Web Design

Voice and Multimodal Search: Is Your Service Website Ready to Be Heard?

A homeowner photographs a cracked foundation and asks their phone what it costs to fix. A clinic manager drives between sites dictating questions about payroll providers. Neither will ever see a search results page; an assistant will hear or see the query, consult the web, and answer. Voice search was over-hyped in 2018 and then arrived late in a different costume: not a separate channel to optimize, but one input mode of multimodal AI assistants that take speech, text, and images interchangeably. This article covers what that shift actually demands from a service-business website, with claims date-stamped June 2026.

Joshua Agonya Pi'Rwot

By Joshua Agonya Pi'Rwot

Founder, Business Growth Accelerator

Executive summary

The voice search hype fizzled, then quietly came true in a different form: multimodal AI assistants that hear questions, see images, and speak one answer. How service businesses should design for being the answer, not the link.

Section 1

What the voice hype got wrong, and what it accidentally got right

The 2018 prediction was that half of searches would be voice and every business needed a 'voice strategy.' That specific future fizzled: smart speakers became kitchen timers, and dedicated voice SEO never produced a distinct discipline. But the underlying bet, that natural language would replace keyword-ese as the way people ask for things, came true through large language models instead. Today's assistants accept a spoken sentence, a photo, a screenshot, or all three, and respond in kind. Stanford HAI's AI Index documents the steep capability gains in language and multimodal systems that made this practical. The lesson for operators is about how queries changed shape: 'plumber tampa' became 'who can fix a tankless water heater in South Tampa this week and what should it cost.' Longer, situational, spoken-style questions are now the input your content competes to answer. For the step that usually comes next, see [The Future of Storytelling in Business: Trends to Watch](/blog/the-future-of-storytelling-in-business-trends-to-watch).

Section 2

The three query modes and what each demands from your site

It clarifies planning to treat voice, visual, and conversational text as distinct modes that converge on the same backend: an assistant reading the open web and choosing what to repeat. The table below maps each mode to design implications, with priority ratings for a typical 5-7 figure service firm. Notice what is absent: no exotic 'voice markup' tricks. Google deprecated most speakable-specific experiments, and its current guidance for AI features is that standard fundamentals, crawlability, useful content, structured data where applicable, carry the load. The leverage sits in unglamorous places: a services page that answers cost questions in plain sentences, a Google Business Profile that agrees with your website, image alt text and captions that say what the photo shows, and pages that load fast on the mobile networks where spoken queries happen.

Section 3

Local service businesses have the most to win, and lose

Spoken and assistant-mediated queries skew heavily local and transactional: near me, open now, how much, who is best. When an assistant answers those, it typically names one to three providers, not ten. That compression is brutal for whoever ranks fourth and golden for whoever is cited. The inputs assistants weigh are mostly verifiable entity data: consistent name, address, and service details across your website and Google Business Profile, genuine reviews, service-area pages that say plainly where you work, and Organization or LocalBusiness structured data binding it together. None of this is glamorous, which is exactly why most local competitors have not done it properly; in our site audits, inconsistent or machine-illegible business data remains the most common failure. Being boringly, verifiably consistent everywhere machines look is currently the highest-ROI 'voice strategy' a local service business can execute. A useful companion to this piece is [The Minimum Viable Website Stack for a Service Business](/blog/minimum-viable-website-stack-service-business).

Section 4

What to do this quarter, in order

First, collect your real spoken-style queries: pull the exact questions from sales calls, intake forms, and chat transcripts; these are what buyers ask assistants. Second, build or rewrite Q&A content so each question gets a direct, complete answer in the first two sentences, with specifics an assistant can repeat accurately. Third, audit entity consistency: website, Google Business Profile, directories, and schema markup must agree on name, services, areas, and hours. Fourth, fix mobile speed; multimodal queries happen overwhelmingly on phones, and web.dev's Core Web Vitals data shows roughly 40% of sites still fail the loading threshold, which is a quiet competitive opening. Fifth, test monthly: ask the major assistants your top ten buyer questions and log who gets named. If your firm never appears, that finding is the agenda for a strategy call. If you are turning this into practice, [Attention Is Not a Lead: Why Service Businesses Confuse Audience With Pipeline](/blog/attention-is-not-a-lead-why-service-businesses-confuse-audience-with-pipeline) maps the adjacent system.

FAQ

Direct answers for operators.

How do I find out what voice questions buyers actually ask?

Skip keyword tools and mine your own operations: sales call recordings, intake forms, chat transcripts, and the questions your team answers repeatedly by phone. Those are verbatim the questions buyers also ask assistants, phrased exactly as spoken. Build one page or section per high-value question, answer it directly in the opening sentences, and include the specifics, price logic, timelines, service areas, that assistants need to repeat you accurately.

Do images on my website matter for multimodal search?

Increasingly, yes. Assistants now accept photos of problems, a damaged roof, a tax notice, a piece of equipment, and match them against web content that names and explains those problems. Give your images descriptive file names, accurate alt text, and nearby captions or text explaining what is shown and what it implies. Real photography of your actual work outperforms stock imagery for both machine matching and human trust.

Joshua Agonya Pi'Rwot

Written by

Joshua Agonya Pi'Rwot

Founder, Business Growth Accelerator · Country Director, AVODA Group Uganda · EMBA

Joshua helps service-business operators turn scattered marketing into a clear path from first attention to booked call. He is Founder of Business Growth Accelerator and Country Director of AVODA Group Uganda.