Guides/Technical fixesupdated 2026-08-04 · 5 min

NAP Consistency: Why AI Gets Your Phone Number Wrong

The short answer

AI assistants assemble a picture of your business from many sources. When your name, address, and phone number disagree across listings, the picture is blurry, and blurry businesses get named less often — not as a penalty, but because the assistant reaches for the business it can describe confidently.

Here is a failure mode we see on most scans, and almost no owner knows it is happening.

The website footer says Carter Plumbing. Google says Carter Plumbing & Drain LLC. Yelp has the old suite number from before the 2021 move. Angi has a call-tracking number from a campaign that ended two years ago. Facebook has the hours from before Saturday service started. BBB has a phone number that now rings a fax line.

Each one, individually, is trivial. Together they are the reason an AI assistant names your competitor.

Why it matters more than it used to

Classic search engines matched pages to queries. A bit of variation in your listings was survivable because the engine was ranking documents, not reasoning about a business.

AI systems are doing something different. They are assembling an entity — a business with properties — from every source they can find. Then they decide whether they know enough about that entity to recommend it, and what to say about it.

When the sources agree, the entity is well-defined and describable. When they disagree, the system has a set of partially-contradictory claims about something that might be one business or might be two.

There is no penalty applied. It is simpler than that: the assistant names the business it can describe cleanly, and that is your competitor.

The audit

Ninety minutes, once. Do it properly and you will not need to do it again for a year.

Step 1: define canonical

Decide, right now, one exact version of each:

  • Name. Exactly as it should appear everywhere. Include or exclude "LLC" consistently. Decide on "&" versus "and." Write it down.
  • Address. One format. "Suite 200" or "Ste 200," not both. "West" or "W." Match USPS formatting.
  • Phone. One number. Your real main line. Formatted consistently.

Additionally: hours, website URL (with or without www, http vs https — pick one and redirect the other), and your business description.

Step 2: find every listing

Search these, and write down what each says:

"Your Business Name"
"Your Business Name" "Your City"
"Your Phone Number"
"Your Old Phone Number"
"Your Old Address"

Searching your old phone number and address is the step people skip, and it is where the stale listings hide.

Then check each of these directly:

Step 3: fix the majors first

Google, Bing, Apple, Yelp, Facebook, BBB, Angi. These are the ones the assistants read and the ones that other directories re-scrape from.

Most allow direct editing once you have claimed the listing. Claiming is the tedious part — postcard verification, phone verification, email confirmation. Do it anyway.

Step 4: fix your own site

Your website is a source too, and it is the one you fully control.

  • Footer NAP matching canonical exactly
  • Contact page matching
  • Schema matching
  • No conflicting phone numbers in page copy
  • Consistent formatting throughout

Step 5: link your identities with sameAs

In your LocalBusiness schema, list every profile URL in the sameAs array:

"sameAs": [
  "https://www.google.com/maps/place/?q=place_id:ChIJxxxxxxxx",
  "https://www.yelp.com/biz/carter-plumbing-tampa",
  "https://www.facebook.com/carterplumbingtampa",
  "https://www.bbb.org/us/fl/tampa/profile/plumber/carter-plumbing-xxxxx",
  "https://www.angi.com/companylist/us/fl/tampa/carter-plumbing.htm"
]

This is the most direct statement available that all of these listings describe one business. It is also the most under-used field in local schema.

The call tracking problem

Call tracking is the most common cause of NAP inconsistency, and you do not have to give it up.

Do: use dynamic number insertion that swaps the displayed number in the browser for visitors from a given source, while your real number remains in the raw HTML and in your schema.

Do not: put a tracking number in a permanent directory listing. That number will be scraped, propagated, and eventually orphaned when the campaign ends, leaving a dead number attached to your business identity in a dozen places.

Do not: put a tracking number in your schema. Schema should always carry your canonical main line.

The other common causes

A move. Old addresses persist in scraped directories for years. Search your old address explicitly.

A name change or rebrand. Both names circulate. Make sure the new one is canonical everywhere you control, and correct the majors.

Multiple locations. Each location needs its own profile with its own address and ideally its own local number, plus its own page on your site with location-specific schema. Do not run two locations off one listing.

A franchise or DBA. Legal name and operating name must both appear somewhere connected, or the state license lookup will not match your listings.

A defunct partner listing. An old lead-gen relationship that created a listing with their tracking number. These persist. Find them.

How to keep it clean

  • Quarterly: search your name and phone, scan the first two pages for drift.
  • Whenever anything changes: update Google first, then the other majors within a week.
  • Never create a listing with a temporary number.
  • Keep the canonical values written down somewhere your team can find them, so nobody invents a new format.

What to do this week

  1. Write down your canonical name, address, phone, hours, and URL.
  2. Search your name, phone, old phone, and old address, and list every result.
  3. Fix Google, Bing, Apple, Yelp, Facebook, BBB, and Angi.
  4. Make your footer, contact page, and schema match exactly.
  5. Add every profile URL to your schema's sameAs array.
  6. Replace any tracking number in a directory with your real line.

Then check what the assistants think

Once your facts agree, ask the assistants what they know about you. Our free scan includes branded questions — "is [your business] any good," "tell me about [your business]" — so you can see exactly what they get right and wrong.

About fifteen seconds, free, no account.

Common questions

How much does NAP inconsistency actually matter?
More than it did for classic SEO. Search engines could tolerate variation because they matched pages to queries. AI systems are reasoning about a business as an entity, and conflicting facts reduce their confidence that they are looking at one real business. Low confidence means the competitor with consistent data gets named.
What about call tracking numbers?
They are the most common cause of inconsistency. If you use call tracking, use dynamic number insertion that swaps the number in the page for visitors while keeping your real number in the HTML and schema, and never put a tracking number in a permanent directory listing.
Do I need to fix every directory?
Fix the ones assistants actually read: Google, Bing Places, Apple Business Connect, Yelp, Facebook, BBB, Angi, and your trade association and manufacturer listings. The long tail of scraper directories matters far less and largely re-syncs from the majors over time.
What if my business moved or changed names?
Fix the majors first, then work outward. Old addresses persist in scraped directories for a long time. The most important thing is that your Google profile, website, and schema agree with each other today — that is the cluster with the most influence.
Stop guessing

See exactly what the five AIs say about your business.

Twenty-five real customer questions across Claude, ChatGPT, Gemini, Perplexity, and Google AI Overviews. Free, about fifteen seconds, no account.

Run the free scan

Keep reading