Stop Chasing Tools. Start With the Problem.

He had 11 AI tools bookmarked.

Eleven. I counted them when he showed me his browser one afternoon. Some he had tried once and abandoned. Some he had never opened at all. One he was paying $29 a month for and could not quite remember why he signed up.

He was actively using two of them. Neither was solving the problem he actually had.

His name is Derek. He runs a small landscaping company in Mississauga — three crews, seasonal work, constantly behind on client communication and invoicing. He had heard about AI everywhere and decided he needed to get on board. So he signed up for things. A lot of things.

Six months later, nothing had changed. He was still writing estimates by hand, still chasing overdue invoices, still losing three hours a week to admin he hated. Just with a longer list of subscriptions he felt guilty about not using.

When I asked him what problem he was trying to solve, he went quiet for a moment.

“I haven’t really thought about it that way,” he said.

That sentence is the whole issue.

How Most People Adopt AI

Most people hear about a tool. They sign up. They spend a week trying to figure out what it does. Then they either make it work for something it was not really designed for, or they give up and move on to the next one they heard about.

The tool comes first. The problem comes later. Sometimes the problem never comes at all.

This is how you end up with 11 bookmarks and two active tools. This is how you end up paying for things that sit unused. This is how you end up convinced that AI is not really for your business, when the truth is you just never asked it to solve anything specific.

There is nothing wrong with being curious about new tools. The problem is treating curiosity as a strategy.

The businesses actually saving meaningful time with AI right now did not start with a tool. They started with a task they hated doing. They found the thing that was draining their week, costing them money, or keeping them up at night. Then they looked for a tool that could help with exactly that one thing.

That order matters more than anything else about AI adoption.

Why Tool-First Thinking Fails

When you start with a tool, you spend your energy learning how it works instead of solving a problem. You adapt your workflow to fit the tool rather than finding a tool that fits your workflow. And when it does not quite work, you assume the tool is the problem and move on to the next one.

This is why so many people try AI, feel underwhelmed, and conclude it is overhyped. They were not underwhelmed by AI. They were underwhelmed by a tool that was designed to solve a problem they had not clearly identified.

Derek had tried ChatGPT once, about eight months before we talked. He had opened it, typed “help me with my business,” got a generic response, and closed the tab. He told me it felt like it was made for someone else.

He was right. It was made for someone with a specific question. He had not brought one.

The Three Questions That Find Your Problem

Before you sign up for anything new, and before you revisit anything you have already given up on, answer these three questions.

What task do you dread most that comes up every week? Not something vague like “my marketing” or “staying organised.” Something specific. The Friday afternoon email you do not want to write. The monthly report that takes four hours and could probably take one. The quote you put off because writing it from scratch every time takes too long. The social media caption you stare at for twenty minutes before giving up and posting nothing.

How much time is this actually costing you? Do the math. If it takes you two hours a week, that is eight hours a month. Roughly 100 hours a year. What would you do with 100 hours back?

What would it cost you if this task just stopped happening? If the honest answer is nothing, it should probably stop happening entirely. If the answer is lost clients, missed revenue, or genuine stress, you have found something worth solving.

When you have clear answers to all three, you have your problem. Then you go looking for the tool.

What Derek Found

I walked Derek through the three questions. His answers came fast once he was thinking in terms of problems instead of products.

The task he dreaded most was writing estimates. Every new job required a custom quote and it took him anywhere from 45 minutes to two hours depending on the scope. He was doing eight to twelve of these a week during busy season. That was up to 24 hours a month spent writing estimates — time that could have been spent on the work itself, on sales calls, on anything else.

That was his problem. Specific. Recurring. Costly.

We did not go looking for a new tool. He opened ChatGPT, the one he had tried once and abandoned, and we built one prompt together. It took twenty minutes to get right.

The prompt asked for a professional landscaping estimate based on the job type, property size, scope of work, and any special considerations. It produced a clean, client-ready draft in under two minutes. Derek adjusted a few numbers, added his company header, and sent it.

He ran it the next morning on a real job. Then again the day after.

Within two weeks it was part of his daily routine. He was spending about six minutes per estimate instead of 90. He got back roughly 18 hours a month. With something he already had access to.

He did not need another subscription. He needed a clearer question.

The Problem-First Framework in Practice

Once you have identified your problem, the framework for finding the right tool is straightforward.

• Start with what you already have. Before you sign up for anything, check the tools you are already paying for. Most people are sitting on AI capabilities inside ChatGPT, Google Gemini, Microsoft Copilot, Adobe Firefly, Canva, Notion, and their email platform that they have never turned on. Derek’s solution was already in his existing toolset. He just had not asked it the right question.

• Test before you commit. Most AI tools have free tiers or trials. Use them for your specific problem before you pay. If the free tier solves your problem, you may not need the paid tier at all.

• Give it four weeks. The first attempt with any prompt rarely produces the best result. Build it, test it on real work, refine it based on what comes back, and save the version that works best. After four weeks of using the same prompt for the same task, you will know whether it is genuinely saving you time or just adding a step.

• Add one thing at a time. Once the first solution is running on autopilot, look at the next problem on your list. Not five problems. One. The businesses that build real AI systems do it one task at a time over several months, not in one frantic week of signing up for everything.

The AI Graveyard in Most People’s Accounts

One of the most useful exercises I do with business owners is what I call the subscription audit. You go through every active subscription and ask one question about each: is this solving a specific problem, or did I sign up because I thought I should have it?

The results are usually uncomfortable.

Most people have at least two or three AI-related subscriptions that fall into the second category. Tools that seemed essential when someone mentioned them at a conference. Tools that had a compelling demo video. Tools that solved a problem the owner had vaguely considered but never clearly defined.

Cancel those. Not because the tools are bad. Because you have not done the work to use them well, and paying for something you feel guilty about not using just adds to the mental load.

The goal is not a comprehensive AI stack. The goal is a small number of tools that each solve a specific, recurring problem so well that you would notice immediately if they stopped working.

Derek now has two. He had eleven and got nothing. He has two and saves 18 hours a month.

How to Know If a Tool Is Actually Working

After four weeks with any new AI system, ask yourself these questions.

• Do I use it without thinking about it, or do I still have to remind myself it exists? Real systems become automatic. If you are still consciously deciding to use it, it has not become a habit yet.

• Is the output quality good enough that I send it with minimal editing? If you are spending more time editing the AI output than you would have spent writing from scratch, the prompt needs work or the tool is wrong for this job.

• Can I tell someone exactly what problem this solves and how much time it saves? Vague answers like “it helps with content” mean the problem was never specific enough. Specific answers like “it drafts my client estimates and saves me 90 minutes a day” mean it is working.

If you cannot answer yes to all three after four weeks, go back to the three questions and clarify the problem before you try a different tool.

The AI Task Filter

Before you make any tool decision, the AI Task Filter helps you sort every task in your business into three categories.

Automate it means AI handles this fully. You review the output and send it. Writing estimates, drafting routine emails, generating social captions — anything repetitive and low-stakes where good enough is genuinely good enough.

Assist it means AI drafts, you refine. Client proposals, sensitive communications, anything where the relationship matters and a generic output would do more harm than good. AI gets you 70 percent of the way. You add the specific details that only you would know.

Keep it human means AI stays out. Pricing decisions, strategic calls, anything requiring judgment built from experience and relationship context. These are not candidates for automation regardless of how capable the tools become.

That clarity is worth more than any specific tool recommendation. Once you know which category a task belongs in, the tool decision becomes obvious.

The AI Task Filter is free. Grab it before you do anything else.

AI Task Filter (Free)

What to Do This Week

• Pick one task. The one that came to mind when you read the three questions. Not the biggest, most complicated thing you want to fix. The one recurring task that drains your energy and could probably be handled with a well-built prompt.

• Run it through the three questions. What is it, how often does it happen, what does it cost you.

• Then build one prompt for it. Use ChatGPT, Gemini, or whatever tool you already have access to. Test it on real work this week. Refine it once. Save the version that works.

Do not sign up for anything new. Not yet.

Find the problem first. The tool is the easy part.

Resources

Drop FILTER in the comments on LinkedIn and I will DM you the one-page AI Task Filter.

• AI Task Filter (Free) — sort every task in your business before you spend a dollar or an hour on any tool.
• AI Growth Advisor (Free) — talk through your specific situation before you make any decisions.
• SME AI Playbook — 30 plus prompts built for specific small business use cases so you stop guessing what to ask.
• Digital Equity Hub — hands-on help for small business owners who want practical support without enterprise pricing.

Book a Discovery Call — 30 minutes, no pitch, just a real conversation about what would actually help your business.

Wilton Otto is the founder of WilGlobo Digital Solutions, helping Canadian small business owners, solopreneurs, coaches, and newcomer entrepreneurs use AI practically without the hype. He publishes Digital Equity Notes every Monday.

Subscribe: Digital Equity Notes on LinkedIn | 📩 Join WilGlobo Signal and stay ahead of what’s next.