Back in January, I sat down to do my normal Monday work. I had to send invoices to three clients, edit a podcast for a client, write two blog posts, and build a quick landing page for a friend’s small business. By lunch, I had finished none of it. I remember looking at my screen and thinking, “I run a one-person business, I pay for twelve different apps, and I still feel buried.”
That’s the day I stopped using AI tools just for fun and started building my real work around them. Six months later, my whole toolkit looks different. I dropped Grammarly, cancelled a video editing app, stopped paying for a virtual assistant service, and let a design software plan expire because I just wasn’t using it anymore.
I won’t pretend this switch was perfect. I once messed up a client project because I trusted an AI result without checking it first (more on that later). But overall, this change saved me around $300 a month in software costs, and it gave me back real hours in my week.
Here’s what I switched, what worked, what didn’t, and where I made mistakes so you don’t have to make the same ones.
Why I decided to switch
I’m a freelance writer and part-time consultant. My work covers writing, basic design, scheduling, talking to clients, and simple bookkeeping. None of that should need a full set of business software, but somehow I had ended up with one anyway.
The real problem wasn’t even the cost. It was switching between apps all day. Every time I opened a new tool, I lost five or ten minutes just remembering how it worked. AI tools that live right where I already work, like my browser, my documents, and my chat apps, cut that wasted time down to almost nothing.
The 10 tools that stuck around
I tried way more than ten apps this year. These are the ones still open on my laptop right now, months later. That’s the real test, not how good they sound on paper.
1. Claude for writing and editing
I use Claude to write first drafts, turn messy client notes into a clear outline, and fix tone problems in a piece of writing. What won me over wasn’t the writing itself. It was asking Claude to explain why a paragraph wasn’t working. That’s the kind of feedback I used to pay a human editor for.
Lesson learned: don’t ask it to copy a client’s exact writing style from just one email. It will guess, and the guess is often a little off. Now I paste in three or four samples of their writing first, and the results are much better.
2. Otter.ai for meeting notes
I record every client call and let Otter turn it into text. The real value isn’t the transcript itself. It’s being able to search it later. When a client says, “remember what we talked about with the budget in March,” I just search for it instead of digging through old emails.
3. Descript for editing audio and video
This one app replaced a $50-a-month video editing tool. I edit podcast audio by deleting text, not by dragging clips around on a timeline. One time, a client gave a bad answer during an interview, and I fixed it without asking them to record it again.
Mistake I made: I once used the “remove filler words” feature too much, and the audio started to sound stiff and robotic. Now I always listen to the full thing at normal speed before sending it to a client, instead of just skimming it.
4. Notion AI for staying organized
I already used Notion, so adding this was easy. It’s great at turning a long client email thread into a short list of action items. It’s not great at anything creative, but I don’t use it for that anyway.
5. Canva’s AI tools (Magic Studio) for quick design work
I’m not a designer, and I never will be. But clients sometimes need a simple social media graphic or a one-page flyer. Magic Studio gets me most of the way there in a few minutes. For anything that needs a real design eye, I still hire an actual designer. This tool isn’t a replacement for that, and I won’t pretend it is.
6. Perplexity for research
When I need to check a fact or get a quick overview of a topic, Perplexity is faster than opening ten browser tabs, because it shows the sources right next to the summary. I still click through and check anything that will end up in a client’s project. I don’t trust AI summaries blindly anymore, not after one awkward correction email I had to send a client.
7. Calendly with AI scheduling
Nothing fancy here, but it quietly got rid of the back-and-forth emails asking, “does Tuesday at 2pm work for you?” The AI part just suggests good meeting times based on my usual schedule. It sounds small, but it removed a lot of annoying emails.
8. ChatGPT (voice mode) for thinking out loud
I use this one a little differently than the rest. I talk through client problems out loud while I’m on a walk, instead of sitting at my desk feeling stuck. A lot of the time, I don’t even use its exact suggestions. Just talking it through helps me think more clearly on my own.
9. Wave for invoicing and bookkeeping
I used to pay a bookkeeper a small monthly fee just to sort my transactions. Wave’s automatic sorting got good enough that I only need to check it once a month, instead of paying someone to do it from scratch.
10. Gamma for client presentations
When a client needs a proposal ready by the next morning, Gamma turns a rough outline into a decent-looking slide deck fast. It won’t win design awards, but it’s much better than what I used to throw together in PowerPoint at 11pm.
How to set this up yourself, step by step
If you want to try something like this without wasting time like I did, here’s the order I’d follow:
- Track your real week first. Before you sign up for anything, write down every task you do for three or four days. You can’t fix a workflow you haven’t actually looked at.
- Start with writing and research tools. These are the safest place to begin. If a research summary is wrong, it’s easy to catch before it causes a real problem.
- Add one audio or video tool only if you need it. Don’t buy a tool like Descript “just in case.” Add it once you actually have a repeated task that needs it.
- Always check anything before it goes to a client. I read or listen to everything before I send it out. No exceptions, even now.
- Cancel old subscriptions right away, not just in your head. I told myself I’d cancel a design tool for two months and never actually did it, because I never opened the cancel page.
Common mistakes to avoid
A few things I’d warn my past self about:
- Don’t trust AI-made numbers. I once let a summary tool give me a stat that was two years out of date. Always double-check anything with a number in it before it reaches a client.
- Don’t skip the final check. These tools save time, but skipping the last read-through or listen-through wastes that time the moment something goes wrong.
- Don’t fully automate client emails. I tried letting AI write full client emails on its own for a week. It felt cold, and one client noticed right away. Now I only use AI for the first draft, and I always add my own final touch.
- Don’t pick the cheapest option just because it’s free. I dropped a couple of tools only because they were free, then switched back to a paid plan because the free version’s limits wasted more of my time than the subscription cost.
Where I ended up
I’m not running my whole business on autopilot, and I don’t think that’s a good goal anyway. Clients hire me because they trust my judgment, not a computer’s. What changed is that the boring, repeated 40% of my work, like transcribing, summarizing, writing drafts, sorting expenses, and scheduling, now takes a fraction of the time it used to.
That extra time now goes toward the parts of the job that actually need a person: strategy calls, tricky conversations with clients, and writing that needs real care behind it.
If you’re not sure where to start, my honest advice is to pick just one tool, probably whichever task eats up the most of your week right now, and give it two full weeks before deciding if it’s worth keeping. That’s about how long it took me to stop second-guessing every AI suggestion and start trusting the ones that actually earned it.