Skip to content

Build in Public

Day 119: Not Every Public Page Deserves AI Visibility

Most GEO conversations assume visibility is additive.

Publish the clearer page. Make the offer easier to cite. Improve the source. Strengthen the proof. Remove ambiguity. Help buyers and answer-led systems understand the business more accurately.

That is often the right work.

But not every public asset deserves more visibility. An expired offer page can keep attracting buyers to terms the company no longer sells. An obsolete pricing page can anchor commercial expectations after the model has changed. A superseded product document can describe a capability that now works differently. A retired event page can look like an active programme. An acquired-brand route can keep sending prospects to the wrong team. An unsupported-market page can create demand the business cannot responsibly serve.

For CMOs, Marketing Directors, and founders, this is the subtractive side of Generative Engine Optimization: visibility lifecycle management. The question is not only, “What should we make easier to find?” It is also, “Which public assets still represent a valid commercial state?”

The answer is not to delete anything old. The answer is to classify the asset, choose a proportionate state change, and be honest about propagation limits.

Day 118: Make Every Google AI Tactic Pass the Search Test

The easiest way to waste a Google AI budget is to buy a shortcut Google has not asked for.

A CMO hears a new proposal: create an AI text file, add special markup, restructure pages for AI Mode, expand schema solely for AI Overviews, or build a separate “AI SEO” layer before the site can be eligible for Google’s AI features. The proposal sounds plausible because the interface has changed. AI Overviews and AI Mode feel different from classic blue-link search results, so the procurement logic drifts: if the experience is new, perhaps the required technical layer must be new as well.

Google’s own guidance does not support that leap.

In its Search Central documentation on AI features and your website, Google says the best practices for SEO remain relevant for AI features in Search, including AI Overviews and AI Mode. It also says there are no additional requirements to appear in those features, nor other special optimisations necessary. To be eligible as a supporting link, a page must be indexed and eligible to be shown in Google Search with a snippet; Google states there are no additional technical requirements.

For CMOs, Marketing Directors, and founders, that creates a practical procurement test.

Before funding any Google AI tactic, ask whether it improves ordinary Search eligibility, usefulness, relevance, clarity, accessibility, or content quality. If it does, it may be worth doing. If it only exists because someone claims Google AI needs a separate mandatory layer, ask for the source.

Day 117: Do Not Buy a GEO Retainer Before the Decision Repeats

A monthly fee can make immature GEO work look more mature than it is.

The buyer wants seriousness. The supplier wants continuity. The dashboard wants a cadence. A retainer sounds like the grown-up answer: track the prompts, watch the surfaces, report the movement, recommend the next content batch, and keep the machine warm.

But recurring spend is only useful when there is a recurring commercial decision behind it.

For CMOs, Marketing Directors, and founders, the better procurement question is not, “Should we buy a GEO retainer?” It is, “What decision are we ready to make now, and does that decision repeat often enough to justify an ongoing workflow?”

If the problem is still uncertain, buy a bounded baseline. If a material gap has been validated and someone can change the underlying public truth, buy a focused sprint. If the same commercially meaningful AI-visibility decision keeps returning across surfaces, markets, campaigns, or buying moments, then a managed workflow may be justified.

Do not buy the recurring shape before the decision repeats.

Day 116: Weight AI Visibility by Buying Consequence, Not Prompt Count

The biggest AI-visibility gap is not always the first one worth fixing.

A dashboard can make the decision look obvious. One buyer-question family has the most prompts. Another has the most missing mentions. A third appears across the largest number of surfaces. The team sorts descending, sees the biggest bucket, and calls it the priority.

That can waste the budget.

For CMOs, Marketing Directors, and founders, the useful allocation question is not, “Where do we have the largest number of missing mentions?”

It is, “Which absence or inaccuracy changes an important commercial decision?”

A small set of procurement, qualification, risk, implementation, or retention questions may deserve attention before a much larger pool of low-stakes awareness prompts. Not because the smaller family proves more demand. It does not. It deserves attention because the consequence of a bad answer is higher, the business understands the decision affected, and the gap is something it can responsibly improve.

Day 115: Stress-Test the AI Recommendation With One Buyer Constraint at a Time

A recommendation can look strong until the buyer becomes real.

The broad question is flattering. A CMO asks for providers that can help with AI visibility, answer-led buyer research, GEO strategy, or commercial diagnosis. The answer names a suitable category. It may even include the company. The comparison sounds plausible enough to screenshot and send around the team.

Then one constraint is added.

The buyer has a limited implementation team. Or needs UK support. Or cannot buy software this quarter. Or uses a stack that changes delivery. Or has a board deadline in six weeks. Or cannot tolerate reputational risk. Suddenly the recommendation set shifts. The answer moves from specialist advisory to monitoring tools, from independent consultants to enterprise integrators, from strategic diagnosis to content production, from provider selection to an internal no-action route.

For CMOs, Marketing Directors, and founders, that is the useful test. AI visibility is commercially meaningful only if the recommendation survives the conditions that decide shortlist eligibility, or changes in a way the business can explain.

The practical GEO question is not only, “Are we recommended?”

It is, “Which single buyer constraint makes the recommendation change?”

Day 114: Decide Whether GEO Must Create the Category or Win It

A GEO programme can be funded to do two very different jobs.

In one market, buyers do not yet have stable language for the problem. They may feel the pain, but they do not know what to call it, what kind of category solves it, which alternatives are sensible, or what would make the issue worth budget. In that situation, Generative Engine Optimization has to help make the problem and the category legible.

In another market, buyers already understand the category. They are asking which provider fits, which route is safer, which option is credible, and which trade-off matters. In that situation, GEO has to help the company win comparative consideration inside an existing buying frame.

Those jobs are not interchangeable.

For CMOs, Marketing Directors, and founders, the leadership question is not only, "Are we visible in answer-led research?" It is, "Are we funding category creation, or are we funding demand capture?"

If that choice is unclear, the team can celebrate the wrong movement: provider mentions inside a category buyers are not yet using, or more educational content in a market where buyers are already choosing between suppliers.

Day 113: Test the Integration Question Before the Demo

A buyer may decide whether the demo is worth taking before they compare the features.

They do not always begin with, "Which supplier is best?" Sometimes the first useful question is smaller and more brutal: "Will this work with the stack we already have?"

For a CMO, Marketing Director, or founder, that question can decide whether good-fit demand reaches sales at all. A buyer may need to know whether an offer works with their CRM, analytics layer, data warehouse, content workflow, security model, region, implementation partner, team size, or procurement constraint. If the public answer is vague, the buyer either books a call that sales must spend correcting, delays the conversation, assumes the fit is weak, or arrives without the right context.

This is not a claim that answer engines control buyer behaviour. It is a narrower commercial point: pre-demo compatibility is an eligibility gate. If answer-led research can surface that gate, Generative Engine Optimization should test it directly and make the honest state easier to understand.

The question is not only, "Are we visible?"

It is, "When a buyer asks whether we fit their real operating environment, what answer do they receive before anyone speaks to them?"

Day 112: Your Buyer May Arrive With an AI-Written RFP

The buyer may not arrive with a blank page.

Before a supplier call, a CMO, Marketing Director, founder, or procurement lead may ask ChatGPT, Claude, Perplexity, Gemini, Google AI features, search results, comparison pages, review sites, and other public surfaces how to choose a provider. They may ask what to look for, what questions to ask, which options to compare, and which risks to avoid. By the time the first conversation happens, the buyer can already have a shortlist, a set of evaluation rules, and a view of what good should look like.

That does not mean an answer engine literally wrote the RFP. It does mean answer-led research can rehearse a buying brief before the formal brief exists.

For CMOs, Marketing Directors, and founders, that is a practical GEO problem. The commercial risk is not only absence from the shortlist. The risk is that the company is compared on criteria designed for another category, a substitute route, or a narrower version of the problem. The buyer may underweight the real buying risk, overweight a visible but secondary feature, or omit the trade-off the offer was designed around.

Generative Engine Optimization should therefore inspect the decision rules being introduced, not only the names being recommended.

Day 111: Your AI Content Workflow Needs More Than a Stop Button

A safe AI content workflow should be able to say no: refuse weak claims, catch stale angles, stop private notes becoming public copy, and preserve the caveat that Google’s AI features rely on core Search ranking and quality systems, not a magic switch called llms.txt or special AI markup.

But a stop is not an operating model.

For CMOs, Marketing Directors, and founders, the commercial problem begins one step later. If a workflow only says “do not publish”, the queue either dies or people learn to bypass the standard. A campaign waits. A sales enablement asset remains unfinished. A buyer-facing claim sits in limbo. The team loses time not because the system was cautious, but because caution produced no recoverable next state.

A mature AI-assisted content and GEO workflow needs more than a stop button. Every refusal should return a decision contract: why the output is unsafe, what evidence or commercial authority is missing, what can happen next, and which shortcut remains prohibited.

Day 110: Ask What Your GEO Baseline Refuses to Promise

A GEO baseline should be useful before it sounds ambitious.

That sounds obvious until procurement begins. The buyer asks for a diagnostic. The supplier wants the work to feel valuable. The proposal starts to stretch. It promises visibility improvement, citation opportunity, ranking movement, revenue upside, competitor displacement, or a clean score that leadership can track. The language becomes easier to approve because it sounds closer to growth.

It also becomes less credible.

For CMOs, Marketing Directors, and founders, the safer buying question is not only, “What will this baseline deliver?” It is also, “What will this baseline explicitly refuse to promise?”

A serious diagnostic can promise observable work: the question window, surfaces tested, retained captures, source labels, access limits, technical inspection, prioritised diagnosis, and a recommended next move. It should not promise deterministic answer-engine outcomes from a snapshot.