A local search now ends in two places.
The map pack, and the answer above it.
The profile still has to win proximity and prominence. But the sentence a customer reads before they ever scroll to the map is now assembled by a model — and it cites whoever it can verify. Local SEO that only optimizes the profile is optimizing half the result.
Built for contractors and multi-location home services operators — the businesses whose next job comes from a homeowner three miles away who has never heard of them. Single truck or national franchise system, the local work decides who gets called.
Local SEO didn't get replaced. It got a second surface to win.
Everything that made a business rank in the map pack still makes it rank in the map pack. What changed is that a generated answer now sits above the map, assembled from sources a model retrieved and checked — and the business that gets named in that sentence is not automatically the business ranked first below it.
Here is the mechanic, in Google's words rather than an agency's. Its generative features rely on retrieval-augmented generation — also called grounding — which uses the core Search ranking systems to retrieve relevant, current pages from the index, then reviews the specific information on those pages to build a response with clickable links back to the sources that support it. Alongside it runs query fan-out: the model issues a set of concurrent related queries behind the one the customer actually typed.
For a local business, that second mechanic is the one worth sitting with. A customer asks a model which company to call for a failed water heater. Behind that single question, the system may be resolving several — who services that equipment, what the work involves, who is licensed, what other people reported. Your business has to be retrievable and corroborated across all of them, not just ranked for the phrase the customer typed.
Google is the only one of these systems that documents how it works. Gemini and Google's AI Overviews and AI Mode draw on the same Search index described above, which is why the fundamentals move both at once. ChatGPT, Perplexity, Claude and Microsoft Copilot publish no equivalent account of how they choose which local business to name, and Copilot is not even reading the same index — it is built on Bing. Everything this page says about those four is therefore observation, not documentation, and is labeled that way wherever it appears.
This is why a business can hold the top map position and still be absent from the paragraph above it, and why a competitor with a thinner profile but a deeper, better-corroborated web presence gets named instead. They are two different retrievals answering two different questions.
What has not changed is the foundation. Google states plainly that there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary. The work that earns AI citation is the same work that earns rankings — done to a standard most local businesses never reached in the first place.
The tactics didn't change. The standard they have to meet did.
Every line in the left column below is real local SEO work that produced real results. None of it is wrong. It is simply calibrated to a result page where being ranked was the same thing as being read.
Local SEO as a placement exercise
- Citations by volume. Submit to as many directories as possible; count the listings.
- Reviews as a star average. Chase the rating number, treat the review text as sentiment rather than as content.
- City pages at scale. One near-duplicate page per town, differing mainly by the name of the town.
- Keywords in the profile name. Append services to the business name to match more queries.
- Rank tracking as the report. A position for a phrase, screenshotted monthly.
- Content written for the crawler. Length and density targets, written to be indexed rather than read.
Local SEO as a verification exercise
- Citations by agreement. Fewer, authoritative listings that state identical facts, because contradiction is what breaks a model's confidence.
- Reviews as retrievable evidence. Volume and rating still count for prominence; the words customers use about specific services are what a model can actually quote.
- Locations that reflect operations. Real service areas, real premises, pages that could only have been written by someone who works there.
- Profile accuracy as the asset. The real-world name, the correct primary category, complete detail — the things Google names as relevance signals.
- Citation share as the report. Which surfaces name you, which name a competitor, and what changed.
- Content written to be quoted. Direct answers a person is satisfied by and a model can lift without distorting — the standard Google sets for helpful, people-first content.
Three factors decide local results. One of them you cannot touch.
Most agency pages on this subject paraphrase a paraphrase. So here is the primary source, read closely — including a wording detail almost nobody reports correctly.
Relevance
Relevance is how well a Business Profile matches what someone is searching for, and Google's stated remedy is to provide complete and detailed business information so it can understand the business and match it to relevant searches. That is the whole instruction. There is no trick beneath it.
- A primary category chosen for breadth rather than accuracy, diluting every query it might have matched
- Services left undescribed, so the profile can only match the category name
- Hours, attributes and detail fields left partially filled for years
- A profile name padded with service keywords — Google requires the name reflect the real-world name used on signage and stationery
Distance
Distance is how far each business is from the customer who is searching, and where the customer has not shared a location, Google uses what it knows about their location. There is no configuration that shortens a distance. Any agency selling a lever here is selling something that does not exist.
- Declaring service areas truthfully — Google's rule is that a service-area business gets one profile for the whole area it serves, not one per town
- Removing the address entirely when customers are not served at it, which Google explicitly instructs
- Understanding that a virtual office or an unstaffed co-working desk is not eligible for a profile at all
- Building genuine prominence, which is what lets a business compete beyond its immediate radius
Popularity
Nearly every article on this topic states that Google's three local factors are relevance, distance and prominence. Google's own help page opens by naming them as relevance, distance and popularity — and then uses Prominence as the heading for the third. Both words are on the page. Only one of them is in the sentence that defines the model.
The distinction is small and it is the point. An industry that has repeated the subhead for years without reading the sentence above it is an industry working from summaries. We read the primary source.
- Prominence means how well known a business is, and prominent places are more likely to appear
- It is based partly on how many websites link to the business and how many reviews it has
- More reviews and positive ratings can help local ranking
- Google states there is no way to request or pay for a better local ranking, and keeps algorithm detail confidential deliberately
The OMNIVIZ™ framework, applied to a local footprint.
OMNIVIZ™ is the operational framework behind every Allegiant engagement. Local work does not get a different framework — it gets the same five pillars calibrated to a business whose customers are geographically bounded and whose reputation is public.
A local business is an entity with a name, a place, a set of services and a reputation. Every system that might cite it — Google's local index, its generative features, an independent model — is trying to resolve those attributes into one confident record. Contradiction is the enemy of confidence. A suite number present in three places and absent in nine, a legal name on the license that differs from the name on the van, a phone number that changed in 2023 and survives in forty listings: each is a small reason for a system to hedge.
This pillar is unglamorous reconciliation work, and it is the highest-yield thing most local businesses can do. Google's guidance is to represent the business as it is consistently represented and recognized in the real world across signage, stationery and other branding — and to keep the address or service area accurate and precise.
- A single canonical fact set for name, address, phone, hours and services
- Contradiction audit across every listing that names the business
- Correct primary category and defensible secondary set
- Organization and LocalBusiness markup that agrees with the profile
- Real-world name compliance against Google's naming rules
- Eligibility review before any premises is listed
Google's guidance for its generative features is blunt about what wins: unique, non-commodity content with a point of view that could only come from experience. Its own example of commodity content is a listicle of generic tips; its example of the opposite is a specific, first-hand account of a decision and its consequences. For a local operator that distinction is a gift, because the experience is the one thing a competitor cannot copy and a model cannot generate.
The practical form is direct answers placed where the question is asked, written to satisfy a reader first. What a failed inspection actually costs to remediate. Why the cheaper unit is the right call in one situation and the wrong call in another. What the permit timeline in a dense municipality really looks like against the published one.
- Service pages that answer the decision, not the keyword
- Question-level content mapped to real customer language
- First-hand operational detail no competitor can restate
- Answers structured to be quotable without losing meaning
- Review language treated as content input
- Location content that reflects genuine operations
The old citation playbook optimized for count. A retrieval system does not count; it corroborates. When several independent, credible sources state the same facts about a business, confidence rises. When they conflict, the system either hedges or reaches for a competitor whose record is clean.
Google's own framing supports the shift. Prominence is based partly on how many websites link to the business and how many reviews it has — but the guidance for generative features warns directly against seeking inauthentic mentions, noting that core ranking systems focus on high-quality content while other systems block spam. Manufactured presence is a known failure mode, not a shortcut.
- Prioritised listing set by authority, not directory count
- Field-level consistency enforcement across every listing
- Duplicate and legacy listing suppression
- Industry and association presence that is genuinely earned
- Review generation built on real customer moments
- Owner responses that add retrievable specifics
To appear in Google's generative features a page must be indexed and eligible to be shown with a snippet, meeting the ordinary technical requirements for Search. That is the gate. Everything sold beyond it deserves scrutiny, and this pillar exists as much to remove work as to add it.
Google states that structured data is not required for generative AI search and that no special schema markup needs to be added — while still recommending it as part of overall SEO because it makes a page eligible for rich results. Both halves of that sentence matter, and most agencies quote only the half that sells.
- Crawlability and index-eligibility verification
- LocalBusiness and Organization markup that mirrors reality
- Core Web Vitals remediation on the pages that convert
- JavaScript rendering checks where content depends on it
- Duplicate content reduction across location pages and ongoing site maintenance
- Search Console verification and inclusion checks
Google publishes a first-party instrument: the generative AI performance report in Search Console, which reports impressions from AI Overviews and AI Mode and excludes Search Labs experiments. That is the measured layer, and it is the one we anchor reporting to.
Everything outside Google's own surfaces is observation. We sample a fixed query set against ChatGPT, Gemini, Perplexity, Claude and Microsoft Copilot, record which businesses each one names, and track how that set moves. Running all five matters because they disagree: Copilot answers from the Bing index rather than Google's, and a business can be named consistently by one assistant and absent from another for the same question. It is real signal and it is not a ranking metric, because no such metric is published. Google's guidance says it directly: no third-party tool has access to its internal ranking or AI systems. Ours included. We label our scoring observational because that is what it is.
- Search Console generative AI reporting as the measured baseline
- Assistant citation sampling on a fixed query set
- Competitor citation share, tracked over time
- Map pack and profile performance reporting
- Call and lead attribution to the surface that produced it
- Explicit evidence labelling on every figure reported
Before anything gets built, the contradictions get found.
Most local engagements begin with a build. Ours begins with an audit of what the open web currently believes about the business — because in almost every case, the first month of gains comes from removing conflicting information rather than adding new content.
What the assessment covers
The assessment is diagnostic, not a sales instrument. It produces findings that are true and useful whether or not the business ever engages us, and several have been handed to in-house teams who executed them alone.
- Profile integrity. Category accuracy, completeness, naming compliance, eligibility of every listed premises, and service-area declaration against Google's one-profile rule.
- Contradiction map. Every listing that names the business, field by field, with conflicts ranked by the authority of the source carrying them.
- Retrieval test. A fixed set of real customer questions run against Google's AI surfaces and against ChatGPT, Gemini, Perplexity, Claude and Microsoft Copilot, recording who gets named and on what basis.
- Technical eligibility. Index status, snippet eligibility, rendering, markup validity and Core Web Vitals on the pages that actually convert.
- Content substance. Whether the pages carry first-hand operational knowledge or restate what every competitor already published, assessed against the competitor set.
- Reputation content. What review text actually says about specific services, and whether owner responses add anything retrievable.
Foundation first, because the compounding depends on it.
The order matters more than the inventory. Prominence work performed on top of an unresolved entity produces signals that argue with each other, which is the specific failure this sequence exists to prevent.
Resolve the record
- Canonical fact set agreed and locked with the operator
- Profile corrected to real-world name, category and complete detail
- Highest-authority contradictions cleared first
- Duplicate and legacy listings suppressed
- Index eligibility and snippet eligibility confirmed
- Search Console verified and generative AI reporting baselined
Build the substance
- Service pages rewritten around the decision the customer is making
- First-hand operational detail captured from the people doing the work
- Question-level answers published where the questions are asked
- Markup deployed to mirror the corrected record
- Review generation built into real customer moments
- Core Web Vitals remediation on converting pages
Extend the corroboration
- Authority listings and genuine association presence pursued
- Owner responses adding retrievable service specifics
- Citation share tracked against named competitors
- Query set expanded as new customer language appears in market research
- Findings fed back into the content plan
- Reporting split into measured and observed layers
The layer that decides eligibility — and the layer that gets oversold.
Technical work for local search is narrower than the market implies. These four items carry most of the weight, and we say plainly where Google's documentation stops supporting the practice.
Index and snippet eligibility
A page cannot be cited by a surface that cannot retrieve it. Google requires that a page be indexed and eligible to be shown with a snippet in order to appear in generative features, on top of the ordinary Search Essentials and spam policy requirements. Blocking crawlers, then wondering why an assistant never names the business, is a more common sequence than it should be.
Also true: meeting every requirement guarantees nothing. Google states that indexing and serving are not guaranteed. We do not promise what the issuing authority declines to promise.
LocalBusiness structured data
LocalBusiness markup helps Google understand the business's details for local features. Deploy it — and deploy it so it agrees exactly with the profile and the page. Markup that contradicts the profile is another contradiction, not an advantage.
Where we stop: Google is explicit that structured data is not required for generative AI search and no special schema exists for it.
Page experience on converting pages
Core Web Vitals are the metrics Google reports for real-world page experience. For a local business the honest framing is that this is a conversion problem that also happens to be a search input — a slow service page loses the call whether or not it costs a position.
We remediate where the money is: the service pages, the contact path, the mobile experience a customer uses standing in front of a broken appliance.
Location pages that are not duplicates
The mass-produced city page is the single most common liability we inherit. Google's spam policies name scaled content abuse — generating many pages primarily to manipulate rankings rather than to help people — and its generative guidance warns specifically against creating separate content for every query variation.
Our rule: a location page exists when there is something true and specific to say about operating there. Otherwise it is a liability wearing the costume of coverage.
The framework is constant. The pressure points are not.
What breaks first differs by business model. These are the patterns we see most often across the verticals we serve.
Home services
Franchise systems
Medical and aesthetics
Legal
An agency that will tell you what not to buy.
There is no peer-reviewed science on how AI systems choose local citations. The field is roughly two years old. What exists is platform documentation, a small body of academic work, and correlational studies from tools with real data — and those are three very different grades of evidence.
Most of the market flattens them into one confident voice. We do not. Every claim on this page is labeled by the kind of evidence behind it — documented by the issuing authority, observed by us, or inferred. Where Google has published a mechanic, we cite Google. Where we are describing a pattern we have watched, we say so.
That is a harder page to write and an easier agency to verify. It is also the only defensible position when the honest answer to several popular questions is nobody has published that.
What separates this from a standard local SEO retainer.
Not a competitor teardown — a description of where the practices actually differ, so the choice can be made on substance.
| Practice | Standard local retainer | AllegiantOMNIVIZ™ |
|---|---|---|
| Directory strategy | Volume — submit broadly, report the count | Agreement — fewer sources, identical facts |
| Conflicting listings | Rarely audited after onboarding | Contradiction map is the first deliverable |
| Location pages | One per town, templated | Built only where operations differ |
| AI surfaces | Sold as a separate add-on discipline | Treated as search, per Google's own guidance |
| Reporting basis | Rank positions and a tool score | Search Console measured, sampling labeled observed |
| Claims about AI ranking | Stated with confidence | Labeled by evidence tier, or not made |
| Work we decline | Rarely itemised | Published on this page, with sources |
Five things you can stop paying for, on Google's authority rather than ours.
Google publishes a mythbusting section inside its own guidance for generative AI features. Each item below is drawn from it. If a proposal on your desk includes one of these as a line item, this is the source to read before signing.
Files written for AI systems. Google states you do not need to create AI text files, special markup or Markdown to appear in Search including its generative capabilities, because Search does not use them. Maintaining one for other systems is fine — it will neither help nor harm Google visibility.
Chunking content into fragments. There is no requirement to break content into small pieces for AI to understand it, and there is no ideal page length. Pages are for the audience.
Rewriting pages in an "AI-friendly" voice. Not needed. The systems understand synonyms and general meaning, which is also why chasing every long-tail variation of a phrase is wasted budget.
Buying mentions. Seeking inauthentic mentions is named directly as less useful than it appears, with core ranking systems focused on quality content and other systems blocking spam.
Schema sold as an AI unlock. Structured data is not required for generative AI search and there is no special markup for it — though it remains worth deploying for rich-result eligibility, which is a different and honest reason to do it.
The pattern underneath all five is the same: there is no separate AI channel to buy. There is search, done well or done badly, and a set of surfaces that read the result.
- Google Business Profile Help — Tips to improve your local ranking
- Google Search Central — Optimizing for generative AI features
- Google Search Central — AI features and your website
- Google — Guidelines for representing your business
- Google — Manage service areas for service-area businesses
- Google Search Central — LocalBusiness structured data
- Google Search Central — Establish your business details
- Google Search Central — Creating helpful, people-first content
- Google Search Central — Spam policies for Google web search
- Google Search Central — Core Web Vitals
- Search Console Help — Generative AI performance report
- Google Search Central — Evaluating third-party SEO advice
The questions operators actually ask.
Answered against primary sources where primary sources exist, and answered honestly where they do not.
Is local SEO still worth doing now that AI answers the question first?
Yes, and the reason is structural rather than optimistic. Google's own position is that optimizing for its generative AI features is optimizing for search, and thus still SEO, because those features are rooted in the core ranking and quality systems. The generated answer is not drawing from a separate index — it retrieves from the same one. Work that makes a business retrievable and credible feeds both surfaces at once. What has changed is that the standard is higher and the tolerance for contradiction is lower. Our AI SEO practice and this local work are the same discipline applied at different radii.
Should we create a separate Google Business Profile for every city we serve?
No, and for a service-area business this is one of the few places Google's rule is unambiguous. A service-area business — one that visits or delivers to customers but does not serve them at its own address — can only have one profile for the whole area it serves. Google further instructs that if you do not serve customers at your business address, the address should be removed from the profile. Extra profiles created per town are a suspension risk rather than a coverage strategy, and cleaning them up is often the first thing we do in profile optimization.
Can we use a virtual office or co-working address to rank in another market?
No. Google's guidelines state that a rented physical mailing address the business does not operate out of — a virtual office — is not eligible for a Business Profile, and that an office in a co-working space does not qualify unless it maintains clear signage, receives customers during business hours, and is staffed during those hours by your business. The honest route into a market you cannot staff is prominence, which is the factor that lets a business compete beyond its immediate radius. That work runs through directory and citation management and reputation, not geography.
Do we need special schema or an llms.txt file to be cited by AI?
No to both, on Google's authority. Its guidance states that structured data is not required for generative AI search and there is no special schema markup to add, and that AI text files are not used by Google Search at all — maintaining one for other systems will neither help nor harm Google visibility. Structured data is still worth deploying, for a different and legitimate reason: it makes pages eligible for rich results. We deploy LocalBusiness markup as part of technical SEO, and we do not bill it as an AI unlock.
How do you actually measure whether we're showing up in AI answers?
In two layers, kept separate on purpose. The measured layer is Google's own generative AI performance report in Search Console, which reports impressions from AI Overviews and AI Mode and excludes Search Labs experiments. The observed layer is our own sampling across ChatGPT, Gemini, Perplexity, Claude and Microsoft Copilot — a fixed set of real customer questions run against the assistants, recording who gets named and how that changes. The second layer is genuine signal and it is not a ranking metric. Google states directly that no third-party tool has access to its internal ranking or AI systems, and that includes ours. Anything reported to you carries its evidence tier on its face, and the sampling method is the same one behind AI visibility monitoring.
Our previous agency built forty city pages. Should we keep them?
Usually not in their current form. Google's spam policies define scaled content abuse as generating many pages primarily to manipulate rankings rather than to help people, and its generative guidance separately warns against creating separate content for every query variation, noting that a high quantity of pages does not make a site higher quality. The test we apply is simple: is there something true and specific about operating in that place? Crew, permits, building stock, response times, local conditions. If yes, the page earns its existence and gets rewritten. If no, it gets consolidated. This is usually handled alongside site structure work.
How much do reviews really matter, and is it the rating or the volume?
Google's local ranking guidance names both: prominence is based partly on how many websites link to the business and how many reviews it has, and states that more reviews and positive ratings can help local ranking. That covers the map. For the generated answer, our observation — labeled as observation, because Google publishes no mechanic here — is that the text matters in a way the star average does not: a review naming a specific service in a customer's own words is retrievable content, while a five-star rating with no text is a number. We build review generation into real service moments through reputation management rather than blasting requests at a list.
What is the single highest-impact thing we could fix this month?
In most audits it is not content — it is contradiction. A business whose name, address, phone, hours or service area disagree across its own listings gives every system a reason to hedge, and no amount of new publishing outruns that. Google's guidance is to represent the business as it is consistently represented and recognized in the real world and to keep the address or service area accurate and precise. After that, primary category accuracy, because a category chosen for breadth dilutes every query it might have matched. Both are fixable inside thirty days, and both are covered in a visibility assessment.
Is AEO or GEO something different from what we're already paying for?
From Google's perspective, no. Its guidance addresses the terms directly, stating that answer engine optimization and generative engine optimization describe work focused on visibility in AI search experiences, and that from Google Search's perspective this is still SEO — it also points readers to its guidance on evaluating third-party SEO advice before buying such services. The terms are useful as lenses. They are not separate channels to be invoiced separately, which is why our generative engine optimization and answer engine optimization work sits inside one framework rather than beside it.
How long before we see something, and what should we expect first?
Profile and record corrections tend to surface fastest because they change what Google can understand about the business immediately — Google's remedy for relevance is complete and detailed business information, and that is a same-week change. Prominence is slower and compounds over months, because links, mentions and reviews accumulate rather than switch on. We will not put a date on a ranking position: Google states plainly that indexing and serving are not guaranteed even when every requirement is met, and an agency that promises what the issuing authority declines to promise is telling you something about itself. What we commit to is a sequence, a baseline, and reporting that separates what was measured from what was observed. That is also how our content programs and Local Services Ads work is reported.
Find out what the web currently believes about you.
A local visibility assessment: what your listings actually say, where they contradict each other, and which businesses are being named when customers ask about your services. Findings are yours whether or not we work together.
- Contradiction map across every listing that names your business
- Profile integrity review — category, completeness, naming and eligibility
- Retrieval test on a fixed set of real customer questions
- Technical eligibility check on the pages that convert
- Prioritised findings with the evidence tier stated on each, plus relevant case studies
Prefer to start with the deeper diagnostic? Request an A.R.C. Report, or review the full services we bring to a local program — including paid search, social, email, CRM integration and conversion rate optimization.
See what your listings actually say, where they contradict each other, and which businesses are being named when customers ask about your services.
No cost, no commitment. We will follow up by email or phone to walk you through the findings.
