How I am organically introducing AI into my writing workflow.
I don’t think it’s a stretch to say that integrating AI into my writing workflow has been one of the biggest productivity boosts.
Not being a native English speaker means I have to spend considerably more brainpower when it comes to writing. Speaking in English comes easily to me, given I have been doing it regularly for the better part of 8 years now.
However, writing has always been a challenge.
That’s why, I was overwhelmed by a rush of “imposter syndrome” when some of my stories on Medium started doing well. It was a surprise to me that people actually enjoyed what I wrote.
Since then, I have gone on a journey to improve my writing — both in style and content.
One Thing I Will Never Do
Using Artificial Intelligence (AI) to give a little boost to all my writing has been one of the biggest productivity hacks that I stumbled upon.
That being said, I will never publish writings written by ChatGPT, Google Bard, or any of the other gazillion bots that are out there today. That defeats the purpose of my writing.
I write because I enjoy writing and sharing my thoughts and ideas with the world. If you have read my other writings, you will definitely know when you come across something that I wrote.
Instead of creating an ‘irtizaGPT” that mimics my writing, I am using AI to make my writing workflow more efficient. The more time AI can save me by automating some of the tedious parts of writing, the more time I can spend thinking, reading, writing, and doing other things that I truly enjoy.
Now, let’s look at 5 ways I am using AI to help me write more efficiently.
(1) Brainstorm Title Ideas
I objectively hate coming up with title ideas for my blog posts. Maybe that’s because I am really bad at it! I have no shame in admitting that.
Some of the best pieces I wrote before performed horribly because of bad titles. I can never strike a good balance between being informative, concise, and engaging, all at once.
Enter ChatGPT and Google Bard.
Here’s my new workflow:
- Focus on writing the best blog post I can
- Copy / Paste everything into ChatGPT or Google Bard console
- Ask it the following prompt — Generate 10–15 titles for this blog post. Make it engaging, and concise, and make sure you focus on <insert_focus_1> and <insert_focus_2>.
- Pick 2–3 that I like the most
- Edit it further to tweak things as I see fit
- Finalize the copy and just go with it 🚂
Since introducing this framework, I spend significantly less time stressing about title ideas.
(2) Gain Insights from Books
There are 2 primary ways I extract key insights from books:
- Highlight as I read; then digitalize the notes in my note-taking app of choice, Bear.
- Use Shortform
Between these two approaches, I can retain most of the critical information that I want for every book I read.
Every Friday, I have a recurring task to remind me to go through a random book’s notes to consolidate the ideas better.
Now, let’s talk about a subset of books that I don’t necessarily enjoy or read cover to cover. There’s also another set of books that are full of “fluff”. I don’t find it worthwhile to sift through all the fluff and extract key information.
For both these sets, I turn to AI.
I feed the following prompt to ChatGPT or Google Bard:
Generate 3 themes, 3 key insights, and 3 action items from Atomic Habits by James Clear.
Of course, I am using Atomic Habits just as an example here. I love that book, and I think in reality it’s the book with the most highlighted text.
Now, you might ask, if I didn’t enjoy the book, or don’t find it worthwhile to finish, why even bother to extract key information from it?
I am very selective when choosing books. So, most likely, I started reading a book because I value the author’s opinion, or it’s a topic that greatly intrigues me. Just because I am not enjoying the activity of reading doesn’t necessarily mean the book doesn’t have any important information to convey.
That’s where the prompt helps. I go over the following information generated by the LLM Model:
- Key Insights
- Action Items
If I find any of these interesting, I put them in my content backlog to do more research.
If not, I just discard everything and don’t ingest them into my second brain.
(3) Write SEO Meta Title and Description
One of the best ways to grow online as a writer is by leveraging the power of Search Engine Optimization or SEO.
Medium, as well as most other blogging platforms, provides the user with custom control over the Meta title and description which are used by search engine crawlers to decide how to order the user’s website and blog post in the SERP (Search Engine Results Page) page.
This lets the user customize their blog posts to target particular “keywords” to get more traffic.
If you are good at this, you can grow your website really fast.
However, I am bad at it. Actually, it’s more accurate to say, I am not really bothered by it. It’s not something that interests me, so I don’t spend any time learning about it.
Until a few months ago, I didn’t even bother to customize my stories to include meta titles or descriptions. I just went with Medium’s default, which is one of the worst things you can do.
Now, instead, I use ChatGPT or Google Bard. I copy/paste my blog post into the console, and use the following prompt:
Given the blog post above, generate a Meta Title (under 60 characters) and Description (under 150 characters).
I want to stress, this is so much worse than actually spending time on SEO to optimize your blog post. However, it’s better than nothing. It’s good enough for me.
(4) Summarize Research Articles
Summarizing long blog posts and research articles is one of the most practical uses of AI. Both ChatGPT and Google Bard are exceptionally good at it.
Numerous times I come across research articles that are interesting, but not interesting enough to go through the whole thing in one go. Also, sometimes the technical details are a little too much for me to understand.
In these cases, ChatGPT or any other Large Language Model (LLM) can be a great help!
Not only do I use LLM to summarize these articles, but I can also ask it questions about the article. Here are a few cool things you can do too if you want to be even more engaged:
- Create a pop quiz based on the article
- Ask for counter-arguments to understand “the other side”
- Generate key insights and action items
- Extract citations so that you can add them to your read-it-later app for future reading
What’s the point of doing all this?
Every content I write is inspired by things I experience, read, listen to or watch. So it’s important for me to ingest things I read into my second brain.
(5) Draft Emails Quickly
The emails span various topics — from programming help to asking for code to sponsorships and advertisements.
I hate not replying to emails. In most cases, people put in their time and effort to reach out to me, so the least I can do is reply back to them.
Previously, the process of replying to most (if not all) emails was a very time-consuming process. However, AI has changed that for me.
Now, I can quickly draft responses to emails using ChatGPT or Google Bard.
It’s about to get significantly better though. Google recently announced a ton of AI improvements to their products, including Gmail. I cannot wait for it! Check out the announcement for more details.
There you go, folks!
I hope you found this valuable.
As with all my other AI writings, it’s important to mention that I am trying out new ways to integrate AI into my different workflows every day. So, everything I write about here is subject to change.
When and if things change, I will make sure to update you folks about it. Who knows? Maybe you will find something interesting to integrate into your life.
Want more insights like this?
You will get early access to all my Medium work and exclusive access to tons of things I don’t post on Medium!
If you enjoy my writing on this platform, join Medium so you can get unlimited access to valuable and beautiful writings from great writers.