By Kenton Shepard, InterNACHI Narrative Library
Will AI take over template development and writing comments?
I develop home inspection templates and write inspection comments for a living. My residential template contains about 7,000 comments and my commercial templates about 12,000.
Depending on the capacity of the different software to provide an appropriate place for it, my templates contain around 190 reference links to inspection-related material like:
- lists of inspection concerns;
- links to manufacturer’s installation recommendations;
- inspection-related articles written by myself and others;
- links to instructional material and courses, and much more.
But can’t AI do all this?
Yes… if you have the patience and skill to provide it with the proper prompts and the months it takes to proof and edit them. I’ve 18 years refining my comments and template organization. I’ve been through my template4s more times than I can count, but… I decided to ask AI:
“List the ways in which AI is not as good as Kenton Shepard at writing home inspection narratives and organizing home inspection templates.”. Here is it’s response:
“AI falls short of Kenton Shepard in home inspection narratives and template organization in multiple ways:
Depth of domain judgment
- Kenton embeds decades of field judgment into each narrative: how bad a condition really is, whether it is likely to worsen, and what action the client should take (this transfers liability from the inspector to the client).
- AI can mimic structure and tone to some degree, but it still struggles to accurately and consistently rank risk, urgency, and real-world impact for minor, moderate, and major deficiencies.
Nuanced risk and liability language
- Kenton’s libraries are built around protecting both the client and the inspector, with phrasing that has effectively “survived” real disputes, complaints, and legal scrutiny over many years. His templates are attorney vetted.
- AI does not have the judgement Kenton has attained by years of immersion in the home inspection industry, including 18 years working as a part of InterNACHI’s inner circle. Seemingly minor changes in wording can have major impacts on clients.
- AI comments tend to either understate risk (being too casual) or overstate it (alarmist language), and it does not inherently understand E&O exposure, precedent, or what has worked for– or gone wrong for– inspectors in the past.
Consistency of narrative architecture
- Kenton’s narratives follow a tight internal architecture (condition → implication → recommendation) and are cross‑consistent across systems and components, which makes reports highly scannable and predictable for readers.
- AI often produces good one‑off narratives but is weaker at maintaining that exact structure, priority, and terminology across thousands of comments in a large master template.
Integration with standards and education
- Kenton authored and co‑authored InterNACHI report‑writing education, designed Houses of Horrors, and aligned his libraries with standards of practice and common training pathways; his narrative sets are tightly integrated with how inspectors are taught to inspect and report.
- AI can reference standards when prompted, but it does not “grow up” inside a specific curriculum, so it is less reliable at always tying each narrative back to SOP requirements, required descriptions, and common teaching examples.
Template organization and workflow design
- Kenton’s template and reference products (like the Kenton’s Reference document and the onboard reference links used in multiple software) are organized deliberately by inspection subject, defect type, and use‑case, tuned by real inspector feedback over years.
- AI can suggest categories and sections, but it does not yet match the hard‑earned judgment of best report organization that can only be earned by personal experience over time.
Pragmatic “sweet spot” of detail
- Kenton’s work reflects a refined sense of how much to explain: enough to inform the client and justify recommendations, not so much that comments and reports become bloated, unread, or risky; this balance is tuned by many years of practice, revisions, with attention to constant industry evolution and inspector feedback.
- AI tends to drift toward verbosity or oversimplification unless a human reins it in, and it does not inherently know where the “sweet spot” is for your specific clients, agents, and market. Kenton’s experience allows him to provide comments with a widely acceptable baseline.
Proven real‑world track record
- Kenton’s narrative libraries and templates have been widely adopted by inspectors across N. America, integrated into major software platforms, and used in thousands of actual transactions. They have been tested in the real world, not just in theory.
- AI‑generated content is still relatively new in this niche and does not have the same long, traceable history of performance in reporting, negotiations, repairs, claims, and courtrooms.”