How I Use AI to Save 10 Hours a Week

I got skeptical of the "AI saves hours" headlines pretty fast. They always lacked specifics. Which hours? Doing what? Show me the workflow.

So I'll show you mine.

I run a consulting business. I am also always testing AI tools because that's the work. Over the past year I've settled into a set of habits that, if I'm being honest about it, probably save me eight to twelve hours a week. Not all at once. A few minutes here, an hour there. It adds up.

Here's what actually works.

Drafting first passes on anything written

I used to stare at a blank email for five minutes before I could start. Now I open Claude and say: "Draft a follow-up to a prospect who went quiet after our second call. Friendly, not pushy, keep it under 100 words."

I get something back in three seconds. It's usually 70% right. I rewrite a few sentences and send it.

The blank page problem is gone. That's the whole trick. I'm not outsourcing my writing. I'm skipping the warmup.

This works for: emails, proposals, meeting agendas, LinkedIn comments, responses to complaints, onboarding instructions. Anything where you know what you want to say but getting started costs you time.

Conservative estimate: one hour saved per week.

Summarizing things I don't have time to read

Articles. Meeting transcripts. Long email chains. PDF reports someone sent me.

I paste them in and ask for the three things I actually need to know. Sometimes I ask "what would I need to act on this?" instead of "summarize this" because I want decisions, not a recap.

This sounds small. It's not. I probably encounter fifteen to twenty pieces of content per day that I used to skim badly or skip entirely. Now I can actually process them.

One hour saved per day, maybe. That's five hours per week right there.

Research that used to take an hour

Before a meeting with a new client, I used to spend thirty to forty minutes reading their website, LinkedIn, recent news. Now I give Claude their company name and industry and ask it to give me the ten things I should know walking into a discovery call.

Then I spend ten minutes reading the output and checking anything that seems off.

It's not perfect. I still verify. But the starting point is so much better that I get to the useful part of my prep in a fraction of the time.

Two hours saved per week if you have regular new prospect meetings.

Building automations I used to pay for or skip

This one took more upfront work, but the return is ongoing.

I have a background process that monitors court dockets for my brother's law firm. It used to take three days a week of manual checking. Now it runs every night in two minutes and sends an alert if anything changed.

I built that with AI help. I'm not a developer. I described what I wanted in plain English, worked through the logic conversationally with Claude, and had a working script in an afternoon.

I'm not saying everyone will do this. But if you have a repetitive process that follows clear rules, it's worth asking "could this run automatically?" The answer is yes more often than it used to be, and you don't need an IT team anymore.

The prep work before I think

Before I make any significant decision now, I use AI as a sounding board. I describe the situation. I ask it to steelman the opposite position. I ask what I might be missing.

It doesn't make the decision. I do. But it catches obvious blind spots and forces me to articulate the problem clearly before I've committed to a direction.

This is harder to quantify in time saved. But bad decisions cost more time than almost anything else. Even if this catches one bad call per month, it's worth it.

What I don't use it for

I don't use AI to replace judgment calls. I don't use it for anything that requires actual relationship knowledge I'd have to teach it from scratch. I don't use it for legal advice.

And I don't use it for things I can do in under two minutes that don't repeat. The overhead of prompting isn't worth it.

The rule is: if I find myself doing the same thing more than three times, I ask if AI can handle the first draft or the heavy lifting.

Start with one thing

If you're not using AI at all right now, pick the thing on this list that sounds most like something you already do. Just that one thing. Try it for a week and see how it feels.

Most people who tell me "AI doesn't save me time" are either using it for things it's bad at, or they're using it wrong. Wrong usually means they're treating it like a search engine instead of a collaborator.

The shift is small. The time savings are real.

If you want to figure out which of your workflows are the best candidates for AI, that's exactly the kind of thing I help small business owners map out. Reach out here.