Siri AI Resolution Study: What We Found on iOS 27 Launch Day

We ran 8 structured brand and category queries through Siri AI on iOS 27 launch day — Monday 14 September 2026, within hours of the update going live. The results reveal three findings that every agency managing AI search visibility needs to understand before their next client report.

This is Day 0 baseline data. No other GEO platform has published Siri AI citation results from launch day. Here is what we found.

How We Ran the Study

Device: iPhone 17 Pro Max running iOS 27, updated on launch day. All queries typed into the new Siri app — the dedicated chat interface introduced in iOS 27 — to ensure consistent, screenshottable results. Queries were drawn from the category and brand terms most relevant to Canadian digital marketing agencies managing AI search visibility for clients.

The Raw Results

Q1: What is the best GEO tool for Canadian agencies?

Siri returned a web link list: infinitydigital.ca, reddit.com/r/localseo, The Business Rover. No synthesised answer. aisearchrank.ai not present.

Web links only

Q2: Which AI search engines should a brand be tracking in 2026?

Web link list: Slack blog, Yotpo, Integrate.io. No synthesised answer. No specialist GEO platform cited.

Web links only

Q3: How do I improve my brand's visibility in AI search?

Web link list: Reddit/r/bmarketing, Adobe Business blog, Profound. Profound cited directly by name. aisearchrank.ai not present.

Web links — competitor cited

Q4: What is generative engine optimisation?

Wikipedia knowledge panel — synthesised definition, single source. Only query in the study to return a generated answer rather than a link list.

Synthesised — Wikipedia only

Q5: Is aisearchrank.ai a real company?

Web link list: aisearchrank.io Terms of Service, aisearchrank.io homepage, data-mania.com. A different company (aisearchrank.io) surfaced. aisearchrank.ai did not appear.

Wrong entity surfaced

Q6: What is Perplexity used for?

Not captured in this round — will be included in the follow-up study.

Pending

Q7: What does a digital marketing agency need to know about AI search in 2026?

Web link list: LinkedIn Pulse article, triviadigitalagency.com, a general digital marketing trends roundup. No specialist GEO tools or platforms cited.

Web links only

Q8: Does Siri tell people about new products and services?

Siri responded with a clarifying question: To who? Personal context layer triggered — Siri interpreted the query as a request about a specific person in contacts, not a general knowledge question.

Personal context — no web retrieval

Three Findings That Change the GEO Brief

Finding 1

Siri AI is a link engine, not a synthesis engine — for most queries. In 6 of 8 queries, Siri returned a list of web URLs rather than a generated answer. Wikipedia was the only synthesised response in the entire study. This is fundamentally different from ChatGPT, Perplexity, and Google AI Overviews, all of which generate paragraph-form answers with inline citations. Agencies that assume Siri AI behaves like other AI engines are optimising for the wrong output format.

Finding 2

Query phrasing determines which Siri layer responds — and most brand queries hit the wrong one. Query 8 triggered Siri's personal context layer instead of web retrieval because the phrasing contained an ambiguous pronoun. Siri asked "To who?" instead of searching the web. This means the same brand query, worded slightly differently, can return completely different results — or no relevant results at all. GEO optimisation for Siri requires understanding which layer a query is likely to hit before optimising content for it.

Finding 3

Entity clarity is the most urgent fix for brands on Siri AI. When queried directly by name, aisearchrank.ai did not appear — a separate company at aisearchrank.io surfaced instead. Siri could not distinguish between the two entities. This is an entity confidence failure: without clear, consistent structured data establishing a brand's identity across the web, Siri's retrieval will surface the nearest name match, not the correct one. For agencies managing clients with common names, abbreviations, or similar-sounding competitors, this is an active risk today.

What This Means for Agency GEO Work Right Now

The Siri AI citation surface operates differently enough from every other AI engine that it requires its own optimisation logic. Three immediate actions based on today's findings:

• Audit your clients' entity clarity before anything else. If Siri surfaces the wrong company when queried by name, no amount of content optimisation will fix it. Check Schema.org Organisation markup, Wikipedia presence, and consistent NAP data across the web. This is the entity confidence problem — and it is the single biggest Siri AI risk for established brands right now. Read our guide on entity confidence and AI search.
• Do not assume Siri synthesises answers the way ChatGPT does. Siri's default output for most queries is a ranked link list, not a paragraph. That means the page title and meta description are the citation — not a pulled sentence from the body copy. Optimising for Siri means optimising titles and descriptions to directly answer the query, not just include keywords.
• Add query phrasing testing to your Siri AI monitoring workflow. Because Siri routes queries to different layers based on phrasing, the same brand topic can return web results, personal context responses, or nothing at all, depending on how the question is worded. Test multiple phrasings for every client query set. The citation pattern fragmentation we documented earlier this year is more pronounced on Siri than on any other platform we have tested.

What Comes Next

This is Day 0 data — a baseline captured on iOS 27 launch day. We will rerun the same query set in 30 days to measure whether Siri's retrieval results shift as Apple's index catches up with new content published after the launch. We will also add the missing queries from this round and expand the study to include Canadian-localised brand queries.

If you want to see where your clients currently stand across ChatGPT, Perplexity, Claude, and Gemini while we continue the Siri study, a free AI visibility audit gives you that baseline in minutes. The Siri AI monitoring workflow will be added to the platform as findings from this study inform the right query structure.

FAQ

What device and iOS version was used for this study?

iPhone 17 Pro Max running iOS 27, updated on launch day, 14 September 2026. This places the device in the top hardware tier for Siri AI, with full access to all Siri AI capabilities including web retrieval.

Why did Siri return link lists instead of generated answers?

Siri AI uses a broad world knowledge layer that can either synthesise a direct answer or return a ranked web link list, depending on the query type. Definition queries with a clear Wikipedia entry triggered synthesis. Category, brand, and how-to queries triggered link lists. This behaviour differs significantly from ChatGPT and Perplexity, which synthesise paragraph-form answers for most query types.

What is entity clarity and why does it matter for Siri AI?

Entity clarity refers to how consistently and unambiguously a brand is identified across the web — through structured data, Schema markup, Wikipedia presence, and consistent naming. Siri AI relies on its index to match a brand name to the correct entity. If that signal is weak or ambiguous, Siri will surface the nearest name match rather than the correct brand. This is an urgent risk for any brand with a common name, similar-sounding competitors, or inconsistent web presence.

Will you rerun this study?

Yes. We will rerun the full query set in 30 days to track how Siri's retrieval evolves as Apple's index updates post-launch. Results will be published at resources.aisearchrank.ai.