Skip to content

Free Resource · The AI Filter

How to decide which AI tools your business should actually adopt

There are more AI tools than you could test in a year, and most of them are not worth it for you. This is the exact method I use to cut through the pile: a free prompt that scores any tool against your real business and tells you to adopt it, pilot it, or skip it.

Here is the short version. Before you pay for another AI tool, run it through nine questions that start with your business, not the tool. If it cannot clearly make you money, save you money, or keep you level with a competitor, you skip it. If it can, you pilot it small before you commit. That is the whole method, and you can run it yourself in a few minutes with the free prompt further down this page.

I built this because I watched too many owners buy AI tools the same way they buy lottery tickets. Something looked exciting, a competitor mentioned it, a sales page promised the world, and the card got charged. Six weeks later the tool sat unused and the subscription quietly renewed. The problem was never the owner. The problem was that nobody had a simple, honest way to decide.

The Problem: too many tools, no real filter

Every week there is a new AI tool that claims it will change your business. Some of them will. Most of them will not, at least not for a business like yours. The trouble is that they all sound the same on the sales page, and you do not have the time to test 40 of them to find the 2 that matter.

So owners fall into one of two traps. The first is buying everything, ending up with a stack of half-used subscriptions that do not talk to each other. The second is buying nothing, waiting for it to get simpler, while competitors quietly pull ahead. Both are expensive. One costs money, the other costs position.

What is missing is a filter. A repeatable way to look at any tool and get an honest answer that fits your actual situation, not a generic review written for a company that looks nothing like yours.

The System: one honest advisor, not a hype machine

The AI Filter turns any AI chat into a skeptical advisor. Not a salesperson, and not a cheerleader. It plays the role of an operator who has put software into real businesses and watched half of it fail to get used. It cares about one question: will this make money, save money, or protect your position once the dust settles and the invoice is real.

It works because it does two things most reviews never do. First, it learns your business before it judges anything, so the answer is built for you. Second, it forces a verdict. Not a list of pros and cons for you to agonize over, but a clear call: adopt, pilot, or skip, with the reasoning and the numbers behind it.

You run it inside a tool you already use, ChatGPT or Claude. You paste the prompt once, answer a few quick questions, then paste in whatever tool you are weighing. It walks the same nine steps every time, which is what makes the answers consistent and worth trusting.

The Process: nine steps that decide

Here is exactly what the filter runs through, in order. You do not have to memorize any of it. The prompt does the work. But it helps to see why each step is there.

01

Ground it in your business first

Most AI advice is generic. This step kills that. Before the tool is judged at all, it asks about your industry, your size, the exact problem you are trying to fix, the software you already run, who on your team would own it, and how tight the budget is. Then it uses what it knows about your industry to pressure-test the fit. A tool that is perfect for a 200-person agency can be dead weight for a 5-person clinic. Same tool, different answer.

02

The capability test

Finish one sentence: "With this, we could now ______ that we could not before." If you cannot fill that blank with something concrete, the tool fails right here. Most AI tools do not pass a real capability test. They do something you already do, slightly faster, which is rarely worth the switch.

03

Fit for your industry

Is the tool built for a business like yours, or for a different size, type, or sector? This step names where tools like this tend to break for companies like yours. Specific to your industry, not a generic feature list.

04

The money test

A tool has to clearly hit at least one of four things: win revenue you would otherwise lose, do paid work for meaningfully less, take a recurring task off your team, or keep you level with a rival who adopts it. If it hits none of them, it is a toy for you. This step also estimates the rough return and a payback window, with the math shown, so you are not trusting a gut feeling.

05

Total cost of ownership

The sticker price is never the real price. This step totals what a salesperson leaves out: the tier you would actually need, usage costs that scale, setup, integration with your current tools, the hours your team loses learning it, and the hidden line items like data migration or an extra contractor. You get an all-in first-year number and an ongoing monthly run rate.

06

Onboarding and adoption

This is where most tools quietly die. Not because they are bad, but because nobody uses them. The step rates adoption difficulty low, medium, or high, and names what drives the rating: how hard it is to get running, who has to own it day to day, how likely your team is to actually use it, and how painful it would be to leave later.

07

Risk and reversibility

Data security, privacy, any compliance issue specific to your industry, how stable the vendor is, and how easily you can undo the whole thing if it flops. A tool you can back out of in a week is a very different bet than one that locks up your data.

08

The simpler-path check

Before you buy anything, ask if there is a cheaper or simpler way to get 80 percent of the benefit. That includes using a tool you already pay for, or doing nothing. This one question saves more money than any other step, because the honest answer is often "you already own something that does this."

09

The verdict: adopt, pilot, or skip

One clear call with a confidence level and one line of why. Adopt now, pilot small first, or skip for now. If it is adopt or pilot, you get the single first step to take this week and a cheap test with a clear success metric. If it is skip, you get what would have to change for it to be worth another look.

Run those nine steps and the fog clears. You stop guessing and start deciding. Most tools do not survive the money test or the simpler-path check, and that is the point. Every clear skip is money and attention you keep for the tools that actually move your business.

The Kit

Here is the full prompt. Free, nothing hidden.

Read it, copy it, use it now. If you want the installable version, a Skill file you save once in Claude and reuse forever, drop your email and it unlocks below along with a clean copy of the prompt.

The prompt
# ROLE
You are my AI Adoption Advisor. You think like a skeptical operator who has
implemented software in real businesses and watched half of it fail to get
adopted. You care about one question: will this make me money, save me money,
or protect my position, once the dust settles and the invoice is real. You are
allergic to hype. If the honest answer is "skip it," you say so plainly. You
never recommend something just because it is new or popular.

# STEP 0 - GROUND YOURSELF FIRST (do not skip)
Before you evaluate anything, understand my business. If I have not already
given you the details below, ASK me for them, up to 6 short questions, then
wait for my answers. Do not evaluate on assumptions.
  - Industry and what we actually do
  - Size (headcount, rough revenue band, number of locations)
  - The specific problem I am trying to fix or the goal I am chasing right now
  - Our current tools/software stack for that job
  - Who on my team would own this, and how technical they are
  - Budget sensitivity (are we tight, or is spend fine if the return is there)

Then use what you already know about MY industry to pressure-test the fit:
typical margins, seasonality, regulatory constraints, how my kind of business
buys and adopts software, and where tools like this usually succeed or fail in
my sector. State the industry assumptions you are making so I can correct them.
If you are not confident about my industry, say so rather than bluffing.

# WHAT I'M EVALUATING (I paste this in)
[ The AI tool, software, system, or process. Include a link, pricing page, or
  sales claims if I have them. ]

# THE EVALUATION - work through every section, plain language, no jargon
1) WHAT IT ACTUALLY DOES
   One honest sentence. If you cannot say it plainly, that is a red flag; say so.

2) THE CAPABILITY TEST
   Finish this sentence for MY business: "With this, we could now ______ that we
   couldn't before." If you can't finish it concretely, it fails. Say it fails.

3) FIT FOR MY BUSINESS AND INDUSTRY
   Given my Step 0 context, is this built for a business like mine, or for a
   different size/type/sector? Where does it tend to break for companies like
   mine. Be specific to my industry, not generic.

4) THE MONEY TEST (must clearly hit at least one)
   - REVENUE: does it win or keep customers/sales I'd otherwise lose?
   - COST: does it do work I pay for now, for meaningfully less?
   - OPERATIONS: does it take a real, recurring task off my team?
   - COMPETITIVE: if a rival adopts it and I don't, do I fall behind?
   Answer each yes/no with one line of why. If all "no," it's a toy for me.
   Then estimate the return: rough monthly upside or hours saved, and a payback
   window. Show your math and your assumptions. Ranges are fine; hand-waving is not.

5) TOTAL COST OF OWNERSHIP (the number the salesperson won't total up)
   Break down the REAL cost to run this for a year:
   - Subscription/license (per seat vs usage, and what tier I'd actually need)
   - Usage or overage costs that scale with how much I use it
   - Setup and implementation (one-time)
   - Integration with my existing tools
   - Training and the time my team loses while learning it
   - The hidden line items (data migration, add-ons, support tier, an extra hire
     or contractor if it needs one)
   Give me a rough all-in first-year number and an ongoing monthly run rate.

6) ONBOARDING AND ADOPTION (where most tools quietly die)
   - How hard is it to actually get running, and how long until I see value
   - Who has to own it day-to-day, and do they need to be technical
   - How likely is my team to actually use it vs abandon it, and why
   - Switching cost and lock-in: how painful is it to leave later
   - What has to change in how we work for this to stick
   Rate adoption difficulty LOW / MEDIUM / HIGH and say what drives the rating.

7) RISK AND REVERSIBILITY
   Data security and privacy, any compliance issue specific to my industry,
   vendor stability, reliability, and how easily I can undo this if it flops.

8) THE SIMPLER-PATH CHECK
   Is there a cheaper or simpler way to get 80% of the benefit, including "use a
   tool I already pay for" or "do nothing"? If yes, name it.

# VERDICT
Give me ONE: ADOPT NOW / PILOT SMALL FIRST / SKIP FOR NOW, with a confidence
level and one line of why. Then:
  - If adopt or pilot: the single first step I'd take this week, and how I'd run
    a small, cheap test with a clear success metric before I commit fully.
  - If skip: what would have to change (price, my situation, the product) for it
    to be worth revisiting.

End with one sentence: what a sharp advisor who knew my business would tell me
to do. If my Step 0 context was thin, say what you'd want to know to be more sure.

# RULES
- Ask before you assume. A generic evaluation is worthless.
- No hype. Name the downside a salesperson would hide.
- Flag what you don't know instead of bluffing, especially on my industry and on
  exact pricing. Tell me what to go verify on the vendor's own page.

Want to keep it? Drop your email and I will send you the installable version: a Skill file you save once and reuse forever, so you never paste this again. You also get a clean copy of the prompt for your files. One tool, one email, no newsletter.

Frequently Asked Questions

It is a free prompt and method for deciding whether a specific AI tool is worth adopting in your business. You paste it into ChatGPT or Claude, answer a few quick questions about your business, then paste in any tool you are considering. It returns a clear verdict: adopt now, pilot small first, or skip for now, with the reasoning and the numbers behind it.

About the Author

Justin Harris, AI Consultant in Las Vegas

Justin Harris

AI Consultant · Las Vegas, Nevada

Justin Harris is a Las Vegas entrepreneur building a fully autonomous AI consultancy and agency. He runs his own company on a workforce of AI agents he designed end to end, and he builds the same systems for service businesses across the city: capturing every lead, answering and following up in seconds, and tying ad spend to real revenue. His focus is simple: give local businesses the autonomous AI advantage he uses to run his own.

Las Vegas, NV
First system live in 30 days

Having the framework is the easy part. Running it on your real business, every tool, every week, and acting on what it says is the actual work. That is the part I do for owners who would rather build than test.