Local search was always more decisive than general search — someone asking for a plumber nearby is much closer to buying than someone reading about plumbing.
Local AI search compresses that further. A map pack shows three businesses and a map. An assistant asked the same question frequently names one.
Consistency is the foundation and it decays
Name, address and phone consistency across directories is old advice that matters more now, because entity resolution depends on it and assistants state facts confidently.
The part teams underestimate is decay. Directories change formats, staff update one profile and not others, an office moves and three listings are missed, an aggregator republishes stale data.
This is quarterly maintenance rather than a project. Treating it as a project is why most audits find inconsistencies eighteen months after the last cleanup.
The Business Profile is not optional
For local questions, the Google Business Profile is among the most directly consulted sources available. Incomplete profiles produce incomplete answers.
Complete every field, not the required ones: primary and secondary categories, individual services with descriptions, attributes, hours including exceptions, service area if you travel, and photographs that are current.
The service list matters disproportionately. It is a structured statement of exactly what you do, in a source that carries weight, and most profiles leave it near-empty.
Reviews as a corroboration signal
Review volume, recency and rating are among the strongest independent signals in local search, and independent signals are precisely what generative systems weight most.
The businesses that do well here are not the ones with the best service. They are the ones that ask consistently, at a defined moment, through a process that does not depend on someone remembering.
That is an operational fix rather than a marketing one, which is why it is so often unowned.
A repeatable ask at a repeatable moment beats an occasional campaign. Volume and recency both matter, and recency is what a sporadic push fails to sustain.
Location pages that are not templates
The failure pattern is unmistakable: forty pages identical except for the town name. Search engines and assistants both discount them, and correctly.
A location page earns its place when it contains something true only of that location — the team there, the service radius, local response times, projects completed nearby, parking and access, referral relationships, area-specific constraints.
For multi-location businesses this is genuinely hard, because it requires local teams to contribute. Without that contribution, fewer and better pages beat more and emptier ones.
Location pages that work and pages that do not
The difference is whether the page could be about anywhere else.
Discounted
- Identical copy with the place name swapped
- Generic stock photography
- No named people or local detail
- Same service list on every page
- Created in bulk on one date
Rewarded
- Named local team and contact
- Actual service radius and response times
- Nearby work referenced specifically
- Local practicalities answered
- Updated when the location changes
Measure per location
A blended visibility number across twelve locations hides which three are carrying the average and which four are invisible.
Run the baseline per location with location-specific prompts, and report per location. It is more work and it is the only reporting that supports a decision.
Common mistakes
- Treating NAP consistency as a one-off cleanup rather than quarterly maintenance.
- Leaving the Business Profile service list empty while investing in on-site content.
- Running occasional review campaigns instead of a repeatable process.
- Publishing templated location pages at volume.
- Reporting a blended figure across locations with very different performance.
- Ignoring service-area structuring for businesses that travel to customers.