
For two years, every Search Central conversation about AI Overviews has hovered around the same phrase: follow our existing guidance. Helpful content. E-E-A-T. Crawlable HTML. Structured data when it fits. Cite reliable sources. The implication was that nothing had changed — generative AI was just another surface drinking from the same well.
That implication just got an asterisk. On May 15, 2026, Google published its first dedicated optimization guide for generative AI features — covering both AI Overviews and AI Mode, according to Search Engine Land's coverage of the release. The new Google Search Central documentation doesn't introduce a new ranking algorithm and it doesn't promise inclusion. What it does is consolidate, for the first time, the specific behaviors that make a page usable by Google's generative systems versus merely indexable.
This post is the small-business translation. It pulls the guide's recommendations out of Search Central voice and into eight concrete actions you can finish in a working month. It also flags — clearly — what the guide does not promise, because reading the document carefully is as valuable as following it. This is the 2026 companion to our broader Answer Engine Optimization guide, and it pairs with our coverage of Google's spam policy update for AI-generated content — together they form the “do this / don't do this” set for the generative search era.
Key Takeaways
- Google published its first dedicated optimization guide for AI Overviews and AI Mode on May 15, 2026, consolidating prior guidance into one document.
- The guide explicitly extends — not replaces — existing helpful-content, E-E-A-T, and structured-data policies.
- Eight concrete actions cover most of what a small business should do this month: be genuinely helpful, structure key facts, tighten entity data, support claims with sources, ship structured data, prove freshness, optimize images, and audit AI Overview triggers in Search Console.
- Google does not promise inclusion in AI Overviews, does not publish a ranking algorithm specific to AI features, and does not treat the guide as a replacement for E-E-A-T.
- Read alongside the spam-policy clarification: optimization guidance and policy enforcement are now explicitly two sides of one strategy.

What did Google's new guide actually say?
Strip the Search Central formality and the guide's argument is small and direct: if you want your content to be useful to AI Overviews and AI Mode, write content that humans find useful, then make the specific facts on the page machine-readable, attributable, and fresh. That is the entire spine of the document.
Three pieces stand out as new or newly explicit. First, Google connects its AI features documentation directly to the long-standing helpful, reliable, people-first content guidance — it is now the named foundation for generative inclusion, not just a tradition. Second, the guide leans noticeably harder on structured data, telling site owners that supported schema types help Google understand the specific facts on a page, which is the level of granularity AI Overviews quote from. Third, the guide ties the optimization story to Google's spam policy clarification from the same day: optimization is encouraged, but content engineered to game generative responses falls under the same spam rules as traditional manipulation.
For a small business, the practical read is that the rules of the road did not change — but the road got signs. Google is no longer asking you to infer what AI Overviews want from a patchwork of E-E-A-T language and structured-data docs. The guide names the behaviors directly, and treats them as a checklist rather than a philosophy. Our analysis in Why your content doesn't appear in AI Overviews holds: ranking on page one is necessary but not sufficient. The new guide explains, in Google's own voice, what the “sufficient” part looks like.
It is also worth saying what kind of document this is. It is published under developers.google.com/search/docs as part of the Appearance section — the same shelf as featured snippets, rich results, and sitelinks. That placement matters. Google is treating AI features as a Search appearance surface, not a separate product, which means the optimization burden lands on the same SEO/content teams already doing the work.

The 8 actions a small business should run this month
The guide implies a worklist; here is that worklist made explicit. Each item is one we believe a 1-to-25-person business can finish — or get visibly under way — inside thirty days. Where the action maps to an existing Search Central page, the link goes to that page rather than a Button Block summary.
1. Be genuinely helpful — on the page, not just in the meta description
The first sentence of the guide is the same first sentence as the helpful-content document: write content for people, not search engines. For AI features, this matters more, not less. Generative systems are trained to identify and quote the part of the page that answers the question. If your page buries the answer below three hundred words of throat-clearing, the AI either skips you or quotes a paragraph that misrepresents you. Lead with the answer; explain after.
2. Structure the key facts where AI can find them
Generative responses are sourced sentence-by-sentence. A page about furnace replacement cost in Allen County needs the specific number — a range, a service area, a year — sitting in a clean paragraph, not buried inside a video transcript or a PDF. Tables, definition lists, and bolded short-answer sentences all help. So does putting the answer in the same paragraph as the question. We covered this pattern in more depth in our writeup of Google's agentic engine optimization playbook.
3. Keep NAP and entity data crisp
The guide repeatedly references “trusted information about your business.” For local businesses, that means your name, address, phone, hours, service area, license details, and ownership should be identical across your homepage, contact page, Google Business Profile, and structured data. Where they disagree, AI systems hedge — and a hedged answer is one that names a competitor instead of you. We treat this as a recurring audit; the Bland Tax: how small businesses escape AI search sameness post explains why entity clarity compounds.
4. Support claims with sources
The guide is explicit: AI features prefer content that is verifiable. For a small-business blog, that means every statistic, comparison, and claim that could be checked should have a source — either an inline link to a primary source or a clear “in our experience” caveat when it is your own observation. We recommend treating sourcing as a non-negotiable: if you cannot cite, soften the claim or remove it.
5. Ship structured data for the content types you publish
Schema is the part of the guide most readers will skip. Don't. Google's structured-data overview lists every supported type. For most small businesses, the meaningful subset is Organization or LocalBusiness, Article or BlogPosting, FAQPage where you have real Q&A, HowTo for genuine step-by-step content, and Product for any e-commerce surface. Use the types that match your content. Do not, as the spam-policy update warns, ship FAQPage schema for a list of questions you wrote to pad the markup.
6. Prove freshness with dateModified — and actually edit the page
dateModified is one of the most-abused fields on the web. The guide does not change that, but the spam-policy clarification does: updating only the date is not a freshness signal, it is a manipulation signal. Use the field honestly — when you make a substantive edit, update the date and add a one-line revision note. When you don't, leave it alone. Pages that age gracefully and update with intent get treated better than pages that pretend to be eternally new.
7. Optimize images for AI captioning
This one is quieter and easier to miss. Generative systems are increasingly multimodal; AI Mode in particular reads the images on your page and considers them as content. That means descriptive alt text, file names that match the subject, and images that actually illustrate the claim (not a generic stock photo of a laptop). For a service business, a real photo of a real job is worth a dozen stock images.
8. Audit AI Overview triggers inside Search Console
You can already see — imperfectly — which of your queries are triggering AI Overviews. Search Console shows AI-feature impressions and clicks within standard performance reports, even though the labeling continues to evolve. Pull the queries where you appear in AI Overviews, look at which pages are cited, and treat that subset as your priority list for the actions above. If a query is showing AI Overviews and your page is not cited, that is the gap to close.

What the guide does not promise
This section matters more than the eight actions, because it protects you from agencies and tools that will read the guide as a contract. It is not a contract. The guide is a description of behaviors Google's systems prefer, not a guarantee of placement, and Google goes out of its way to say so.
The guide does not guarantee inclusion in AI Overviews. Following every recommendation perfectly is a precondition, not a result. Google's generative systems decide which queries to answer with AI Overviews on a per-query basis, and they decide which sources to cite using a combination of relevance, authority, and the freshness signals discussed above. Many queries never trigger AI Overviews at all. Many of those that do change their cited sources over time, sometimes weekly. Search Engine Land's diagnostic coverage of the AI Overviews citation gap captures the volatility well.
The guide does not publish an AI-features ranking algorithm. This is worth repeating because it is the question every client asks first. There is no “AI Overviews score,” no separate index, and no documented ranking factor with weights. The systems use the same underlying understanding of the web that Search does, applied through generative interfaces. The optimization work is therefore one funnel, not two.
The guide does not replace E-E-A-T. The Google Quality Rater Guidelines remain the canonical document for what Google considers experience, expertise, authoritativeness, and trustworthiness. AI features inherit those criteria. The new guide layers on specific implementation recommendations; it does not soften the bar for who is allowed to be cited. Topical authority and trust signals stay load-bearing even as the optimization surface shifts.
The guide does not promise that AI features will continue in their current form. Google has reorganized its AI surfaces twice in two years. AI Mode launched, expanded, and was repositioned. AI Overviews tightened its citation behavior after early backlash. The guide is a snapshot of preferred behaviors as of May 2026; it will change. Treat the eight actions above as fundamentals — they would still be good advice if Google rebranded AI Mode tomorrow — rather than as a moat.
How this pairs with the spam policy update
Google published two documents on May 15, 2026, and they should be read together. The optimization guide tells you what to do. The spam-policy clarification tells you what counts as overreach. The line between the two is, in Google's framing, intent.
Adding FAQPage schema to a genuine FAQ is optimization. Wrapping a list of synthetic questions in FAQPage markup so it gets quoted in AI Overviews is spam. Updating dateModified after a substantive rewrite is optimization. Updating it weekly without changing the content is spam. Citing a real source you read is optimization. Inventing an analyst quote and attributing it to a real firm is — and this is now explicit — both deceptive and a spam-policy violation enforceable in the AI surfaces, not just the blue links.
That symmetry is the most useful part of the May 15 release. Our coverage of the spam-policy update walks through the specific behaviors Google is now enforcing against generative-AI responses. Reading the two posts as a pair gives you both halves: the affirmative checklist and the negative one.

A small-business 30-day rollout
The eight actions are not equally urgent for every business. We recommend the following sequence for a 1-to-25-person business that publishes content occasionally and has a Google Business Profile.
| Week | Focus | Outcome |
|---|---|---|
| Week 1 | Audit NAP and entity data across homepage, Contact, GBP, and existing schema. Reconcile every disagreement. | One source of truth for who you are. |
| Week 2 | Add or fix Organization/LocalBusiness schema. Add FAQPage schema only where you have real Q&A. Ship Article schema on the blog. | Machine-readable facts where AI can find them. |
| Week 3 | Rewrite the top three pages (by GSC impressions) to lead with the answer. Source every claim. Update dateModified honestly. | Pages that hold up as a quoted paragraph. |
| Week 4 | Pull GSC AI-feature impressions. Identify the queries triggering AI Overviews where you are not cited. Pick three to address next month. | A real measurement loop, not a vibe. |
We use a version of this sequence with our own clients. It is intentionally unambitious — the goal of the first month is to remove blockers, not to chase every recommendation in the guide. Our Information gain audits post covers what comes after, once the cleanup is finished and the question is what unique value can only this business publish.
What we'd recommend for Button Block clients
If your business already has clean structured data, consistent NAP, and a habit of citing sources, the new guide is reassurance more than a roadmap. Keep doing what you are doing, audit the eight actions quarterly, and watch the GSC AI-features data as it matures.
If your business has none of those — a marketing site built five years ago, no schema, a Google Business Profile that disagrees with your homepage on closing time — start with Week 1 above. We help small businesses through this kind of cleanup as part of our Answer Engine Optimization engagements; we are also happy to point you at the right Search Central pages and let you run it in-house. Either way, the trap to avoid is treating the new guide as a magic incantation. It is a description of competent behavior. Most of the work is doing competent things consistently for long enough that Google's systems notice.
If you would like a second pair of eyes on your AI Overviews readiness, reach out. We will tell you honestly which of the eight actions you have already done, which can wait, and which one is costing you citations right now.

Want help running the 30-day rollout on your site?
Button Block's Answer Engine Optimization engagements start with the exact rollout above: NAP audit, schema cleanup, page rewrites, and a Search Console measurement loop. We'll tell you honestly which of the eight actions are urgent, which can wait, and which one is costing you citations right now.
Frequently Asked Questions
- Is Google's new AI features guide a ranking algorithm?
- No. The guide is a description of optimization behaviors that help Google's generative systems understand and use your content. It does not document a separate ranking algorithm, and Google explicitly says following the guide does not guarantee inclusion in AI Overviews or AI Mode. Treat it as a checklist of fundamentals, not a formula.
- Do I need to add FAQPage schema to every page now?
- No, and you should not. The guide encourages structured data that matches your content. Adding FAQPage schema to a page that does not contain a genuine FAQ falls under Google's May 15, 2026 spam-policy clarification and risks being treated as deceptive markup. Ship FAQPage only where the questions and answers are real.
- How is this different from the existing helpful-content guidance?
- The new guide is the first time Google has consolidated AI-features optimization into a single document, but its foundation is the existing helpful-content guidance. Where it goes further is in tying specific implementation details — structured data, freshness signals, image optimization, Search Console diagnostics — directly to AI Overview and AI Mode inclusion behavior.
- Will following the 8 actions get my site cited in AI Overviews?
- Maybe. The actions raise the probability your content can be used by AI features when a relevant query triggers them; they do not guarantee citation. Many queries never trigger AI Overviews. Many that do change their cited sources over time. Treat the eight actions as removing reasons to exclude you, not as a guaranteed path to inclusion.
- How often should we audit our AI Overview readiness?
- We recommend a quarterly audit cadence for small businesses that publish content regularly, and a monthly audit for businesses in fast-moving verticals — healthcare, legal, finance, or anywhere AI Overviews are visibly active for your core queries. The audit should cover all eight actions plus a fresh pull of Search Console AI-feature impressions to see which queries have started triggering AI features since the last audit.
- What if my industry's AI Overviews show competitors with worse content than mine?
- This is common, and it usually means your competitors have either cleaner entity data, more structured-data coverage, or more linked citations to their pages — even when the human-readable content is weaker. Audit the eight actions on the cited competitor's pages, then close the gap on whichever signal you lack.
Sources & Further Reading
- Search Engine Land: searchengineland.com/google-publishes-guide-on-optimizing-for-generative-ai-features-477671 — Google publishes guide on optimizing for generative AI features (2026-05-15).
- Google Search Central: developers.google.com/search/docs/appearance/ai-features — Helping Google understand and use your content for AI features.
- Google Search Central: developers.google.com/search/docs/fundamentals/creating-helpful-content — Creating helpful, reliable, people-first content.
- Google Search Central: developers.google.com/search/docs/appearance/structured-data — Introduction to structured data markup in Google Search.
- Search Engine Land: searchengineland.com/google-updates-search-spam-policies-to-clarify-it-applies-to-generative-ai-responses-477657 — Google updates search spam policies for generative AI responses (2026-05-15).
- Search Engine Land: searchengineland.com/why-content-doesnt-appear-in-ai-overviews-473325 — Why your content doesn't appear in AI Overviews (2026-04-02).
- Search Engine Land: searchengineland.com/agentic-engine-optimization-google-ai-director-474358 — Agentic engine optimization: Google AI director outlines new content playbook (2026-04-15).
- Google: services.google.com/fh/files/misc/hsw-sqrg.pdf — Search Quality Rater Guidelines.
