How to Write Defect Narratives That Protect You

A weak narrative is a liability magnet. The findings that come back to bite inspectors are usually not missed defects, they're found defects described so vaguely that nobody acted on them. The fix is a four-part structure applied to every finding: location, condition, implication, direction.

The four-part defect narrative

1. Location: pinpoint it

"Bathroom wall" helps no one. "Primary bathroom, north wall, adjacent to the shower valve" lets the client, the agent, and the contractor find the exact spot without you on the phone. Precision here also protects you: the finding is anchored to one place, with no room for "that's not what he meant."

2. Condition: what you observed

State the evidence, not a diagnosis. Not "mold on the wall" but "visible dark staining, roughly 12 by 18 inches, below the shower valve." You're reporting what a camera would agree with. Diagnosis belongs to the specialist you're about to recommend.

3. Implication: why it matters

One sentence connecting the observation to consequence: "The staining suggests moisture intrusion, which can lead to material damage and indoor air quality concerns if unaddressed." Without this sentence, clients can't rank findings and agents will rank them for you.

4. Direction: what to do next

End with an action and an actor: evaluate, repair, or monitor, and by whom. "Recommend evaluation by a licensed plumbing contractor to determine the moisture source and repair as needed." Don't estimate repair costs and don't offer to do the work, both create conflicts you don't want.

The failure modes that create risk

Vague locations

"Kitchen sink" could mean the drain, the faucet, the supply lines, or the cabinet floor. Every ambiguity in the report is a future argument. Write "kitchen sink, cold supply line, at the shut-off valve."

Hedging until the finding says nothing

"Appears to be" and "may indicate" are honest when you genuinely can't confirm. Stacked three deep on clear evidence, they gut the finding. If water is actively dripping, write "active leak observed." A report that hedges everything protects no one, including you.

Boilerplate that contradicts the photo

Copy-paste comments drift out of sync with evidence: the text says "minor staining" while the photo shows a saturated ceiling. When a narrative and its photo disagree, the report impeaches itself. Every comment should match the specific conditions in the frame next to it.

Alarmist language

Your job is to inform, not to detonate the transaction. "This panel could burn the house down" is opinion theater. "Panel cover missing, energized conductors exposed, a safety hazard, recommend correction by a licensed electrician" is a fact pattern with a direction. Objective severity reads as more credible, not less.

Standards are your backstop

Whatever your state's standards of practice require, and whatever your association (ASHI, InterNACHI) publishes as reporting guidance, follow it consistently. Consistency is itself protective: a report that follows the same defensible structure every time is much easier to stand behind than one that improvises.

Where AI drafting fits

The four-part structure is exactly the kind of work AI does well. With a tool like InspectAI, you supply the observation at the defect (a photo and a voice note saying what and where), and Gemini drafts the narrative in the full structure: location, condition, implication, direction. The judgment stays yours, the formatting discipline becomes automatic, and boilerplate drift disappears because each comment is generated from that finding's actual evidence. More on what the AI does and doesn't do in AI report writing for home inspectors.

FAQ

What's the most important part of a defect narrative?

The one you're most tempted to skip. In practice that's the implication sentence. Location and condition get captured naturally, but a finding without "why this matters" reads as noise and gets ignored.

Are boilerplate comments bad?

Templates for structure are fine. Comments pasted without adjustment are the risk, because sooner or later one contradicts the photo beside it. Tailor every comment to the observed condition.

How does AI keep narratives specific instead of generic?

Good AI drafting works from your actual inputs, the photo and your dictated observation, so the text describes that defect in that place. You review every word before the report goes out.