A different-looking local result can change how someone contacts you. A statistic from another company’s test cannot tell you how many customers you lost. If your calls are down while a rank report looks steady, inspect the actual search experience and the path from result to inquiry.
This article previously turned a reported “32% fewer businesses” finding into a broad forecast for Seattle businesses. It also implied that a percentage change in the number of businesses shown could predict an individual company’s appearance rate. That calculation was not justified. This revision explains the source and gives you a test you can run on your own market.
Where the 32% came from
Sterling Sky’s article on local SEO and AI reports that AI local packs show 32% fewer unique businesses than traditional packs. The same article discusses call data from 179 Business Profiles and a separate Places Scout observation about AI-pack prevalence. These are industry observations, not a government dataset or an official Google guarantee about every local result.
The linked page does not supply enough methodology alongside the 32% statement to reproduce a Seattle-specific estimate: you would need the query set, geographic coverage, dates, devices, result definitions and business-counting method. The number can justify a closer inspection. It cannot establish your lost calls, market share or probability of appearing.
Do not merge distinct findings into one study. Call data, result-layout counts and the frequency of an interface are different measurements. A decline in calls can also involve demand, hours, competition, phone routing or seasonality. An interface change is one explanation to investigate.
Name the result you are measuring
A standard local pack, a Maps result, an AI Overview, an AI Mode response and a third-party chatbot recommendation are different surfaces. A report that labels all of them “AI ranking” can hide the exact problem you need to fix. Save the screenshot and describe what appeared rather than assuming a label used by one analyst is Google’s universal product name.
Google’s AI-feature documentation explains that AI Overviews and AI Mode can use different models and techniques, with varying responses and links. This supports recording the surface and conditions when you test. It does not establish one permanent set of AI-local positions.
Build a repeatable worksheet
| Record | Example or instruction |
|---|---|
| Query | Use a real service query, with its exact wording |
| Location | Record the searcher’s actual or test location |
| Environment | Date, time, device, account state and search surface |
| Visible result | Save screenshot; note named businesses and contact options |
| Your business | Present or absent; correct facts; working destination |
| Follow-up | Open cited links and test the call or booking route without submitting a fake inquiry |
Start with five high-value queries from two actual customer areas. Observe them on more than one date and separate mobile from desktop where relevant. This is a manageable diagnostic exercise, not a representative survey of every Seattle search. Record “no AI result shown” as an outcome rather than discarding it.
If you count appearances, include the denominator: “Named in 6 of 20 recorded observations” is clearer than “30% AI visibility.” Preserve the prompts and screenshots. Do not describe observations from several platforms as independent users or multiply a small sample into a forecast of monthly impressions.
Inspect the contact path
For each result where you appear, ask what a customer can do next. Is there a call option, a website link, a directions link or a booking action? Does the linked page show the correct phone number? Does the number route to a person during advertised hours? Can a customer book on mobile without an overlay obscuring the form?
This catches a problem that a position report cannot: visibility without a functioning next step. A restaurant may appear for private dining but send the customer to an ordinary table-reservation form. A contractor may advertise emergency work while its phone routes to an unattended voicemail. Both deserve attention before another visibility campaign.
Use documented local fundamentals
Google’s local-ranking guidance points to relevance, distance and how well known a business is. Keep the profile accurate, choose categories that describe the real business and maintain current hours. Explain actual services on the website. These are sensible corrections with documented support, even though they do not guarantee an AI recommendation.
Do not create fictitious locations to increase your chances of appearing. Do not ask customers to insert prescribed content into reviews. Do not treat LocalBusiness markup as an inclusion switch. Structured information should describe reality; a claim that a particular markup change caused more recommendations needs its own evidence.
Connect observations to calls
Compare like periods and account for business changes. If operating hours changed, record the date. If a phone provider changed, inspect missed and failed calls. Separate qualified inquiries from spam and vendor solicitations. Match inquiries to service type so a decline in lucrative work does not disappear inside a stable total.
A useful investigation has three columns: search observations, contact-path checks and actual inquiries. If all three deteriorate in the same period, you have a stronger lead to investigate. You still need to test the explanation. If search presence is stable but missed calls increase, operations may be the immediate constraint.
Decide what to fix first
Fix incorrect hours, broken links and failed call routing immediately. Next, clarify missing service facts and booking requirements. Then expand the observation set if a pattern remains. Paid search may be worth a separate, measured experiment, but an industry headline is not sufficient reason to increase ad spend or abandon organic work.
Ask any agency reporting a visibility decline for the raw observations, dates and definitions. Ask which part of the finding applies to your business and what evidence would change the recommendation. A useful audit leaves you with a reproducible record and a prioritized action—not a frightening percentage without a denominator.
Updated October 30, 2026: corrected the scope of the 32% claim, removed unsupported universal rollout and call-button assertions, and replaced ranking speculation with a reproducible audit. Platform guidance and the original industry source are linked above.
Written by the Seattle Organic SEO Team
The Seattle Organic SEO team combines more than 30 years of hands-on client experience across SEO, paid media, email marketing and affiliate marketing. Our team has worked for leading technology companies, including Amazon, Expedia and Google-acquired DoubleClick. We bring that experience to source-based research, transparent analysis and practical guidance for Seattle businesses. Learn more about our team and approach.
Want to inspect your actual results?
We can check your important local searches, review the contact path and compare the findings with qualified inquiries. Bring the service and locations that matter to your business.