WWDC 2026: What Siri AI Means for Search Visibility

At WWDC on June 8, Apple previewed Siri AI and said it will sit inside Spotlight on iPad and Mac. The same month, Google began rolling out a Search Console report that measures AI Overview visibility. One new answer surface is arriving with a measurement layer. The other is arriving with none. That gap is the part worth your attention.

What Apple actually announced

Strip away the keynote framing and three concrete things are changing for anyone who cares about search visibility. All of it is previewed, not yet live: Siri AI is in developer beta now, with a public beta due later this year and full release in the fall alongside iOS 27.

Siri will read the live web and answer in place. The rebuilt assistant is designed to pull up-to-date information and generate answers to broad topics, rather than just handing you a list of links. Apple’s own Applebot documentation, rewritten the same week, now states that crawled data may provide context for AI-generated output, and that those answers may include links to the sources used. May. It does not say when, how often, or for which queries.

Siri will live inside Spotlight. On iPad and Mac, the box that launches apps is also set to answer questions. This is the quiet one. A Mac user who used to open a browser tab for a quick factual lookup could get the answer without the tab, and without the click that tab would have produced for whoever ranked for that query.

Visual Intelligence creates queries with no results page at all. Point a camera at a product or a storefront, ask Siri about it, and the entire interaction happens without a SERP. For e-commerce and local businesses, that’s a new query type where the sourcing is completely opaque.

Worth being precise about: This is a staged rollout, not a finished launch. Siri AI arrives as an English-only beta later this year, and even within the US it’s gated by hardware (broadly the iPhone 15 Pro line, iPhone 16 lineup, and newer) and an opt-in waitlist. It will not be available in the EU on iOS and iPadOS at launch, and is unavailable in China pending regulatory approval. So the near-term impact is concentrated in the US and other non-EU markets, which, if anything, sharpens the relevance for US-focused teams rather than softening it.

The measurement problem is the whole problem

Here’s the thing we keep coming back to. A new answer surface is only as manageable as your ability to see yourself on it. With Google’s AI Overviews, that visibility was missing for two years, and as of this month, it partly exists. With Siri, it doesn’t exist at all.

There is no Apple webmaster console. No impressions metric for Siri answers. No citation report. No stated referrer string you can isolate in analytics. If Siri answers a question by drawing on your page and the user never taps a source link, there may be nothing in your logs to record that it happened. You can do excellent work and have no signal that it landed.

This is not a reason to panic, and it is definitely not a reason to buy a “Siri optimization playbook.” Nobody has one yet, and anyone selling one in June 2026 is guessing. It’s a reason to be honest about what you can and can’t know, and to set client and stakeholder expectations accordingly before someone asks you for a Siri visibility dashboard you cannot build.

Set that against what Google did in the same window. The divergence is striking, with one caveat worth stating plainly: Google’s new report is a partial rollout, going to a subset of website owners first, starting in the UK. So not every team can see it yet, either. The difference is direction. Google is building the measurement layer; Apple, so far, has not described one at all.

Google — AI Overviews & AI Mode (rolling out from Jun 3)	Apple — Siri AI & Spotlight (no reporting surface)
Dedicated Search Console report for generative AI features	No console, no first-party visibility report
Impressions, pages, countries, devices, dates	No impressions, no citation count, no device split
Still no clicks or CTR, impressions only, for now	No documented referrer for Siri answer links
Rolling out to a subset of owners, UK first	“Links may appear”, frequency unspecified

Both surfaces answer questions in place, so neither will reward a referral-traffic mindset. But Google has at least given you a presence signal to baseline against. Apple has given you the surface without the instrument panel. If your measurement model only counts clicks and rankings, both of these surfaces are blind spots; Apple’s is simply the darker of the two.

The one technical decision actually worth making this week

Most of the Siri news isn’t actionable yet. This part is, and it’s where we’d correct a simplification that’s been circulating.

You’ll see the advice framed as a binary: stay open to Apple and feed Siri, or block Applebot and disappear from it. That framing misses how Apple’s controls are structured. There are two separate user agents doing two different jobs.

Applebot — the search and retrieval crawler

Powers Spotlight, Siri, and Safari Suggestions, and now feeds context into Siri’s AI answers. Blocking this is what removes you from Apple’s answer and search surfaces entirely. For most businesses, you want this allowed.

Applebot-Extended — the training opt-out token

A separate directive that governs only whether your content trains Apple’s foundation models. Disallowing Applebot-Extended opts you out of training while keeping you fully visible in Siri, Spotlight, and Safari. It is not a crawler and makes no requests of its own.

The Googlebot fallback that may already be deciding for you

If your robots.txt names no Applebot rules but does have Googlebot rules, Apple’s crawler follows the Googlebot instructions. So your existing directives may already govern your Siri and Spotlight exposure without anyone having chosen that on purpose. Worth an audit.

So the real decision isn’t “in or out.” It’s three smaller, separable choices: do you want to be retrievable in Apple’s answers (almost always yes), do you want your content used for model training (a content-policy and IP question), and do you know what your existing robots rules are silently telling Applebot to do. The nosnippet tag adds a further lever, stopping a page from being used as an answer context while keeping it indexed. None of these requires new markup on the page itself.

The defensible move: Audit your robots.txt this week for explicit Applebot handling rather than relying on the Googlebot fallback. Make the training decision (Applebot-Extended) deliberately. Then stop, there is no further Siri “optimization” to do that isn’t just guesswork, and the developer beta will produce the first real evidence of how Siri cites sources before any playbook is worth writing.

What we’d resist concluding too early

It’s tempting to assume that because Siri AI runs on Gemini, the content Gemini surfaces will be the content Siri surfaces. We’d hold that thought loosely. Apple describes its models as custom builds, the answers are assembled through Apple’s own stack, and nothing Apple has published links Gemini visibility to Siri visibility. Treating them as one pipeline is an assumption the available facts don’t support, and acting on an unverified assumption is exactly the failure mode that produces wasted effort in AI search.

The honest summary: a large new answer surface is coming to devices people already own, it reads the open web, and it sits where people already type their questions. That much you can plan around. How often it cites anyone, whether those citations send measurable visits, and whether Gemini presence carries over, those are open, and the teams that do best will be the ones who hold the open questions open instead of filling them with confident guesses.

What to actually do

Three things this week, none of them heavy:

One, audit your robots.txt for explicit Applebot and Applebot-Extended handling, and decide each lever on purpose. 
Two, check whether you’re in the rollout for Google’s new generative AI report in Search Console, it’s going out to a subset of owners first, so you may not have it yet. If you do, baseline your AI Overview impressions now while the data is early; if you don’t, the absence tells you nothing about your site, only about the rollout. 
Three, set expectations with your stakeholders that Siri visibility is, for now, unmeasurable, so nobody builds a forecast on a number that doesn’t exist.

The teams that come out ahead in the next year of search won’t be the ones with the longest checklist of Siri tactics. They’ll be the ones who can tell the difference between a surface they can measure and one they can’t, and who invest accordingly.

See where you stand across the surfaces you can measure

Quattr’s AI Search Visibility platform tracks citation share, mention coverage, and brand gaps across AI Overviews, ChatGPT, and Perplexity, so the measurable surfaces stay measured while the new ones settle.