When a local business moves, the new address shows up in Google Business Profile within days. But ask ChatGPT, Claude, or Perplexity where that same business is located three months later, and there's a decent chance you get the old address — the strip mall you left, the suite you outgrew, or the neighborhood you escaped. Customers show up at the wrong door. Some leave. Some don't bother at all.
This isn't a bug in one AI system. It's a structural problem that hits every business that relocates, and understanding it is the first step to fixing it.
Why the Old Address Persists in AI Answers
AI assistants don't pull your address from a live database the way Google Maps does. They work from a snapshot — a training corpus built from web crawls, directory exports, review platforms, and structured data collected months before the model shipped.
When you moved, that snapshot already existed. Every citation of your old address in that corpus still counts. A Yelp listing, a Yelp-cached aggregator, a local chamber directory, a citation on a blog post from three years ago — they all contributed to what the model "knows" about where you are.
Even retrieval-augmented systems like Perplexity, which pull live search results, are only as current as the sources they retrieve. If the top three results for "Riverside Dental on Oak Street" still reference your old building, that's what gets cited.
The Aggregator Multiplier Problem
One stale address in one place is manageable. The real damage comes from data aggregators — companies like Localeze, Acxiom, and Foursquare that redistribute business data to hundreds of downstream directories. When your old address lives in an aggregator's database, it gets re-syndicated continuously. New directories that launched after your move get the old address from the start.
A family-owned bakery that moved two blocks in 2024 might have a correct Google listing but still show the old address on forty-plus directories the owner has never heard of. When AI models train on web-wide data, those forty incorrect citations outweigh the one correct one.
This is the same dynamic explained in why AI gets your business info wrong — consistency of signal matters more than the authority of a single source.
Which Sources Matter Most After a Move
Not all citations carry equal weight. Prioritize in this order:
Google Business Profile — Update this first. It's the most heavily indexed source and feeds multiple AI retrieval pipelines. Verify the new address; don't just edit and assume.
Apple Maps Connect — Siri and Apple Intelligence pull from Apple's own business graph. A stale Apple Maps pin is invisible to Google's tools and requires a separate correction.
Bing Places — Copilot (Microsoft's AI assistant) weights Bing's index heavily. Many local businesses skip this and pay for it in Copilot answers.
Core aggregators — Submit directly to Neustar Localeze, Data Axle, and Foursquare. These three feed the widest number of downstream directories. One correct update here cascades further than manually fixing fifty individual sites.
Industry directories — A plumber who moved should prioritize Angi, HomeAdvisor, and Thumbtack. A dentist should prioritize Zocdoc and Healthgrades. These vertical directories often rank highly in the retrieval results AI systems use.
Structured Data Is Your Fastest Signal
Directory corrections take weeks to propagate. Your website is something you can fix today. Adding or updating LocalBusiness schema on your homepage gives AI crawlers — and retrieval systems — a machine-readable, unambiguous address signal.
The key fields after a move:
{
"@type": "LocalBusiness",
"name": "Riverside Dental",
"address": {
"@type": "PostalAddress",
"streetAddress": "412 Maple Ave",
"addressLocality": "Springfield",
"addressRegion": "IL",
"postalCode": "62704"
}
}
Don't leave old address structured data on any page. If your contact page still has the old suite number in schema markup, that's a direct contradiction that AI systems will weigh against your corrected listing.
The 90-Day Window
Most AI training cycles and retrieval index refreshes operate on roughly a 30–90 day lag. The work you do in the first three months after moving determines how quickly AI answers correct themselves.
A landscaping company in a mid-size city moved its office in January. By March, ChatGPT still cited the old address. The owner updated Google and the website but had done nothing about aggregators or Bing. After correcting the core aggregators and adding schema markup, AI answers started showing the correct address by late April — roughly ten weeks later.
That's a best-case scenario for a business that acted quickly. Businesses that wait six months to clean up stale citations often find the incorrect address baked into AI answers for a year or more.
One More Signal That Gets Overlooked
Your "moved" announcement itself becomes a citation. A post on your Google Business Profile saying "We've moved to 412 Maple Ave" gets indexed. A press release, a neighborhood blog mention, a local news blurb — these all create dated references to the new address that retrieval systems can cross-reference against the old data.
Brief, factual announcements on channels that get indexed (not just Instagram Stories, which AI can't read) help establish a clear before/after timeline in the data. The AI isn't reading your post to understand the narrative; it's counting citations and their recency.
Run a free scan to see what address AI assistants are currently returning for your business — it takes thirty seconds and shows you exactly which AI systems have the wrong location before your next customer ends up at the wrong door.