Generative Engine Optimisation (GEO) & AI Search Optimisation (ASO)
Be the business AI recommends — AI Knowledge Signal
Be seen by customers in AI search, don't get filtered out.
AI systems are now the primary interface through which people access knowledge. The organisations that shape how they are represented inside those systems will define their industries.
What is GEO?
Generative Engine Optimisation (GEO) — also called AI Search Optimisation (ASO) — is the discipline of engineering how your organisation is represented in AI systems.
AI systems do not discover organisations through search ranking alone. They ingest, classify, and cite content based on structural, epistemic, and technical properties — properties most organisations have never had reason to optimise.
The AI systems your customers are asking
See how it works
Traditional search · worldwide
Google still dominates the open web
~90% of all search engine queries still run through Google.
Update: across 300,000 keywords, the top-ranking page’s click-through rate was 58% lower when an AI Overview appeared, up from 34.5% in April 2025 (Ahrefs, February 2026). This is a correlation, not a controlled effect: keywords without an AI Overview also lost clicks over the same period.
Note: a full-year analysis of 53 brands and 2.43 billion impressions (Seer Interactive, April 2026) found the AIO-related CTR decline is levelling off rather than continuing to fall — recalibrate against Q1 2026 actuals before setting H2 traffic targets. That finding covers AI Overviews; AI Mode is a separate surface, and in October 2026 Reddit reported search referrals down by more than 1 million a week on average, attributing the fall mainly to AI Overviews and AI Mode (CNBC, 29 Oct 2026).
AI assistant app usage · worldwide · end of May 2026
Measured by active users, ChatGPT has fallen below 50% for the first time as Gemini and Claude surge
The AI system answer leaderboard
ChatGPT46.4%
Google Gemini27.7%
Claude10.3%
Share of active mobile-app users. Grok, Perplexity, Copilot and every other assistant each hold less than 5% (roughly 15.6% combined).
ChatGPT still leads on raw scale (1.1B monthly mobile-app users to Gemini's 662M and Claude's 245M, Sensor Tower, May 2026; OpenAI itself reports 1B+ weekly active users, September 2026) — but the answer layer is no longer a one-engine race. Being cited across all of them is what now matters.
These are the systems answering your customers’ questions
Each one decides for itself what to read, believe, and repeat — so visibility has to be earned in all of them, not just one.
A single cross-engine strategy is now a documented liability — the Framework structures your content for each engine's distinct preferences, informed by Goodie’s June 2026 analysis of more than one million AI answers.
AI chatbot referrals · worldwide
ChatGPT still leads the traffic it sends back
Of the visits AI sends to the open web, ChatGPT leads by a wide margin — Gemini is the clear second, ahead of Perplexity.
ChatGPT’s referral lead dipped, then recovered — Gemini is the steady gainer
ChatGPT’s share of AI referrals fell to 76.85% in April 2026, then rebounded to 80.48% by September. Gemini has climbed from 2.31% (April 2025) to 10.91%.
What AI systems actually cite — and it isn’t the Google leaderboard.
AI citations no longer mirror Google rankings — 28.3% of ChatGPT’s most-cited pages have zero organic visibility (Ahrefs).
Zero-click is accelerating — AI Overviews reduce clicks to top-ranking content (see the Pew figures on slide 5).
Fresh content wins — 89.7% of identifiable ChatGPT-cited pages were updated in 2025 (Ahrefs).
YouTube is the strongest AI visibility signal — correlation of ~0.737 between YouTube brand mentions and AI visibility (Ahrefs, 75,000 brands).
“Best X” listicles dominate discovery — 43.8% of ChatGPT source page types for top-of-funnel queries (Ahrefs, 26,283 source URLs).
Sources: Ahrefs research articles linked above. Figures are from separate studies, each with its own method and dates.
Time to citation · ChatGPT & Claude
How long does it take AI to cite a new page?
Among new pages that ChatGPT or Claude cited, half were cited within 6.8 days and 90% within about 37 days. So judge results at day 30, not at week three.
Source: Profound agent-log data (ChatGPT + Claude, March–May 2026; n ≈ 900 new pages cited within a 60-day window), published in LinkedIn — Unlocking AI Search Visibility: The B2B Marketer’s Guide to LinkedIn, October 2026. Redrawn from the published median, P75 and P90, with straight lines joining them. Counts only pages that were cited; pages never cited are not in the sample. Vendor-reported.
The method
AI Knowledge Signal Framework
A structured methodology for engineering how organisations are represented in AI systems.
Crawlable, clear, original, credible, usable — then maintained
Each phase feeds the next, and the loop keeps running as the AI systems change.
Self-assessment
How AI Systems Evaluate Your Organisation
✓Crawlability & Technical Access — Can AI systems reach, access, and ingest your content?
✓Structural Clarity & Machine Readability — Is your content shaped so machines understand what your organisation actually is?
✓Knowledge Uniqueness & Contribution — Are you contributing original knowledge, or just repeating what already exists?
✓Authority, Evidence & Trust — Are your claims credible, consistent, and verifiable across the web?
✓AI Usability & Retrieval Readiness — Can AI systems actually use, extract, and cite your content?
✓Maintenance, Monitoring & Improvement — Are competitors being cited instead of you, and are you improving over time?
The measurement
Our tool measures you before and after
We use the AI Knowledge Signal tool to score your pages and record how the big AI models describe you — before we start and again after the work. That re-measure is what the second 50% of the fee depends on.
The result
Then you fix it — and watch the score move
Rewrite with the framework and score again. This page went from 1/10 to 7/10.
Answer a few quick questions and we'll follow up with where to focus first — no obligation.
Free Audit
Audit your AI Knowledge Signal
Results in under 30 seconds
✓Enter any public URL.
✓The audit runs a technical pre-analysis — checking schema markup, heading structure, canonical tags, and metadata — then scores your content across five AI training readiness dimensions: Crawlability & Technical Access, Structural Clarity & Machine Readability, Knowledge Uniqueness & Contribution, Authority, Evidence & Trust, and AI Usability & Retrieval Readiness.
✓Returns a Corpus Survival Likelihood rating and a structured report with Framework phase priorities. Corpus Survival Likelihood is the estimated probability that a piece of content survives AI training data quality filters rather than being removed or down-weighted before model training.
One free audit per URL each month — re-audit any time to track your improvement. We'll email you the full report, and you can print or share it too. No spam; unsubscribe any time.
Audit blocked
What this means for GEO/ASO:
Suggested fix:
Need help fixing this?
This is exactly the kind of problem we help organisations resolve. The fact that your site can't be audited is itself a finding — and one that's costing you AI visibility right now.
We'll send a branded report with your full audit and the priority fixes. One email, no list.
Report sent. Check your inbox — if it doesn't arrive within a minute, check your spam or junk folder. Share this link:
Free audit — partial analysis. This audit covers technical signals and the first 9,000 characters (~1,500 words) of your page — see FAQ for full methodology details. To score full-length content or unpublished drafts, use the Chrome and Edge extension and the framework which provides full suite of 170+ levers aligned to the AI Signal Publication 6-Phase Framework.
out of 10
This audit measures the readiness of the content you own. Third-party surfaces — Wikipedia, Reddit, review sites, news — shape AI visibility to a comparable degree (Goodie, June 2026 prompt-output study), and are assessed separately.
Dimension Scores (5 = publication-grade)
Strengths
Weaknesses
Top Recommendation
How to fix these issues
Your priority framework phases are readyBook a discovery call to walk through the exact phases to address — and get the full framework by email.
Your platform-specific fix list is ready. Book a $99.99 discovery call at aiknowledgesignal.io to get the full framework and walk through the exact steps.
Step 2
Do AI systems actually know you?
Step 1 scored your page. This asks two AI systems who you are — Perplexity, which searches the live web, and Claude, which answers from what it learned in training. Being findable and being known are not the same thing.
We’ll use the email you entered above. One free check per brand each month.
Research basis: GEO: Generative Engine Optimization, Aggarwal et al., KDD 2024 (arXiv 2311.09735) — up to ~40% visibility uplift from citations, quotations and statistics · Navigating the Shift: Web Search and Generative AI Response Generation, Chen et al. (2026) · Evaluating Verifiability in Generative Search Engines, Liu, Zhang & Liang (2023)
How this audit works: see aiknowledgesignal.io/faq for the methodology, scoring, and dimensions.
The audit tool itself is a worked example of GEO optimisation in practice — content structured so AI systems can ingest, cite, and accurately represent it.
Your buyers build their shortlist inside AI, before they reach your site
For professional questions, LinkedIn is where AI systems most often look. What gets cited is named experts with something specific to say, not company pages repeating the brochure.
94%
of B2B buyers use generative AI in their research.
6sense, 2025
#1
LinkedIn is the most-cited domain for professional queries in AI search.
Profound, 2026
75%
of LinkedIn citations come from individual member profiles. Company Pages account for 25%.
Meltwater, 2026
6.8 days
median time to first citation, among new pages that ChatGPT or Claude went on to cite. 90% of those were first cited within about 37 days.
Profound agent logs, ~900 pages, Mar–May 2026
What to do, mapped to the framework
4.1Publish under named experts, not just the brand. Give the people behind the work complete profiles, and link the site and LinkedIn in both directions. Authorship guide →
3.2Say something only you could say. The pattern you spotted first, the customer story that reframes an assumption. Structure helps, but it does not get you cited on its own. Contribution guide →
2.1Answer first. Open each section with a short, direct answer, then expand. Direct-answer guide →
4.2One anchor article, several posts. Publish one in-depth article as the canonical source, then three to five short posts that each carry one idea back to it. Hub-and-spoke guide →
6.3Track prompts, not keywords, and give it 30 days. Measure share of voice, sentiment and which sources get cited. Don't judge the programme at week three. Monitoring guide →
Summarised from LinkedIn’s Unlocking AI Search Visibility: The B2B Marketer’s Guide to LinkedIn (October 2026). LinkedIn is writing about its own platform, and most of the figures above come from companies that sell AI-visibility tracking. Treat them as directional. Meltwater’s wider data (9.5M citations, six AI platforms) ranks LinkedIn second to YouTube, so check your prompts before shifting budget. The extension’s Prompt Insights tracks share of voice, sentiment and cited sources for the prompts you choose.
Contact Form
Answer a few quick questions and we'll follow up with where to focus first — no obligation.
How we compare
Pay for results, not retainers
The big SEO suites bolt AI tracking onto enterprise plans, and the specialists are agencies on retainer. AI Knowledge Signal gives the framework away, and when we implement it for you, half the fee is paid only once your brand measurably improves in AI answers.
Brand Radar add-on, roughly $828–$1,347 / mo depending on engines covered
Yes
No
Semrush
SEO suite with AI visibility (Semrush One)
Mention / share-of-voice tracking
No
AI visibility included in Semrush One, from $199 / mo
Yes
No
Peec AI
GEO / AI-visibility analytics
Mention + source-gap tracking
PartialFlags YouTube as a citation source only
From roughly €70–€89 / mo (EUR, geo-priced since Aug 2026)
Yes
No
Onely
Technical-SEO / AI-search agency
Service-delivered
No
Custom retainer (~$2,500+ / mo)
No
No
iPullRank
SEO / GEO agency (Relevance Engineering)
Service-delivered
No
Custom / enterprise quote
No
No
Competitor features and pricing as publicly listed or reported, October 2026, and subject to change; ranges depend on plan, region and engines covered. Euro prices are shown in EUR. Confirm current pricing with each vendor.
The offer
The framework is free to start. You pay us when it works.
$99.99discovery call
Includes the full framework, emailed when you book. If you want our help implementing it, the work is billed hourly: you pay 50% upfront, and the other 50% only when a re-measure shows your brand has improved in AI answers. The How-To Guides and the audit stay free.
