Where Are Your People Still Typing?

Early in my career I sat through an interview that I wanted badly and did not get. The hiring manager asked me a strange question: on a scale of one to ten, how often do you touch the mouse when you are working in Excel? I said four or five. I had no idea it was a real question with a real answer.

I did not get the job. I have no proof the answer cost me, but it stuck with me. At the time I was working a junior accounting job while learning to automate spreadsheets in an intro systems class, and I made it a personal mission to almost never reach for the mouse again. I got good at it. In an analyst training program a couple years later, they ran a contest for who could work through a set of Excel instructions keyboard-only. I won. That habit, staying in flow instead of stopping to point and click, is one of the biggest reasons I have gotten a lot done fast over the years.

I have been thinking about that interview lately, because the same idea is back and it is much bigger than a keyboard shortcut. The question is no longer how often you touch the mouse. It is how often your skilled people stop working to type.

The multiplier most businesses are underusing

AI already does an enormous amount of work. If you write software, or run agents, or use the current crop of tools, you know the feeling of getting a week’s worth of output in an afternoon. That is the first multiplier.

The second one is voice. Tools like Wispr Flow now turn plain talking into clean, formatted text, good enough that I rarely think about punctuation or cleanup anymore. I just think out loud. When I am driving three or four coding agents at once, being able to speak instructions and feedback instead of typing them is not a small convenience. It stacks on top of what AI already does. A multiplier on a multiplier.

That part gets attention in software circles. What interests me more is where it has barely landed at all.

The reframe: stop asking where you can add AI. Ask where your best people are still typing when they could be talking.

The opportunity is in the trades, not the tech

Picture a technician at a busy auto repair shop. He has a backlog of cars, and as he finishes each one he is supposed to document what he did, what he didn’t, what the customer declined, and what to watch next visit. That note matters. It drives the invoice, the follow-up, the next appointment, and increasingly it feeds whatever AI the shop uses to personalize customer messages later.

But he is a technician, not a typist. So one of two things happens. He rushes a thin note between cars, or the shop’s software hands him canned text he clicks to insert. Canned text is fine for the immediate job. It is close to useless for anything downstream, because every car gets the same three sentences and none of them say what actually happened.

Now change one thing. While he is working, he talks. He describes the repair the way he would explain it to a coworker. The system takes that speech, checks it against the required format, strips anything that shouldn’t be in a customer-facing note, and produces a detailed, specific record. He never stopped working. The shop got documentation it could not have gotten before.

This is the pattern, and it is everywhere in service and field work: skilled people producing thin records of valuable work because the cost of writing it down properly, in the moment, is too high. Voice removes that cost.

A test for finding these spots in your own business

Not every typing task is worth converting. The ones that pay back share a few traits. When I am looking at how a business actually runs, these are the questions I ask about any repetitive writing task. The more that are true, the bigger the opportunity.

1. Is the person doing it skilled and expensive? A technician, a nurse, a field engineer, a project lead. Their hour is worth far more than data entry, and every minute spent typing is a minute not spent on the work only they can do.
2. Does it interrupt physical or focused work? Stopping mid-task to type carries a context-switch cost on top of the typing itself. Voice lets the record happen alongside the work instead of pausing it.
3. Does the output feed something downstream? Invoices, follow-ups, reporting, customer communication, an AI workflow. When the note is an input to something else, better notes compound into better everything after.
4. Is it currently skipped, rushed, or reduced to canned text? That is the tell. When a task keeps collapsing into the laziest possible version, it is not because people are careless. It is because the friction is too high to do it well. Remove the friction and the quality comes back.

If you run a business, walk one workflow with those four questions in hand. You will find a spot faster than you expect. It is usually hiding in plain sight, in the part of the job everyone quietly hates.

Do the math, honestly

Time savings are the easy part to see, so start there, with the clear caveat that the numbers below are illustration, not a promise. Your mileage depends on your people and your workflow.

Say voice-first capture saves one employee 20 minutes a day. That is about 1.7 hours a week. Call it 87 hours a year. Across a team of ten, you are near 870 hours returned annually, and it is likely more if your people are not fast typists to begin with.

But hours are the smaller half. The bigger return is the one that does not show up on a timesheet: the documentation you now have that you did not before. Richer repair notes mean smarter follow-up. Specific records mean you can reach back out to a customer with something real instead of a generic reminder. Clean, consistent data means the AI workflows you build next actually have something good to work with. Better notes today are better decisions, better retention, and more revenue later. That compounding is hard to put on a spreadsheet, which is exactly why most businesses miss it.

Where this does not apply

To be fair about it: voice is not always the answer. Highly structured data entry, the kind that is really a form with ten fixed fields, is often faster to tab through than to narrate. Anything requiring exact figures or codes read off a screen wants eyes and keys, not speech. Loud environments break transcription quality. And if nobody checks the output, voice just produces long, confident, wrong notes faster.

The sweet spot is narrative capture by a skilled person in the flow of their real work, where the record matters and typing it well is the thing that keeps not happening. That is a specific shape. Once you can see it, you see it in half the businesses you walk into.

The better question

Most companies right now are asking, “Where can we use AI?” It is the wrong question, or at least too broad to answer. It sends you shopping for tools instead of looking at your own work.

Here is the better one: where are your people still typing when they could be talking? Where is skilled, expensive time getting spent on writing things down, badly, because doing it well costs too much in the moment? That is where the quiet opportunities are. They are not glamorous. They rarely make the AI headlines. But they return time and improve the business in the same move, which is more than most of the flashy stuff can say.

Finding those spots, the places where good people are stuck on repetitive work that a system could carry, is most of what I do at Joint Lab Solutions. You do not need me to start. You need one walk through one workflow, and the four questions above.