Voice-First AEO for Local Services: Siri, Alexa and Google Assistant
A voice query does not return a page of options. It returns one business, spoken aloud. For a plumber, a clinic or a restaurant, that is the difference between a booking and silence. There is no position four in a spoken answer.
How do you optimise a local business for voice search?
Get your listing data identical everywhere, answer questions in the exact words people speak, and mark up hours, services and location in structured data. Voice assistants pull from local business records and structured content far more than from page copy, so consistency across sources beats keyword density every time.
Why voice queries behave differently
Spoken searches are longer, phrased as full questions, and carry stronger intent. Nobody says two words into a phone. They say where can I get a boiler serviced this week near me. Three consequences follow: conversational phrasing matters more than short keywords, the winner takes everything because only one result is read out, and local signals dominate because most voice queries have implicit location.
The foundation: consistency across every listing
Name, address and phone number need to match exactly across your website, your Google Business Profile, Apple Business Connect, Bing Places and every directory that carries you. Not approximately. Exactly, including whether you write Street or St.
Apple Business Connect deserves specific attention because Siri leans on it and most businesses have never claimed their listing there. It is a common gap and a quick win.
Structured data that voice assistants use
- LocalBusiness with the correct subtype, full address, geo coordinates and telephone
- openingHoursSpecification as real data, including holiday exceptions, not an image of a timetable
- Service entries for each thing you actually do, named the way customers say them
- areaServed covering the towns and neighbourhoods you genuinely cover
- FAQPage answering the questions people ask before booking
- Speakable marking the passages best suited to being read aloud
Write the way customers speak
Your services page probably says Domestic Heating Solutions. Your customers say boiler repair. Voice matching works on natural phrasing, so the page needs to contain the phrase a person would actually say, ideally as a heading followed by a short spoken-style answer.
Build the question list from real sources: the calls your team takes, the messages you get, the autocomplete suggestions on your core services. Those are the queries, in the customer''s own words.
The details that decide close calls
When two businesses are equally relevant, assistants fall back on signals of reliability: recent reviews, a rating above the local average, complete profile information, correct current opening state, and responsiveness. Hours accuracy matters more than most owners expect. Being named as open when you are closed is the fastest way to lose a customer permanently.
How to test it
Say your ten most valuable queries into Siri, Alexa and Google Assistant, on a phone that is not signed into your own accounts, from the area you serve. Write down which business gets named. That list is your gap analysis, and it is more useful than any rank tracker for this channel.
The FAQ schema guide covers the markup in detail, and a free audit checks whether your structured data is valid.
Frequently asked questions
How do I optimise my business for voice search?
Keep name, address and phone identical across every listing, answer questions using the exact words customers speak, and publish LocalBusiness schema with accurate opening hours, services and areas served. Consistency across sources outweighs keyword density.
Does Siri use a different source from Google Assistant?
Yes. Siri leans heavily on Apple Business Connect, which many businesses have never claimed, while Google Assistant draws on Google Business Profile. Claiming both is a common and quick win.
Why is only one result read out in voice search?
Because a spoken answer has no room for a list. The assistant selects the single best match, which makes voice a winner-takes-all channel where being second returns nothing.
What schema matters most for local voice search?
LocalBusiness with the correct subtype, openingHoursSpecification as real data including holiday exceptions, Service entries named as customers say them, areaServed, FAQPage, and Speakable for passages meant to be read aloud.
How do I find the questions customers actually ask?
Use the calls your team takes, the messages you receive and autocomplete suggestions on your core services. These give you the phrasing in the customer''s own words rather than your internal service names.
How can I test my voice search visibility?
Speak your ten most valuable queries into Siri, Alexa and Google Assistant from within your service area, on a device not signed into your own accounts, and record which business gets named each time.
Want this done for your site?
Run a free audit and see exactly what to fix for Google and AI search.