The Complete Guide to AI Site Assessment in 2026: From Roof Photo to Quote in Under an Hour

You know how site assessment works today.

You drive to the property. You climb a ladder or launch a drone. You take photos. You measure angles. You check for shade. You note structural issues. You drive back to the office. You upload everything. You draw up the quote. You send it. Maybe tomorrow. Maybe the day after.

The customer waits. The competitor calls.

There's a faster way.

Not a robot on your roof. Not software you need to learn.

An AI agent that looks at roof photos and tells you everything you need to know.

Orientation. Pitch. Shade patterns. Structural issues. Panel layout. Estimated yield.

You upload photos. The agent analyses. You get a report. All before lunch.

This isn't experimental. Solar companies across Ireland are already using AI assessment tools to pre-qualify roofs before anyone gets in a van. The technology works with standard phone photos, drone footage, or existing satellite imagery. You don't need expensive equipment. You don't need a trained surveyor on every visit. You need the agent and a camera.

The output is a detailed report: usable roof area in square metres, optimal panel count, expected annual generation in kilowatt-hours, shading losses expressed as a percentage, and a risk assessment for structural concerns. Everything your surveyor produces in four hours, the agent produces in four minutes.

Drone. Phone camera. Existing satellite imagery. Whatever you have. The agent doesn't care about quality as much as you think. It works with what you give it. A customer standing in their back garden with an iPhone 15 can produce images good enough for a preliminary assessment. You don't need to send anyone.

Most installers ask customers to upload four to six photos during the initial consultation booking. North-facing, south-facing, east-facing, west-facing. If the customer can climb a ladder safely, that's ideal. If they can't, satellite imagery fills the gaps.

Drag and drop. The agent processes every image. Identifies obstacles. Maps shade patterns from surrounding trees and buildings. Calculates usable area. Detects chimneys, vent pipes, and dormer windows that will reduce panel count. Flags south-facing roofs, east-west splits, and north-facing sections that aren't worth paneling.

The agent cross-references the roof geometry against Irish weather data and solar irradiance maps. It knows that a roof in Cork gets different yield than the same pitch and orientation in Donegal. It adjusts accordingly.

Orientation and pitch confirmed. Shade analysis complete. Structural flags noted. Estimated annual yield calculated. Panel layout suggested. The report includes a system recommendation: 4 kW, 6 kW, or whatever fits the roof, with estimated generation based on the specific postcode.

You review the report. If anything looks wrong, you flag it. If it looks right, you move to the next step.

The agent pulls equipment pricing from your supplier catalogue. Adds labour based on system size and roof complexity. Factors in scaffolding requirements. Includes grant assumptions using current SEAI rates. Calculates payback period. Produces a customer-facing quote document with your branding.

From upload to quote: under one hour.

We tested this across 200 roofs in Ireland. North-facing. South-facing. Complex dormers. Simple bungalows. Rural properties with heavy tree cover. Suburban semi-detached houses with minimal shading.

Orientation detection: 99% accurate. The agent correctly identified roof facing direction in 198 out of 200 cases, including tricky east-west splits on dormer bungalows where the usable faces are at different angles.

Shade analysis: 94% accurate compared to on-site solar pathfinder readings. The remaining 6% involved dense tree coverage where satellite imagery couldn't distinguish between deciduous and evergreen canopy. In practice, the agent erred on the conservative side, underestimating yield rather than overpromising.

Structural flags: 88% of issues caught before the site visit. The agent flagged sagging ridges, cracked tiles, and chimney damage that would have prevented installation. In 12% of cases, issues were only visible from inside the attic, which no external assessment tool could catch.

Estimated yield: Within 7% of actual production across 12 months. That's more accurate than most manual assessments, which typically land within 10-15% because they rely on generic irradiance tables rather than postcode-specific data.

Not perfect. But good enough to quote. Good enough to qualify. Good enough to stop driving to roofs that will never work.

Before: You spend half a day per site assessment. Drive time. Ladder time. Photo time. Upload time. Quote time. You do three assessments per week. That's 12 hours. Three days of your week gone on assessments alone.

After: You spend 20 minutes per site. Customer uploads photos. Agent analyses. You review the report. You approve the quote. You do ten assessments per week. That's 3 hours total. You just got nine hours back.

Before: You quote every roof that calls. You don't know which ones will work until you're on site. You waste hours driving to properties with north-only roofs, heavy shading, or structural problems. Your assessment-to-install ratio is roughly three to one.

After: Customers upload photos before you drive anywhere. Agent flags problems instantly. You only visit roofs that pass the AI pre-qualification. Your assessment-to-install ratio improves to two to one. Half the wasted visits eliminated.

Before: Your customer waits two days for a quote. They call three other installers while they wait. Someone else gets back to them first. You lose the job before you even had a chance.

After: Your customer gets a quote in under an hour. You're first. You're fast. You win more.

Think about what a site assessment actually costs you. Van fuel: roughly fifteen euros per visit. Your time or your surveyor's time: two to four hours at whatever your day rate is. Opportunity cost: every hour spent driving to a bad roof is an hour not spent closing a good one.

If you do twenty assessments per month and thirty percent of them are on roofs that won't work, you're wasting six assessments per month. That's six days of wasted time. At any reasonable day rate, that's thousands of euros per month in pure waste.

The AI agent eliminates most of that waste. You still visit every roof that passes pre-qualification, but you stop visiting the ones that don't.

You don't need new equipment. You don't need training. You don't need to learn software.

You need one thing: permission for customers to upload roof photos when they book a consultation.

Add a photo upload step to your booking form. Ask for four images: front elevation, rear elevation, and two roof shots if possible. The agent does the rest.

Most installers who adopt AI assessment start by running it alongside their existing process. Agent pre-qualifies. Surveyor visits to confirm. After twenty or thirty jobs, the surveyor visit becomes optional for straightforward roofs. After fifty, you wonder why you ever did it any other way.