My Goal for AI Is to Do Less of What I Dislike

That is the whole ambition. Not disruption, not replacing anyone, not ten times output. Just a redistribution of my own week toward the parts of this work I actually want to be doing.

What I do not like

Reformatting a document because the client uses different headings. Writing the fourth version of an email I have already written three times. Pulling numbers out of one system to paste into another. Summarizing a call I was present for. Turning notes into something presentable. Writing the boilerplate section of a proposal that is identical every time.

None of that is difficult. All of it takes real hours, and none of those hours produced anything a client would notice or pay extra for.

What I do like

Sitting across from somebody who has just started a business and working out what is actually going wrong. Finding the one decision that unlocks four others. Explaining something technical to somebody who was told it was too complicated for them. Building the thing once it is clear what the thing should be.

Those are the parts I would do for free, and they are also the parts clients get the most from. That alignment is not a coincidence. I am better at the work I like, because I have spent more attention on it.

Where the framing needs a filter

Here is where I have to be honest, because "do the things I dislike" is a rough proxy and it fails in one specific way.

Some of what I dislike, I dislike because it is uncomfortable rather than tedious. Following up on an unpaid invoice. Telling a client the thing they want is a bad idea. Deciding what to charge. Sitting with a problem I do not immediately know how to solve.

None of that should be offloaded, and the temptation to offload it is strong precisely because it is unpleasant.

So the test is not whether I enjoy it. The test is whether doing it personally produces something only I can produce.

• Tedious and nobody could tell who did it, offload it entirely
• Tedious but the output needs my judgment, let the tool draft and I edit
• Uncomfortable but consequential, do it myself, and do it sooner rather than later
• Enjoyable and valuable, protect the time aggressively

The second category is the largest and the one people get wrong most often. A first draft is not the work. Deciding whether the draft is right is the work.

What this looks like in a week

Concretely, because the principle is worthless without it.

I keep a document of my own material. How I describe what I do, my standard proposal sections, the tone I write client emails in, the objections I hear most often and how I answer them. When I need something written, that document goes in with the request. What comes back sounds like me because it was built from me.

Meeting notes become a summary and an action list in about a minute rather than twenty. I read the summary because the summary is where I catch the thing I missed, but I am not the one typing it.

Recurring emails are templates now. Not generic templates. Templates built from the four best versions I already sent.

The research I used to put off because it would take an afternoon, I now do, because it takes twenty minutes. That one is the surprise. The gain was not the time saved on things I was already doing. It was the things I had been quietly avoiding because the cost was too high.

What I do not hand over

I write the thing you are reading. I decide what a client should do. I have the difficult conversation. I check anything with a number in it before it leaves my hands.

Not out of principle. Because those are the parts where being wrong is expensive, and because a tool that has never met my client has no view worth having on my client.

There is also a quieter reason. If I stopped doing the thinking parts, I would get worse at them. Skill is maintained by use. The people who will struggle in a few years are not the ones who refused to use these tools. They are the ones who used them for the parts that were keeping them sharp.

Why this is the right goal for a small business

If you have one or two people, you do not have a productivity problem in the abstract. You have a specific problem, which is that the person who should be doing the valuable work is instead doing the administrative work, because there is nobody else to do it.

Every hour reclaimed from that is an hour returned to the thing you are actually good at.

That is a smaller ambition than most of what gets written about this technology, and I think it is the correct one. I am not trying to build a business that runs itself. I am trying to spend more of my week on the reason I started this in the first place.

Fifteen years in, the work I like has not changed. There is just less standing between me and it.