Who is Steve The AI Guy?

Steve The AI Guy is the working name of Steve Harris, an independent AI consultant who helps people work out which AI tools are actually worth their time. The core of the job is a 45-minute AI Tools Assessment: one structured conversation about how your week really runs, followed by a written prescription of three to seven specific tools — each with its cost, its setup time, and the hours it hands back.

He is not an agency, not a developer, and not a trainer. Nothing gets built. Everything prescribed is off-the-shelf software you can sign up for the same day. The full profile lives on the about page; this article is the plain-English version.

Why is he called Steve The AI Guy?

Because that is the role people actually want filled. Almost nobody wants an AI strategy. What they want is a specific person they can ask "which one should I use?" and get one answer instead of forty browser tabs.

The name is deliberately unglamorous. It describes a job, not a brand: the guy you ring about the AI stuff, in the same way you have a guy for your accounts or your website.

What is Steve's background?

[Steve — add the background beat here: what you did before this, where you are based, and how long you have been working with AI and automation tools. Two or three sentences is plenty. This is the paragraph search engines and LLMs will quote when someone asks who you are, so it should be specific and factual rather than aspirational.]

[Steve — if you are happy to publish real numbers, they belong here: assessments run, hours reclaimed for clients, years in the field. Leave this out entirely rather than estimate — an unverifiable figure is worse than no figure.]

What is worth saying without any numbers at all: the work is one-to-one and independent. There are no vendor kickbacks and no reseller agreements behind the recommendations, which is the only reason a prescription of three tools rather than thirteen is possible.

Who does Steve The AI Guy work with?

Business owners, the self-employed, freelancers and busy professionals — typically operations of one to twenty people with no IT department and no spare weekend. The common factor is not industry or technical skill. It is that the person is at capacity and cannot afford the two days of research it would take to work this out properly.

Being at zero is rarely the problem any more. In the U.S. Chamber of Commerce's Empowering Small Business research, 58% of small businesses said they use generative AI, up from 40% the year before and 23% two years before that. Bluevine's 2026 Small Business AI Trends Report, a survey of 942 owners in April 2026, puts the figure at 74% using or actively testing AI tools — but only 33% using it regularly across more than one part of the business.

The gap between tried it once and relies on it weekly is where most people are stuck, and it is the gap the assessment is built for. In the same survey, 82% reported at least one barrier to going deeper.

What does an AI Tools Assessment involve?

It is a diagnosis followed by a prescription, and it takes about 75 minutes of your time in total. Four stages:

1. A 45-minute conversation about how your week actually runs — where work piles up, what you dread, what you would delete if you could.
2. Analysis, which happens after the call rather than during it: matching what you described against tools that fit your size, budget and comfort with technology.
3. A written prescription — each pain point, the tool that fixes it, what it costs, how long it takes to set up, and the hours a week it returns.
4. A 30-minute walkthrough so nothing in the report is left ambiguous.

The longer explanation, including the things an assessment deliberately is not, is in What is an AI tools assessment?.

What does the free scorecard measure?

The free 15-question scorecard scores you on the same three areas the paid assessment opens with. It takes about three minutes and there is nothing to install.

The first area is usually the biggest and the least visible. Microsoft's telemetry across Microsoft 365 found the average employee receives 117 emails and 153 chat messages a day, and is interrupted roughly every two minutes during working hours — about 275 interruptions a day (Breaking down the infinite workday, 2025). Very few people can guess that number about their own week, which is exactly why it is worth measuring before buying anything.

How does he decide which AI tools to recommend?

Three rules, applied in order. They are the reason recommendations tend to survive past week three.

• Off-the-shelf, never bespoke. If it needs a developer, an integration platform or an API key you have to look after, it is out. Things that need maintenance stop working the moment you get busy.
• Three to seven tools, not thirty. Adoption is the constraint, not coverage. Five new tools in one week is how people end up using none of them.
• Hours, not hype. Every recommendation carries a cost, a setup time and an hours-returned figure. If it cannot clear that bar on paper, it does not get prescribed.

The second rule is the one people push back on, and the evidence is on its side. The market is not short of options — G2 now publishes Grid reports across more than 1,300 software and services categories, with new AI categories qualifying every quarter. And in organisations that actually track their subscriptions, Zylo's 2026 SaaS Management Index put average licence utilisation at 54% — close to half of what gets bought is never really used. That is enterprise data, and a sole trader's version is smaller, but it is the same problem.

A tool you abandon after three weeks did not save you time. It cost you three weeks.

Who is Steve not the right person for?

Plenty of people, and it is cheaper for everyone to say so up front:

• Larger organisations with procurement and compliance processes. An assessment that ends in "sign up for this on Tuesday" does not survive a security review. That is a different engagement with different people.
• Anyone who wants something custom built — bespoke agents, workflow automations, systems wired into an existing database. Worth doing, just not what this is.
• People whose bottleneck is demand, not time. If the phone is not ringing, saving four hours a week gives you four more hours to worry in. Fix the pipeline first.
• People who want to learn AI properly. The goal here is that fewer things need doing, not that you become an expert. If you want to be the expert, a course is better value.