Siri Now Runs on Gemini. When a Customer Asks Out Loud, Only One Location Gets Named.

A customer stands outside, phone in hand, and says: “Hey Siri, find me a good coffee shop near here.” Six months ago, Siri handed them a list to scroll. Now it names one. Just one. And if that one is not your branch, the customer never sees that you exist.

That shift moved from theory to product on 8 June 2026, when Apple used its WWDC keynote to unveil a rebuilt Siri running on Google’s Gemini models. For multi-location brands, this is not a search update. It is a change to who gets chosen.

What actually changed at WWDC 2026

Apple and Google first confirmed their partnership on 12 January 2026. According to widely reported coverage, Apple is licensing a custom Gemini model from Google to power the new Siri, at a cost reported at roughly one billion dollars a year. At WWDC on 8 June, Apple made it official on stage.

The practical detail matters for planning. Apple has indicated the new Siri launches first in US English, with broader general availability expected in late 2026, and it is not available in the EU or China at launch. So the timeline is uneven by market, but the direction is set. Voice on the iPhone is becoming a conversational answer engine rather than a command line.

The reach is the part brands underestimate. Siri reportedly handles around 1.5 billion requests a day, and Gemini is also rolling out as the assistant layer across Google’s own Android base. A very large share of “ask out loud” moments now flows through the same answer logic.

Voice does not return a list. It returns one name.

A screen can show ten results. A spoken answer cannot. When a person asks a question out loud and listens to the reply, the assistant has to choose. It synthesises across your website, your business profile, review platforms, directories and local discussion, then says a single recommendation.

This is the quiet break with classic local SEO. Ranking third in the local pack still earned you clicks. Being the third-best match for a voice query earns you nothing, because the customer never hears positions two through ten. The gap between first and second is no longer a traffic difference. It is the difference between existing and not existing in that moment.

Why this hits multi-location brands hardest

For a single-site business, voice selection is one contest. For a brand with dozens or hundreds of branches, it is a contest that runs separately around every customer, in every suburb, all day.

Two things have to be true at once. Your brand has to be trusted enough at a national level for the assistant to consider it, and the specific nearest location has to have clean, complete, well-reviewed data so the assistant is comfortable naming it. A strong brand with one weak branch will simply lose that branch’s customers to a competitor the assistant trusts more in that postcode.

• Inconsistent data sinks the nearest branch. If the closest location has wrong hours, a missing category, or a thin profile, the assistant routes the customer to a cleaner-looking rival nearby.
• Reputation is now pooled. AI answer engines read sentiment across your whole brand. A handful of poorly managed locations can drag down how the model perceives all of them.
• You cannot see the loss. A spoken recommendation leaves almost no trace in traditional rank tracking. The branch just gets quieter.

The data the assistant reads before it picks

Gemini-powered answers are only as good as the structured, verifiable information they can find about each location. On the iPhone, that pulls from Apple’s ecosystem as well as the open web. Apple unified its business tools into a single Apple Business platform in April 2026, and Apple Maps leans on its own review and data signals rather than Google’s, so the iPhone and Android answer paths do not read identical sources.

That means the inputs you control directly are now the battleground: accurate listings on every platform, complete and consistent location data, reviews managed across every surface the model can reach, and location pages with real, machine-readable detail rather than a city name swapped into a template.

What to do now, before the rollout widens

The window between a US-English launch and broader availability is preparation time. Use it.

• Audit every location’s data for accuracy and completeness across Google and Apple, not just your strongest markets.
• Close the gap on your weakest branches first. They are the ones quietly losing voice moments today.
• Treat reviews as a brand-wide signal. Respond, resolve, and keep velocity steady across all locations.
• Make location pages genuinely distinct and structured so an answer engine can lift specific, trustworthy facts about each site.

How Social Places helps you be the named location

Being chosen by a voice assistant comes down to data the model can trust about every branch you run. Listings keeps every location accurate and consistent across Google, Apple and beyond, so the nearest branch is always a clean candidate. Reputation keeps review quality and velocity strong across the whole estate, not just your best sites. Store Locator and Local Pages give each location structured, machine-readable detail an answer engine can quote, and Brand Listening shows you what is being said across the surfaces these models read. If you want to be the one name the assistant says out loud, that work starts now. Contact Us.