Roofer near me, spoken out loud.
Conversational queries are longer and more specific. That changes what a page has to answer.
How spoken queries differ
Typed queries compress: roof repair cost. Spoken ones expand: how much does it cost to repair a small leak on a tile roof. The spoken version contains the roof type, the severity, and the intent, which means the page that wins it answers a narrower question than a general service page ever does.
What to publish
Six page types that match how people actually ask.
Near-me is a local problem, not a voice problem
A near-me query is answered by the map pack and the local results underneath it, which means the Business Profile and the city pages do the work rather than anything specific to voice. If you want the near-me queries, that cluster is where the effort goes, and the local search guides cover it.
Why answering the narrow question wins the broad one too
A page that answers how much a small tile-roof leak costs to repair also serves the person who typed roof repair cost, because it contains a real number and a real scenario. The reverse is not true: a general page with no specifics answers neither. Specificity is not a niche play, it is how the general question gets answered credibly.
The overlap with answer engines
This is the same discipline as the rest of this cluster. A page that answers one specific question in its heading and answers it immediately is what a voice assistant reads aloud and what a retrieval system extracts. Two efforts, one piece of work, which is why they are in the same pillar rather than separate ones.
What not to do
Do not write conversational filler on the theory that voice rewards a chatty tone. Do not generate a page per phrasing of the same question. One page answering it properly is better than six answering it thinly. And do not add a question heading you then answer in paragraph four, which is the most common way this work gets done badly.
How to find the real questions
Your own phone. Ask whoever answers it to write down the actual wording of the questions people ask for a fortnight. That list is more accurate than any keyword tool for this purpose, because it is your market speaking in full sentences, and most of it has never been answered on your site or anyone else's.
Assistants answer from different places
Four routes, and they behave differently.
Hours and after-hours are voice questions
A large share of spoken queries are practical: are they open, do they do emergencies, how fast can someone come out. Those answers should exist as plain text on the site and match the Business Profile exactly, because a contradiction between the two is how a system ends up saying it is not sure. Stating the after-hours arrangement in words converts better than an unanswered phone.
One page per question, not per phrasing
How much does a tile roof repair cost, what does tile roof repair cost, and tile roof repair pricing are one question. Answer it once, properly, with a real range. Six thin pages competing for phrasings of the same question split the signal and give a retrieval system three mediocre candidates instead of one good one.
Related guides.
Two acronyms for the same shift: being the source an answer engine quotes rather than a link it lists.
The retrieval and attribution mechanism, described without invented statistics.
How AI answer engines retrieve and cite sources, and what actually makes a roofing site citable.
Find out if your site is citable.
$2,500, and you see every answer a competitor is cited for and you are not.