Skip to content

Your client portal is here, with your reports, updates and tasks under one roof. Sign in to Lumen

Search + AI

Local SEO Is AI Search Infrastructure

Learn how locations, business profiles, NAP consistency, people, services and reviews create the entity infrastructure AI systems use for local answers.

In this article 19 sections

Local search depends on clear business information: what you do, where you work, how customers reach you and whether those details are current.

Your website, business profiles, directories and structured data should describe the same business. Address changes, inconsistent hours and tracking numbers can make that harder to maintain. Treat these details as connected business information. The checks below help you find contradictions while preserving useful call tracking and accurate service-area information.

A local business exists in many places at once

A local business can be represented by:

  • its website,
  • Google Business Profile,
  • Bing Places,
  • Apple Maps,
  • industry directories,
  • review platforms,
  • professional associations,
  • social profiles,
  • local news,
  • chambers of commerce,
  • and dozens of old citations nobody remembers creating.

This distributed identity is why local SEO has always involved consistency. AI-assisted search raises the stakes because answers can synthesize information across sources.

The website should be the strong first-party source

Your website should make the basics unambiguous.

For each real location:

  • Business name
  • Address
  • Phone
  • Hours
  • Services
  • Relevant people
  • Service area where appropriate
  • Map/location context
  • Directions/access information where useful
  • Canonical location URL

If the website does not clearly describe the location, do not expect third-party profiles to reconstruct the business perfectly.

Location pages are entity pages

A good location page is not:

“Personal Injury Lawyer Phoenix” repeated 22 times.

It represents a place. That means it can contain genuinely local information.

Who works there?

What services are available?

What neighborhoods or courts are relevant?

Where should visitors park?

What number reaches that office?

What are the hours?

Is it accessible?

Does the firm actually occupy the location?

Real entity information creates useful local content. City-name replacement does not.

Mithril LabsTry it.
Observe it.
Learn from it.

Try it on your site

The four-source location check

For each office, compare four sources:

Check:

  • Name
  • Address
  • Phone
  • Hours
  • URL
  • Category/service description

Then classify mismatches:

  • Expected tracking variation
  • Minor formatting difference
  • Stale information
  • Material contradiction
  • Missing data

This diagnostic is simple enough to perform quarterly. It can prevent remarkably strange machine answers.

Phone tracking requires governance

Call tracking is essential for many service businesses. It also creates multiple numbers intentionally. That means “NAP consistency” needs nuance. A tracking number is not automatically bad.

The important questions are:

  • What is the canonical business number?
  • Which numbers are DNI?
  • Which numbers are ad assets?
  • Which belong to individual campaigns?
  • Which belong to specific locations?
  • Does structured data use the intended stable number?
  • Does the website have sensible fallbacks?
  • Can humans and machines understand the relationship?

The answer is not:

“Never use tracking.”

The answer is:

Track without destroying identity.

People make local entities stronger

For professional services, offices and people are related. A law firm’s Phoenix office may have specific attorneys. A dental practice location may have specific doctors. A consultancy may have market leads.

Link:

  • Location → Person
  • Person → Location
  • Person → Service
  • Service → Location where relevant

Those relationships help both humans and machines understand the organization. They also create useful conversion paths.

Reviews are evidence, not just stars

Reviews can communicate:

  • service quality,
  • recurring strengths,
  • locations,
  • staff names,
  • and customer experiences.

But avoid treating review count as a magical AI ranking factor.

The safer position is:

Authentic reviews are valuable reputation evidence.

They can also create third-party information about the entity. Do not manufacture them. Do not stuff review markup with first-party testimonials and assume Google treats them as independent. Reputation is larger than schema.

Myth BustedA popular claim.
A closer look.

Follow the evidence

Myth: AI search makes Google Business Profile less important

What the evidence says

Your Business Profile remains a key place to maintain accurate information for Google’s local experiences.

Keep it aligned with your website’s contact details, services and hours. Review the information customers actually encounter when they search for your business.

Bing Places deserves more attention than it gets

Many local businesses obsess over Google and barely look at Bing. That made more sense when Microsoft’s ecosystem was mostly viewed as traditional Bing search. Copilot changes the strategic picture.

Bing’s search and grounding ecosystem matters to Microsoft AI experiences. That makes accurate Bing-facing business and website information more valuable. No, this does not mean Bing Places is a secret Copilot ranking button.

It means the Microsoft ecosystem deserves clean business data too.

Service areas need clarity

A business can have:

  • real offices,
  • service areas,
  • markets served remotely,
  • and locations mentioned for marketing.

Do not blur them. If you serve Tucson but have no Tucson office, say that clearly. Do not build a fake “Tucson office” entity because somebody wanted a map-pack shortcut.

Machine systems are getting better at reconciling location evidence. More importantly, users deserve accurate information.

The “near me” answer is an entity problem

When a user asks:

“Best personal injury lawyer near me”

the system may consider:

  • location,
  • service relevance,
  • business identity,
  • reviews,
  • proximity,
  • prominence,
  • and other factors depending on the platform.

Your website cannot control all of that.

It can control whether:

  • the correct office exists,
  • the address is accurate,
  • the service is clear,
  • the people are connected,
  • and the page is crawlable.

Do the controllable part well.

Mithril LabsTry it.
Observe it.
Learn from it.

Try it on your site

The local entity graph

Sketch:

The local entity graphOrganization → office → people → servicesOne office, its people and its services, with the relationship between each.
  1. Organization
  2. has locationPhoenix Office
  3. staffed byAttorney AAttorney B
  4. providesPersonal InjuryWrongful Death

Now check whether the website actually represents those connections through:

  • pages,
  • internal links,
  • visible copy,
  • and structured data.

If the graph exists only on a whiteboard, the site may not communicate it clearly. This is a useful way to review local architecture.

Canonical location URLs matter

Avoid creating multiple competing pages like:

  • /phoenix/
  • /locations/phoenix/
  • /phoenix-office/
  • /personal-injury-lawyer-phoenix-office/

unless they have distinct jobs. A real location should generally have one canonical entity page. Service-location combinations can exist when they provide meaningful distinct value. But do not create a matrix so large that the business becomes a spreadsheet with stock photos.

Myth BustedA popular claim.
A closer look.

Follow the evidence

Myth: You need a page for every city you serve

What the evidence says

Create a location page when it has a distinct purpose and useful local information.

Explain real offices, coverage and service differences accurately. A city-name swap adds little when the underlying offer and answer are identical.

How AI answers expose local inconsistency

Traditional search may show several results and let the user reconcile them. Generative answers may present one summarized answer. That makes wrong information feel more definitive.

If the system picks:

  • the old office,
  • wrong hours,
  • or the former attorney,

users see the error more directly. This is why local data maintenance becomes part of AI-answer quality.

Mithril LabsTry it.
Observe it.
Learn from it.

Try it on your site

The former employee check

Professional-service sites often accumulate stale people data.

When someone leaves:

  • Remove or update team pages.
  • Handle the person’s canonical URL appropriately.
  • Update structured data.
  • Update location associations.
  • Review author pages.
  • Update business profiles if needed.
  • Review old high-ranking content authored by them.

Preserve historical accuracy where appropriate without implying current employment. This is entity governance. It matters.

LumenWhat a tool can check

How Lumen can help

Compare the business name, address and phone number in visible pages with structured data and the intended location identity. Include the telephone link and tracking fallback in that comparison.

An illustrative mismatch might involve a location page displaying a local number while its schema references the main office. Confirm the intended contact route, preserve useful tracking and align the location’s current information. Carry that finding into the technical audit plan.

The takeaway

Keep business details current across your site and key profiles. Clarify real locations and service areas, check contact-number behavior and correct contradictions at their source.

Sources and primary references

  1. Google, Optimizing for generative AI featuresdevelopers.google.com
  2. Google, LocalBusiness structured datadevelopers.google.com
  3. Bing Webmaster Guidelinesbing.com
  4. Bing Webmaster Toolsbing.com

The AI search guide

Where this fits

AI Search OptimizationThe starting point: how websites get crawled, retrieved, understood and cited.
  1. Access

    AI Crawlability
  2. Retrieval

    How AI Search Finds Sources
  3. Understanding and evidence

    Entities and Evidence
  4. Citation and outcome

    Measuring AI Search Visibility
Technical AI Search AuditThe capstone: the audit that tests every stage.
Cloudflare and AI PolicyTimely: Cloudflare’s controls as of September 18, 2026.