How I Use AI
Practical uses for normal people
Everywhere you look there is an influencer or thought leader telling you how they use AI to 10x their life. They now have a virtual assistant, an accountant, an editor, an idea generator, a content repurposer. The last one's not a real word, but you get the point. AI can do all of the things, we get it. But how can a normal person use AI, practically?
Personally, I have been through a journey with AI. I set out earlier this year to be comfortable and capable with it. Now that I feel like I am both of those things, I'd like to share how I actually use AI and maybe inspire some ways you can leverage it too. I am not using AI to build AI Agents or anything extreme. I am simply using AI to improve a few key workflows in my life. If AI does not do them well most of the time, I tend not to use it for that thing.
This article was written in response to reader requests from Using AI to Be Human.
Editing
My first and most used feature of AI is editing. Mainly, I use it to edit my long form writing, like this newsletter. I also use it to edit long emails and documents.
For this newsletter, I use Notion AI's "Fix spelling and grammar" feature. Every idea for this newsletter is stored in Notion, in a "Content" database. I collect all my thoughts and resources on that page's body. The first and second drafts are written in IA Writer. When I am happy with the draft, I copy and paste the full letter into the page of the original idea in Notion. I then highlight all of the letter and press the "Ask AI" button, and select "Fix spelling and grammar".
Notion has hundreds of thousands of words of my writing in it, so it understands my tone and writing style. It takes that into account and edits the letter in seconds. A 1,000 word letter takes about 10-20 seconds. I then read over this edited draft and make any further refinements. If I make a lot of refinements, I will let Notion AI fix spelling and grammar for that section again.
This feature alone has helped immensely with my writing. As a person with dyslexia, more often than not, I do not always see errors in my writing. This simple review from Notion AI gets any of the kinks out, without changing the tone or structure of the letter.
My wife does a deeper read and edit after this. It is rare that she needs to make any edits since using Notion AI to edit. It's pretty great!
Meeting Transcripts
I have never been a fan of most meeting recorders. Over the last few years, I have tried a few. They always felt bulky and intrusive. Admitting them to the meeting and having them take up a seat on a video call has always felt unnatural. When Notion launched AI Meeting Notes a few months ago, I gave it a try. Honestly, my hopes were not high.
I was blown away and have used it ever since. I have a Meetings database in Notion. My default template now has this block type added by default.
Firstly, I like how it is not admitted to a meeting. I still ask if the other guests in the meeting mind if the meeting is recorded, but it is not a robot on your screen. It just runs in the background on the recorders’ computer.
Secondly, I appreciate how there is a notes section for open capture. This is where I add the meeting objective, topics, notes, and action items. You can take notes manually, while Notion AI does. No meeting should ever occur without a clear objective and topics to discuss.
After the meeting is over, Notion reviews the transcript and your notes to build the summary. The summary is extremely accurate and rarely misses anything. If you think it did miss something, you can query the full transcript for anything specific, right from the meeting note.
Proposals and Admin
When I was consulting, I took a lot of discovery calls. When I first started taking these calls, creating a proposal and follow-up emails was a long and arduous process.
I'd have to take very detailed notes. Then, I'd have to review them. After review, I'd have to put together the proposal and send them a follow-up email. I would personalize that email so it felt as custom as possible. This took anywhere from 30 minutes to 2 hours, depending on the complexity of the project.
By using Notion AI Meeting Notes and a few templates with AI blocks in them + Notion Mail with snippets, this process was under 10 minutes almost every time.
Notion AI would transcribe the call and produce the summary. I would copy the link to that page and spin up a proposal template. On that page, there would be a Notion AI block with the following prompt:
Create me a proposal based on these examples:
- example 1
- example 2
- example 3Using this meeting transcript: {insert page link} Work can start on: {date} Project timeline (estimate): {weeks} Project Cost: {price}
I would then create a near perfect proposal in about 20 seconds. After, I would review it in detail, making adjustments as needed.
Once the proposal was buttoned up, I would change the proposal's sharing settings and draft an email.
On the original meeting page, there was a Notion AI block with a simple follow-up email prompt. I would press that button to generate the email, launch Notion Mail and use the follow-up discovery call snippet. I would paste in the custom follow-up email that Notion AI generated and link the proposal.
I had about 3–5 other similar workflows, but this was the most used, as it produced the most consistent results. The results were consistent because there were a lot of examples to reference as a starting point.
Search
For the last 25+ years, Google has dominated search. While it has done it successfully, it is not the best way to search in 2025. While AI can hallucinate, and the skill of search—and more widely, research—is very important for us all to have, AI is more efficient most of the time. Using AI for a simple search is great. It gets you right to the answer, not a sea of links, which is then filled with ads. I will just preface with, choose when to use AI search. If you think it is the most accurate way to accomplish the task, use it. But if you think it's the easiest, but not the most accurate, comb the links yourself.
I enjoy using AI research to point me in a direction. A lot of the time, when we start research on something, we don't know where to start looking. This is when AI shines. While it might not always be right, it doesn't hesitate.
Like any AI prompt, giving it all the context is key. For deep research, I use Claude or Notion research mode. Don't treat this as a general search, tell it who you are, what you do, why you're doing it. Feed it documents and links. Tell it what steps to take, such as "review x, then wait". After it has all of the context, give it its marching orders.
Using a research mode in any of these tools takes time. This is the key difference. It is reasoning at a deeper level. Understanding parts of the task before moving on to the next part. Some research can take 5-10 minutes.
To be clear, never use AI as the final truth. Always double check it. Review its sources in detail. The more explicit you can be with the sources the better. As an example, point it to specific data or papers, rather than giving it free rein to search all of the web.
Pro-tips
Keep AI at the ready
If you are on a Mac, install the Chat GPT Mac app. If you are on mobile, download the app and put the widget on a screen. On Mac, press option + space to bring up Chat GPT from anywhere. If you use Claude or Notion AI, they have similar options. Personally, I use Chat GPT as my general search engine and Notion AI for more complex research.
Keep a Prompts Database
I originally got this idea from Jeff Su. Keeping a note, database, or spreadsheet of your prompts is great for consistency. I keep mine in Notion, tagged with the tool. In the page, I have the original prompt, my customizations, and how I have used it in the past.
If you use Notion, you can even use the custom AI block on a page, prefilled with the prompt of your choice. I use these all over my workspace. Below is an example for creating a comma separated list for newsletter subscribers.
Starting this year, I would have never guessed that AI would become so entrenched in my workflows. I was very hesitant at first. I understand the environmental, ethical, and societal impacts of AI. It will be a long time before it is regulated to any meaningful extent, and there will be a lot of hurdles to build, then jump through to get there, but the genie is out of the bottle.
If you aren't using the tech in its early days, you're going to find it harder to adopt when it is not optional. Start finding small, practical ways you can use AI in your life. Before you know it, it will be second nature, like Googling something is.