AI, or Just a Chatbot? A Buyer's Guide for Restoration Operators
Photo by jonakoh _ on Unsplash

AI, or Just a Chatbot? A Buyer's Guide for Restoration Operators

Everyone says they have AI now. Five questions separate software that knows your operation from a chatbot with your logo on it.

Verinode Research·June 24, 2026·4 min read·Print / PDF

More restoration software now ships with AI attached. A lot of it answers calls and summarizes email. These five questions separate a chatbot from software that actually knows your operation.

The word AI is on almost every restoration tool now. Your estimating software has it. Your phone system has it. New products show up every month built around an "AI layer" or an "AI operating system" for restoration. Some of it is real help. Some of it is a thin coat of paint. You are the one paying for it, so it is worth knowing how to tell the two apart before you sign.

This is a plain guide to that question. No vendor names, no takedowns. Just the questions to ask, and why the answers matter for a restoration business.

What the thin layer usually is

Most of what gets sold as an "AI layer" today is one of two things, often both: an answering service that picks up the phone and books the first notice of loss, and a reader that turns a call or an email into clean notes.

Both have real value. An assistant that answers every call at 2 a.m. during a storm and writes a tidy intake note saves your office real time and stops leads from slipping. That is good automation, and if your phones are a problem, it is worth buying.

But notice what it is doing. It is handling the front door. It listens, it transcribes, it summarizes, and it files. It does not know anything about how your business actually runs. It has never seen your margins. It cannot tell you that the job it just booked is the kind you lose money on.

That is the line. Answering the phone and summarizing a call is automation. Knowing your operation is something else.

The tell: does it know your numbers?

Here is the simplest way to separate the two.

A chatbot answers questions from general knowledge. Ask it about drying standards or how to phrase an email to an adjuster and it will give you a reasonable answer, the same answer it would give any operator in the country. That can be handy. It is also the same thing a free chat tool does, with your logo on it.

Software that knows your operation reads what your business already produces, the email threads, the books, the estimates, the field notes, and reasons over your actual numbers. Ask it about your water jobs and it can tell you what your gross margin on them was last quarter, because it has read your data, not because it knows water jobs in general.

One talks. The other knows. When a vendor says "AI," that is the first thing to pin down.

The five questions to ask any AI vendor

Before you buy anything with AI attached, ask these. The answers tell you fast which kind you are looking at.

1. Does it know my numbers, or does it only talk? If it cannot reference your own gross margin, your own vendor spend, your own cycle times, it is a general assistant with your name on it. Useful, sometimes. But you are not paying for intelligence about your business.

2. Can it compare me to operators like me, or only repeat my own data back? Showing you your own numbers in a nicer chart is reporting. The value is in knowing whether your number is good or bad. That takes a peer benchmark, and a benchmark takes a network of operators, not a single dashboard.

3. Does it reason over what is in my business, or transcribe and rephrase? Transcribing a call and summarizing an email are reading tasks. Catching that one carrier pays nine days slower than the rest of your book, or that an equipment rental is running above the typical rate, is reasoning. Ask for an example of the second kind.

4. Does it decide and act, or wait for me to ask? A chatbot waits for a prompt. A tool that earns its keep brings you the few things worth deciding this week, with the work already drafted, and gets out of your way the rest of the time. Ask whether it comes to you, or only answers when spoken to.

5. Where does its data come from, and where does mine go? If the answer is general training data, you are getting general answers. If the answer is your own operation plus an anonymized peer network, you are getting something built for you. And ask the second half plainly: who else sees my data, and is it ever sold. You are owed a straight answer.

Why the difference shows up in dollars

A transcript reader cannot tell you that you are paying above the peer median on a software subscription you renew without thinking. It cannot tell you that one of your carriers is quietly the slowest payer in your book. It cannot tell you that your margin on a whole category of jobs has been sliding for three quarters. None of that lives in a phone call. It lives in your numbers, and it only means something when it is measured against operators like you.

That is the work that takes an owner hours every week, the digging through tools and spreadsheets to find the one thing worth acting on. A chatbot does not save you that work. It just gives you one more place to type a question.

The honest version of the question

The data that runs your business is already sitting in your inbox and your tools. The real question to ask any AI vendor is not whether they have AI. Everyone says yes to that now. The question is whether their AI actually reads your data and works for you, or whether it answers the phone, writes a nice summary, and calls that intelligence.

Ask the five questions. The answers will sort it out quickly.

Related Reading