Is AI consulting worth it in your industry?

This is the most common version of the question and it has a weak answer, because industry is a poor predictor. Two companies in the same sector, the same size, can be complete opposites on whether this pays.

What predicts it is the shape of your processes. Four characteristics, and you can score yourself in ten minutes.

Does the same thing happen many times?

The strongest single indicator. High frequency justifies a build; low frequency does not, however irritating the task is.

A process running two hundred times a week is a candidate even when each instance is trivial. Something happening twice a month almost never is, and this is where most disappointment originates, because companies automate the thing that annoys the loudest person rather than the thing that happens most.

Is a lot of your work text and conversation?

Current systems are strongest on language: reading it, writing it, extracting structure from it, holding a bounded exchange.

Businesses where information arrives as messages, calls, forms, and documents have more surface area here than businesses whose core work is physical or heavily specialist. That is why service businesses, professional practices, and anything with a busy inbound channel tend to score well, and it has nothing to do with the sector name.

What is your exception rate?

The characteristic people ignore and the one that most often kills a project.

A process that is uniform ninety-five percent of the time is a good candidate. One that requires judgment a third of the time is a bad one, because the design has to handle the judgment cases and that is a much larger build. Worse, the exceptions are usually where the money is.

Estimate honestly. Ask the person who does the job, not the person who manages it.

What happens if it gets one wrong?

The veto. If an error costs a customer relationship, breaches a regulation, or sends money to the wrong place, that process is a bad place to start regardless of how well it scores on the first three.

That does not mean never. It means not first, and it means human review in the design.

Where industry does genuinely matter

Three places, so this is not a dismissal.

Regulatory constraints. Healthcare, finance, legal, anything handling sensitive personal information. Not a barrier, but it changes what is permissible, where data can be processed, and how much review is required. It also means your consultant needs to know the constraints or be willing to learn them properly.

Seasonality. Businesses whose demand arrives in synchronized waves get more value from capacity that does not queue, because their problem is simultaneous load rather than daily volume.

How much of your value is the human relationship. In some businesses the contact is the product, and removing a person from it is a cost recorded as a saving. Know which of your interactions are in that category before anybody maps them.

The sector-specific pitch to be careful with

Firms selling an industry specialization sometimes have genuine depth and sometimes have a template with your sector's vocabulary substituted in.

The test: do they ask about your volumes, your exceptions, and your systems, or do they present what companies like yours typically do? The second is a template, and templates are why so many roadmaps read as generically true and produce nothing. How to spot it.

Depth in your industry is worth something. It is worth much less than curiosity about your particular operation.

One more thing worth separating: some of what you gain here is a durable advantage and most of it becomes ordinary within a year or two. Which is which changes where the money should go.

Score yourself

Take your three most frequent processes. For each: how often, is it mostly language, what share needs judgment, what happens if it is wrong.

Anything high on frequency, high on language, low on judgment, low on consequence is worth pursuing now, whatever business you are in. If nothing scores that way, your opportunities are elsewhere and honest advice would tell you so. Which problem class you are in.

Why we niched anyway

Worth explaining, since we sell one product to one trade and consulting to anyone.

We build agents for HVAC companies specifically because that trade has a distinctive combination: demand arrives in weather-driven waves, the work makes it physically impossible to answer a phone, and a routine inquiry is indistinguishable from a large one. That is a process argument, not an industry argument, and it happens to describe HVAC unusually well.

The consulting side is deliberately not restricted by sector, because the four characteristics above travel and the sector name does not. How that engagement works, and what we build for the one trade we specialize in.