The IAB has just published its Measuring visibility in the AI era framework. It is the third best practice related guide for generative engine optimisation (GEO) measurement, following on from AMEC’s GEO Principles in June, and the PRCA’s Beyond Visibility: How to measure GEO guide in July. That is a lot of best practice for a discipline barely two years old where we don’t even have an accepted term for it. Is it generative engine optimisation (GEO), answer engine optimisation (AEO), AI visibility measurement or our preferred term: AI as a stakeholder?
The temptation for a busy chief communications officer or agency head is to wait for the dust to settle. The risk of delay is that AI answers are already shaping reputations. There is no historic data, so your benchmark is from when you start, which needs to be as soon as possible. You can adapt and improve later, but you can’t afford not to start now.
PRCA Beyond Visibility: How to Measure GEO guide
Before I compare the strengths and weaknesses of all three, I should declare an interest. I was one of six co-authors of the PRCA guide, Beyond Visibility: How to Measure GEO. The six authors were Nathalie Agnew (Muckle Media), Jonny Bentwood (Golin Ketchum), Stuart Bruce (Purposeful Relations), Alison Spray (Burson), Anna Wilson (Tangerine) and Sanjiv Winayak (Milk & Honey). Purposeful Relations is also the PRCA’s official AI partner and a member of both the PRCA and AMEC, and I am a fellow of both.
The IAB Measuring Visibility in the AI Era Framework
The IAB framework looks at GEO from the perspective of both brands (organisations) and publishers.
The authors of the IAB framework come from multiple industry sectors and include brands, advertising agencies, performance marketing, publishers and measurement providers. However, it is very much from a brand, advertising and measurement vendor perspective, and lacks any real input across a whole organisation as there is no representation from corporate communications or public relations to provide that 360-degree perspective.
It takes a simplistic approach of simply measuring AI outputs as part of generative engine optimisation (GEO). It makes no attempt to understand the complexities of AI as a stakeholder, which should be front of mind for chief communications officers and corporate affairs leaders.
The IAB framework says there are more than 20 companies that now sell AI visibility measurement. That’s a rather bizarre statement, which immediately makes me question its relevance and credibility as in reality there are at least 100 tools on the market. It implies IAB’s definition of an ‘AI visibility measurement tool’ might be quite restricted. It doesn’t say anywhere what these 20 tools might be, so we have no way of knowing whether they are the new generation of GEO tools like Profound and Peec.ai, the SEO tools that have pivoted such as SEMrush, or the new tools from specialist PR and communication tool vendors like Onclusive, Vuelio, Notified, Reptrak and Meltwater.
It appears to be mainly intended as an aid to the procurement of GEO measurement tools and it is clear that its scope is just “organic and non-paid AI visibility measurement” and not how to act on the data. It doesn’t address GEO or AEO optimisation so is only relevant to a limited part of managing AI as a stakeholder.
The framework addresses the huge challenge that Purposeful Relations explained in our Impact of Generative Relations and Communications white paper. None of the myriad of tools have access to robust data as AI companies don’t make AI prompts and AI answer data available. This means it can only ever be a directional indicator as it doesn’t have data in the same way as media monitoring, social listening and search tools do.
All of the GEO measurement tools use different prompt libraries, different platform coverage and different scoring rubrics. Two of them can measure the same brand in the same category and produce materially different answers. Neither is lying. They are measuring different things and calling them the same name.
The risk is that people take these bogus visibility metrics and give them a value they don’t deserve, much like the decades where many were fooled by advertising value equivalents (AVE). A fraudulent metric filled a vacuum, clients accepted it because something beat nothing, and it took the profession years to dislodge. The GEO vacuum is being filled right now with ‘visibility’ metrics.
The IAB Measuring visibility in the AI era framework creates four key metrics – the four Ps of AI visibility. Confidence in the guide is somewhat undermined by the incorrect greengrocer’s apostrophe in the title of one of its most important sections!
The 4 Ps of AI Visibility
Presence: Does the brand appear or is the publisher cited?
Prominence: Where and how prominently?
Portrayal: In what context, and with what accuracy?
Persuasion: Does visibility drive action?
What’s notable about all four Ps is the unanswered question: where does responsibility or ownership sit? There is an argument for all four to sit within corporate communications or public relations, but the strongest to sit with communications is with portrayal – what does it actually say and mean.
The fact that this framework is being published by the IAB shows that CCOs have a limited time window to claim responsibility. The risk if they don’t is that their organisation’s presence in AI answers is reduced to a crude metric on a martech dashboard which stores up reputational issues for the future.
What are the three frameworks, and how do they differ?
They differ less in what they measure than in who they are written to serve.
| AMEC GEO Principles | PRCA How to measure GEO | IAB Measuring visibility in the AI era | |
| Published | June 2026 | July 2026 | August 2026 |
| What it is | Best practice principles | Applied methodology | Procurement instrument |
| Written for | Measurement professionals and PR, comms and corporate comms professionals. | PR, comms and corporate affairs teams | Brands, agencies, publishers, vendors |
| Core structure | Seven principles, three evidence domains (upstream, downstream, search and content readiness) | Five groups: presence, reach, quality, behaviour, commercial | Four Ps: presence, prominence, portrayal, persuasion |
| Looks upstream? | Yes, required before interpreting outputs | Yes, as the first analytical layer | No |
| Ethical guardrail | Explicit | Explicit | Absent |
| Certification planned | No | No | Yes |
In short: AMEC says what good looks like. The PRCA says how to do it. The IAB says how to buy it.
What do all three agree on?
More than the differing vocabularies suggest. This consensus is now, in effect, the industry position.
- AI outputs are directional, not definitive. Identical prompts return different answers. A single response is a sample of one.
- No single tool sees the whole market. Any vendor claiming total AI visibility is overclaiming.
- Method must be documented and repeatable. Prompt text, platform, date, market, language and the saved output itself.
- Branded prompts flatter you. The unbranded, category-level questions people ask before they know your brand are where you find out where you actually stand.
- Set up before you need it. There is no back-search in GEO. Your baseline begins the day monitoring starts, not the day the crisis does.
That last point is the one I would put in front of a board this quarter. Every month of delay is a month of trend data you can never recover.
Where is each framework strongest, and where does it fall short?
| Framework | Strongest | Weakest |
| AMEC | Connects visibility to outcomes and impact via the Integrated Evaluation Framework; the only one that treats accuracy, risk and licence to operate as central | Principles-level abstraction; cannot be handed to a team as a method |
| PRCA | The most operationally usable; a team can run a programme from a standing start | Blanket “directional only” stance is safer but weaker at defending spend; lacks drift detection, variability baselines and any separation of hallucination from factual inaccuracy |
| IAB | Quantified thresholds, vendor disclosure schedule, model-drift protocols, and reporting language that survives a CFO | Measures only what the model says; no upstream reputation or content readiness; attribution deferred; no ethical guardrail |
The PRCA and AMEC documents are stronger than the IAB framework on meaning, but the three are more useful together than in isolation.
Why should a CCO care about the upstream information environment?
Because the alternative is a measurement programme with no diagnostic function.
The IAB framework can tell you how invisible you are, with confidence intervals. It cannot tell you why, because it never examines the earned coverage, third-party commentary, reviews, community discussion and owned content that models actually draw on. Nor does it check whether your site is crawlable, indexed and structured well enough to be cited at all.
Both AMEC and the PRCA insist you assess that environment before interpreting a single AI answer. This is not methodological purity. It is the difference between a report that names a problem and a report that identifies a remedy.
What should CCOs look for in GEO advisers?
Five things, and none of them require buying an unproven new tool.
- Ensure the adviser complies with best practice standards. If they don’t say they use the AMEC GEO Principles and PRCA guide then question their ability to offer world-class counsel.
- Interrogate the details. Ask which platforms and model versions are covered, how the prompt library was built and refreshed, whether data comes from prompt simulation or panel observation and how many responses per prompt. Where a provider will not answer, treat the silence as the answer.
- Separate hallucination from factual inaccuracy. A hallucination is fabricated by the platform. A factual inaccuracy is an accurate reflection of a source that is wrong. Same reputational damage, entirely different remedies. This is the strongest idea in the IAB document and the clearest gap in the AMEC and PRCA guidance.
- Report in tiers. A board wants objective and result. A comms team needs actionable insight. An analyst needs the full prompt and platform complexity. Sending all three the same deck fails all three.
What happens if the PR and comms profession do not move?
The IAB is working towards a vendor certification programme, developed with Microsoft, Walmart, WPP Media, Acxiom, the Alliance for Audited Media and others. Neither AMEC nor the PRCA has anything comparable.
If that certification becomes the procurement default, GEO measurement standards will be set by advertising. Reputation, accuracy and the upstream environment become optional extras rather than requirements. That is the search engine optimisation story repeating itself, only this time with better governance, which makes it considerably harder to argue against.
The answer is not to compete. It is to align deliberately, on our terms, while there is still a seat at the table.
FAQs
Q: What is Generative Engine Optimisation (GEO)?
A: GEO is the practice of improving an organisation’s visibility and reputation in AI-generated answers, similar to how SEO works for search engines. However, it is too crude to think GEO replaces or modernises SEO. PR and comms professionals need to think of AI as a stakeholder.
Q: Why are there multiple GEO measurement frameworks?
A: There aren’t three frameworks. AMEC has published a set of GEO principles. The PRCA a how to guide for GEO, and the IAB has published a framework just to help procure AI measurement tools. Each serves different audiences and purposes including measurement professionals, PR and communications teams, advertising and marketing, and procurement specialists.
Q: Are GEO metrics definitive?
A: No. All three frameworks agree that AI outputs are directional, not definitive. Identical prompts can return different answers, so results should be interpreted as trends, not absolutes.
Q: What’s the risk of not adopting GEO standards early?
A: Delaying adoption means missing out on valuable trend data and potentially letting advertising-focused standards dominate, sidelining reputation and accuracy.
Q: How should organisations choose a GEO adviser?
A: Look for advisers who follow best practice standards, are transparent about their methods, and can clearly separate hallucination from factual inaccuracy.
Sources
Sources used: IAB Measuring visibility in the AI era; AMEC GEO Principles; AMEC A Practitioner’s Guide to GEO Measurement; PRCA Beyond Visibility: How to Measure GEO.
