Phones Are Diagnosing Home Repairs in 10 Seconds. Your Photos Decide Whether You Get the Call.

The customer pointing a phone at their water heater right now does not need you to tell them what is wrong. Their phone already told them. Your only job left is to be the one they call.

Three things shipped or got commercialized in the last 60 days that should change how every contractor thinks about a service call:

• Gemini for Home rolled out faster smart home control, AI descriptions on camera timelines, and an experimental "COSMO" assistant on May 1. Nest cameras now classify activity automatically and send "your HVAC condenser fan stopped" level alerts to homeowners.
• AptRepair, a BYU-built AI repair diagnostic that analyzes photos of a home and identifies repair issues, estimates costs, and connects users with local contractors, is being commercialized right now.
• Apple Intelligence Visual Intelligence and Gemini Live both now work on real photos and live video. A homeowner walks around their house, points the camera at anything broken, and gets a plain English diagnosis in 10 seconds.

The smartphone is officially a diagnostic tool. The repair conversation starts before the homeowner ever talks to you.

The new customer journey

It used to be: notice a problem, search Google, call three contractors, take three quotes, pick one.

The new flow looks like this. Homeowner points their iPhone at a wet ceiling stain at 8:14 PM. Visual Intelligence tells them it is likely a roof leak or upstairs plumbing failure, suggests turning off the water at the valve, and lists three roofers and two plumbers in a 15-mile radius with high citation scores. Total elapsed time: 90 seconds.

The customer is not calling three companies anymore. They are calling whoever the phone surfaced first. Your shop either is one of those names or it is not.

Why this is different from regular AI search

AI search starts with text. Mobile AI starts with reality.

The phone has the photo, the GPS coordinates, the time of day, the weather conditions, and increasingly the homeowner's history through Apple Health, calendar, and connected accounts. When Visual Intelligence sees a leaking pipe joint, it knows the homeowner's address, the zip code, the time, the urgency level, and the contractors that match.

The recommendation gets more specific than text-based AI search by an order of magnitude. And it gets recommended only once because the homeowner is in the middle of a problem, not a research phase.

What contractors get wrong about this

Most contractors think mobile AI is a Google Maps update. It is not. It is a fundamentally different funnel.

The signals that win text AI search — schema, NAP consistency, review volume — also matter for mobile AI, but two new ones get layered on top:

Photo-rich profiles. Apple Visual Intelligence and Google Lens both prioritize businesses with current, well-tagged photos of their work. A roofer with 60 photos of recent jobs in different neighborhoods, with proper alt text and geotags, beats a competitor with three stock photos every single time.

Visual confirmation. When the AI shows the homeowner a recommended contractor, it tries to show evidence the contractor has done the specific kind of work in the photo. Pipe burst photo? AI looks for plumbers with pipe replacement photos in their profile. Cracked shingle photo? AI looks for roofers with shingle work photos.

If your portfolio is generic, the AI cannot match you to the customer's actual problem.

The 30-day mobile AI prep

Photograph every job from now on. Before, during, after. Different angles. Tag with neighborhood and the specific service. Upload to Google Business Profile, Yelp, and your website. Aim for 40-plus current photos in the next 30 days.

Use plain-English photo captions. Skip the jargon. "Replaced 30-year architectural shingles after hail damage in Eastover" beats "Roof replacement project completed." The AI reads the caption and matches it to the customer's described problem.

Add real before-and-after pairs. Visual AI engines cluster these and treat them as evidence of capability. They are also the highest-converting content on a service page anyway.

Get reviews that mention specific symptoms. "They fixed the slab leak under our kitchen" carries 10x more visual AI weight than "Great service." Train your office to ask for symptom-specific reviews after every job.

Test it yourself. Take a photo of a real problem in your home. Ask Visual Intelligence or Gemini Live what to do. Note who gets recommended. If your shop is not in the answer, you have your roadmap.

The window is right now

Adoption of phone-based visual AI roughly tripled in the last six months. AptRepair-style apps that estimate repair costs from photos are about to get commercialized. By the end of 2026, "snap a photo, get a contractor" is going to be the default repair flow for homeowners under 45.

You either start now and become the visual AI default for your service area, or you spend the next two years bidding against Local Service Ads to fight for whatever scraps the phone-first customers reject.

Photograph the work. Caption it like a human. Get reviews that name the symptom. The homeowner's phone is going to ask the question. Be in the answer.