largest single share of where the buyer actually searches.

Floor

Median composite · un-optimized 2026 brands

120+

Brand audits in the calibration corpus

Enginesper engine presence

Per-engine presence bars.

0–100 per engine. Below 30 on the highest-weight engine is the priority leak. Distribution across all 5 engines is more durable than concentration on one.

Coverage5 engines × queries

Query coverage matrix.

Lit cell = brand cited by that engine on that query. The matrix shows where the brand has earned position vs where the gap is concentrated.

Biasconcentration vs distribution

Source-set bias.

Standard deviation of per-engine scores. High bias = concentrated on one engine, exposed to its re-weight cycle. Low bias = distributed, durable.

TrajectoryT+0 → T+120

Composite lift curve.

Surface moves at T+30. Engines re-weight 4–8 weeks later. Tier-D pages rebuild before re-auditing; Tier-C reach Tier-B at T+60 with the right lift plan.

The model

Three inputs in. Three outputs out.

Inputs

Input 01

Brand domain. The root domain to audit. Subdomain-level audits (blog.brand.com vs brand.com) run separately because generative engines treat them as different sources.

Input 02

Commercial query set. 10–25 queries the brand wants to surface on. Question-form + comparison + how-to queries land the highest generative-engine credit.

Input 03

Optional · target geo. US default. Audit can run on UK, EU, IN, BR locales because generative engines surface different sources by geo.

Outputs

Output 01

Per-engine presence (0–100). Five numbers: Google AI Overview · Bing Copilot · SearchGPT · Perplexity · You.com. Each rolls up cited + mentioned + absent across the query set.

Output 02

Composite score + source-set bias. Composite is weighted across engines. Bias = standard deviation of per-engine scores; high bias means concentrated citations on one engine, low bias means distributed.

Output 03

Per-engine lift plan. Each engine has different priors. Plan names the highest-leverage move per engine for the largest composite lift.

Definition

What does a GEO audit actually measure?

Google's AI Overview, Bing Copilot, SearchGPT, Perplexity, and You.com all generate synthesized answers with cited sources instead of a ten-blue-link result page. A GEO audit measures whether a brand is one of the cited sources across each engine on a target commercial query set. Returns per-engine presence (0–100), a weighted composite, source-set bias, and a prioritized lift plan. Differs from AI SEO Audit (per-page, 6-dimension surface audit) and AEO Checker (5-LLM chatbot scorecard) — GEO is the search-engine generative layer.

Scenarios

Four sample audits across brand types.

Each card is a real audit cell from the FORKOFF managed-engagement ledger. Run the live tool above on your own brand to score it.

Scenario 01

Tier-C

AI agency · pre-GEO

Established agency ranking #6 on Google for primary commercial query. Cited by Perplexity on 3 of 15 queries; absent from Google AI Overview entirely.

Composite

Google AI OverviewFail
PerplexityMixed
SearchGPTFail

Recommendation

Top priority is Google AI Overview (32% weight). Add HowTo + FAQPage schema, earn 3 inbound links from .edu / .gov / Wikipedia adjacency. Re-audit at T+45.

Scenario 02

Tier-C

Web3 protocol comparison page

Comparison page cited by Perplexity + SearchGPT, absent from Google AI Overview, weak Bing presence.

Composite

Google AI OverviewFail
PerplexityPass
SearchGPTMixed

Recommendation

Strong long-tail engine fit. Strengthen Schema + freshness signals to land Google AI Overview. The 32% weight gain is the largest unlock.

Scenario 03

Tier-D

Pre-launch SaaS

New domain, no rankings, full schema. Zero generative-engine surface. Brand mentions across 1 of 25 queries.

Composite

Google AI OverviewFail
PerplexityFail
SearchGPTFail

Recommendation

Tier-D = pre-engine. Build authority signals (case studies, primary research, citations from cited domains) before re-auditing. Pre-launch GEO is rebuild, not optimization.

Scenario 04

Tier-A

FORKOFF · baseline

forkoff.xyz · 25 commercial queries · cited on 18 by Perplexity, 14 by SearchGPT, 9 by Google AI, 12 by Bing.

Composite

Google AI OverviewMixed
PerplexityPass
SearchGPTPass

Recommendation

Defend. Push Google AI presence above 75 by earning 3 more inbound links from .edu adjacency. Re-audit quarterly.

Comparison

Why a 5-engine generative-SERP audit beats a rank tracker.

Rank trackers measure position #1–10 on the classic blue-link SERP. They are blind to the AI Overview that paraphrases the answer without sending the click. GEO Audit covers that surface.

Feature	FORKOFF GEO Audit5-engine generative-SERP audit + per-engine lift plan	AEO CheckersChatbot citation share (5-LLM)	AI SEO AuditsPer-page 6-dimension AI-native audit
Measures Google AI Overview presence			partial
Tracks Bing Copilot + SearchGPT + You.com		partial
Returns per-engine lift plan with prioritized moves		partial
Surfaces source-set bias (concentration vs distribution)
Free, no signup, no domain claim
Best for classic SERP rank monitoring	partial
Best for chatbot citation share monitoring	partial

Fit map

Who the audit is built for.

If your decision lands in the left column, the audit is the right surface. If it lands in the right, use a different tool.

Built for this

·Brands ranking on Google but losing clicks because the AI Overview answers the query without sending traffic.
·Marketing teams measuring AI search exposure and need a per-engine breakdown rather than a brand-level chatbot score.
·AI-native SaaS where the buyer asks Google + ChatGPT before clicking. The generative-SERP layer is the new shelf.
·Agencies running AI-search-optimization engagements that need a 5-engine audit artifact for the kickoff.
·Brands with classic-SEO infrastructure that want to extend coverage into the generative SERP layer without rebuilding.

Not the right fit

·Pre-rank brands. The audit assumes some classic SERP presence; sub-100-page domains usually fail every engine.
·Brand-only / navigational queries. The audit measures commercial-intent surface, not brand recall.
·Replacing classic rank tracking. Pair with Ahrefs / SEMrush; the audit covers the generative layer that they don't.
·Daily monitoring. Generative engines re-weight on 4–8-week cycles; daily re-audits are noise.
·Page-level fix-list. Use the AI SEO Audit (/tools/ai-seo-audit-free) for per-page 6-dimension scoring.

FAQ

GEO Audit. Questions answered.

Generative Engine Optimization audit. Measures whether a brand surfaces as a cited source across the 5 generative-SERP engines (Google AI Overview, Bing Copilot, SearchGPT, Perplexity, You.com) on a target commercial query set. Returns per-engine presence, weighted composite, source-set bias, and a per-engine lift plan.

SEO targets classic SERP rank position (#1–10 blue links). AEO (Answer Engine Optimization) targets chatbot citations (ChatGPT, Claude, Gemini, Grok). GEO targets the generative-SERP layer — the AI-synthesized answer at the top of search results, with cited sources. SEO is rank, AEO is chatbot, GEO is search-engine-generative. The three are complementary; a healthy 2026 brand scores in all three.

Because it sits on the highest-traffic generative surface in 2026. Google AI Overview shows on roughly 30% of commercial-intent queries in the US (rising) and consumes 60% of the click volume those queries used to send to organic results. Missing the AI Overview means missing the buyer at the moment of intent. The 32% weight is calibrated against managed-engagement attribution data on the FORKOFF ledger.

Next read

Other toolsin the FORKOFF stack.

01 / 04Open

AI Search Optimization service

The managed engagement that takes the GEO audit + AI SEO audit and runs them across the portfolio with weekly reporting.

02 / 04Open

AI SEO Audit

Per-page 6-dimension AI-native audit. Where GEO is brand-level engine surface, AI SEO Audit is page-level fix-list.

03 / 04Open

AEO Checker

Brand-level 5-LLM chatbot citation share. GEO covers the search-engine-generative layer; AEO covers the chatbot layer.

04 / 04Open

Qualified-View Auditor

Per-post engagement quality audit. Same audit-ledger philosophy applied to a single piece of content.

The brand line

Stop ranking on a SERP that hides your link.

5-engine generative audit · weighted composite · per-engine lift plan · re-audit at T+30 / T+60 / T+120.

Run an AI Search Optimization engagement

Talk to a strategist

Run the audit

From the FORKOFF blog

Receipts, deep dives, and playbooks.

Read all

ETH NYC 2026 Activation Playbook: The 3-Week Motion

ETH NYC 2026 hits June 8 to 14 (ETHConf + ETHGlobal NY). The 3-week activation playbook that lifted sponsor day-30 dev activations 8.9x across 11 audited teams.

By FORKOFF

Read