Voice Search and AEO for Home Service Pros

Voice search predictions of "50% of searches will be voice by 2020" turned out to be wrong by an order of magnitude. But voice queries did not disappear; they merged with AEO. Today, when a homeowner asks Alexa, Google Assistant, Siri, or — increasingly — ChatGPT Voice "find me an interior decorator near me," the engine answers from the same generative pipeline it uses for typed AEO queries, just with a TTS layer on top. Optimizing for voice in 2026 means optimizing for AEO with two specific phrasing adjustments. This piece covers what those adjustments are and why studios that nail them are getting an outsized lift on in-car and in-home queries.

How voice queries differ from typed queries

Two things change when a query is spoken instead of typed.

First, the query is longer. Typed queries average 3-4 words; voice queries average 7-9. "Decorator Langley" becomes "find me a decorator in Langley with good reviews." The implication for content: pages that win voice queries answer the full conversational question, not just the keyword stem.

Second, the question word is usually present. Typed queries rarely begin with "what," "how," "where," "who"; voice queries do, more than 60% of the time. The implication: H2s phrased as full questions ("How much does an interior decorator cost in Langley?") win voice surfacing in a way that noun-phrase H2s ("Pricing") do not.

Beyond those two shifts, voice queries are AEO queries. Same engines under the hood, same source-selection logic, same trust signals. Studios that have done the AEO work this site has covered will surface in voice answers; studios that haven't, won't.

The device mix in 2026

Smart speakers (Echo, Google Home, HomePod) plateaued around 2022 and have been a steady but small share of queries since. The growth surface is in-car (Android Auto, CarPlay, in-vehicle assistants), in-AirPods/in-earbuds (especially Siri and the new ChatGPT Voice mode), and in-home running through TVs and screens that now ship with always-on voice assistants.

The implication for designers: in-car queries skew toward "near me now" intent — homeowners running errands, suddenly thinking about a project. In-earbuds queries skew toward "answer my question" intent — researching mid-task. Both surfaces care about answer quality and entity disambiguation more than they care about brand awareness.

The phrasing adjustments

Two specific changes to your existing AEO pages.

One: every commercial page gets an FAQ block where the questions are phrased exactly as a homeowner would speak them. Not "Pricing" but "How much does interior decorating cost in [city]?" Not "Service Area" but "Do you serve [neighborhood]?" The conversational phrasing matches the voice query, and the FAQPage schema (covered in our schema post) lets the assistant lift the answer cleanly.

Two: every answer paragraph gets a one-sentence "spoken summary" near the top — a sentence that, if read aloud, completely answers the question. Voice assistants prefer summaries that fit in a 5-8 second TTS window, which works out to roughly 25-35 words. "Decorator.tv is a licensed Langley interior decorating studio with five-year warranties, serving the Fraser Valley since 2009." 19 words, complete answer, voice-friendly.

These two changes do not require a redesign. They require an editing pass on existing pages. A studio with twenty service pages can do the editing pass in two evenings.

The local pack and voice

Voice queries with "near me" intent route to the same local pack data Google uses for Maps. Everything in our local SEO post applies: complete GBP, NAP consistency, review velocity, weekly Posts. There is no separate voice search ranking; it's local SEO with a TTS skin.

The one nuance: voice surfaces typically read the top one or two local pack results, not the top three. Position 1 and position 2 are dramatically more valuable than position 3 in voice. The compounding effect of being already-ranked for voice is real.

What to ignore

Three voice-search tactics from the 2018-2020 era should be retired.

"Optimize for natural language by writing conversationally throughout your pages." This was always overstated. Pages should be readable, but the pages that win voice are the pages that win AEO — structured, specific, schema-supported. Conversational filler hurts extractability.

"Target featured snippets at all costs." Featured snippets matter, but the snippet pipeline merged with AI Overview and Bard answer pipelines years ago. Optimize for AEO and you get featured snippets as a byproduct.

"Build voice apps / Alexa skills." For 99% of design studios, building a custom voice app produces zero ROI. The discoverability surface is the assistant's web search, not the assistant's app store.

Measuring voice impact

You cannot directly measure voice queries — Google and Apple deliberately do not break voice out in analytics. The proxies:

Direct phone calls without a referrer URL spike when voice surfacing improves, because voice users tap "call" rather than "open website."

Long-form, question-style queries in Google Search Console that show high impressions and low click-through indicate the page is being read aloud and answering the user without requiring a click.

Mentions in branded search after a sustained AEO push indicate the brand is being heard in voice answers.

The 60-minute voice optimization sprint

For each of your top five service-plus-city pages, do four things:

Add or rephrase the FAQ block so each question reads as a full spoken sentence (15 minutes per page).

Add a one-sentence spoken summary at the top of the answer paragraph (5 minutes per page).

Validate FAQPage schema is present and live (5 minutes per page).

Re-submit the URL to Google's Indexing API to refresh the cached version (2 minutes per page).

That's roughly 27 minutes per page, 2 hours and 15 minutes for five pages. Voice surfacing improvements show up in 4-6 weeks. There is no other voice-specific work that pays back faster.