Start With the Work, Not the Tool

The Question Everyone Wants Answered

One of the first questions business owners ask me about AI is: “Which AI tool should I use?” It sounds like the right question. But usually, it is too early to answer. Before I recommend a tool, I need to understand the work. What are you trying to accomplish? What is taking too much time? Where are customers getting confused?

What keeps being forgotten? What work depends entirely on one person? What information is scattered across emails, documents, notebooks, and someone’s memory? What are you doing repeatedly that could become a process? Until we understand those things, choosing a tool is mostly guessing. And guessing is not a good reason to spend money.

We’ve Been Trained to Shop for Answers

When a problem appears in business, our first instinct is often to look for something to buy. A new website. A new customer relationship system. A new scheduling program. A new marketing platform. A new AI subscription. Sometimes the purchase is helpful. But buying a tool and improving the business are not the same thing. I have seen businesses own software they barely use.

The program may be capable of doing wonderful things, but no one clearly decided what job it was supposed to perform. Eventually, the software gets blamed. People say: “It didn’t work.” Sometimes that is true. But sometimes the tool never had a fair chance because the work had not been defined.

A Tool Needs a Job

Think about walking into a hardware store and asking: “What is the best tool?” The person helping you would probably ask: “What are you trying to do?” Are you hanging a picture? Building a deck? Fixing a pipe? Repairing an engine? There is no best tool without a job. AI works the same way. If you want help writing an email, that is one kind of job.

If you want to organize customer information, that is another. If you want to answer common questions, follow up with prospects, analyze sales calls, create a social media calendar, or turn a conversation into a training guide, each requires a different kind of process. The tool should be chosen for the work. The work should not be invented to justify the tool.

Begin With What Already Happens

You do not need to imagine some enormous AI transformation. Look at the work your business already does. A customer calls. Someone answers the phone. Questions are asked. Information is written down. An estimate is prepared. A follow-up message is sent. The work is scheduled. The customer is served. Payment is collected. A review may—or may not—be requested.

That entire journey contains work. Some of it requires human judgment and personal attention. Some of it is repetitive. Some of it is frequently forgotten. Some of it depends on information that already exists somewhere else. Those are the places worth examining. The first question is not: “Where can I insert AI?” The first question is: “What is happening here now?”

Watch for Repeated Work

One of the easiest places to begin is with work you repeat. Maybe you answer the same five questions every week. Maybe every new customer needs the same instructions. Maybe you write similar follow-up emails after every appointment. Maybe your staff searches through old documents for information that should be easy to find. Maybe you turn every recorded conversation into notes, action items, and a proposal.

Repeated work often reveals the beginning of a system. AI may help make that system faster or more consistent. But first, you need to notice the repetition. Business owners become so accustomed to doing certain tasks that they stop seeing them as processes. They simply think: “That is part of my day.” Sometimes the first benefit of exploring AI is not automation.

It is finally seeing the work clearly.

Don’t Automate Confusion

Imagine that customer information is collected differently every time someone calls. One employee writes it on paper. Another sends themselves an email. Someone else enters part of it into a spreadsheet. Important details live inside text messages. Follow-up depends on who remembers. A business owner might say: “We need AI to fix our follow-up.” Perhaps AI can help.

But the first problem is not a lack of artificial intelligence. The first problem is that there is no agreed-upon process. If we add automation before understanding the work, we may simply move confusion faster. We may send the wrong information more efficiently. We may create more messages without creating better communication. AI works best when it supports a process people understand.

Find the Friction

When I look at a business, I pay attention to friction. Friction is where something feels harder than it should. A customer has to repeat information. An employee has to enter the same details twice. A lead waits three days for an answer. The owner cannot find the latest version of a document. A task depends entirely on someone remembering to do it.

Good people spend hours completing routine work while important decisions wait. Those moments are telling us something. They show us where the work may need to be simplified, documented, organized, or supported. AI may become part of the answer. But even before we use it, we have learned something important about the business. We have found a place where the system is creating unnecessary effort.

Begin With One Small Job

Business owners sometimes believe they need a complete AI strategy before they can begin. I don’t think that is necessary. Start with one useful job. Not the entire company. Not every department. One job. Perhaps AI helps turn a recorded customer conversation into a clear summary. Perhaps it helps draft a follow-up email based on that conversation. Perhaps it organizes the most common customer questions.

Perhaps it turns one video into several social media posts. Perhaps it compares this month’s customer feedback with last month’s. A small, clearly defined job teaches you more than a large, vague ambition. You can observe the result. Correct mistakes. Improve the instructions. Decide whether it actually saves time or produces better work. Then you can build from what you learned.

Keep People Close to the Work

The people doing the work often understand the problems better than the person buying the software. They know which questions customers ask. They know which information is usually missing. They know where delays occur. They know which steps look simple from the outside but require experience and judgment. Before introducing AI into someone’s job, talk with them. Ask what slows them down.

Ask what they repeat. Ask what they wish they had more time to do. Ask which parts of their work require a human relationship. AI should support good people—not be dropped on top of them without explanation. When employees understand the purpose, they can help shape a much better system.

What Are We Really Trying to Improve?

Not every improvement should be measured only by speed. Faster is useful. But sometimes the real goal is: Fewer missed details. More consistent follow-up. Clearer communication. Better organization. More time with customers. Less dependence on one person’s memory. Greater confidence in decisions. AI can help a business become more efficient. But efficiency should serve something meaningful. Saving ten minutes matters most when we know what we want to do with those ten minutes.

The goal is not simply to make people work faster. The goal is to help them do better work.

The Lesson

Do not begin your AI journey by shopping for tools. Begin by looking carefully at the work your business already performs. Find what is repeated. Find what is confusing. Find what is being forgotten. Find where customers wait. Find where valuable knowledge is trapped inside one person’s head. Then choose one small, useful job to improve. Once the work is understood, the right tool becomes much easier to recognize.

Reflection Questions

Think about an ordinary week inside your business:

• What work do I or my employees repeat most often?
• Where do customers have to wait, repeat themselves, or ask for clarification?
• What important tasks depend on someone remembering to do them?
• Where is information scattered or difficult to find?
• What work requires human judgment, empathy, or personal responsibility?
• What is one small job AI might help us perform more consistently?
• What result matters most: saving time, reducing mistakes, improving communication, or serving customers better?

You do not need to transform your entire business. Begin by understanding one piece of work.