A slate of writing tools powered by artificial intelligence (AI) have recently made headlines as a new and more effective way to produce content. Most tech companies have even been implementing AI into their own tools as well, like we did earlier this year when we launched our AI assistant.
I’ve written about my conflicted feelings about these very AI writing tools, but I’ve been making an effort to push past my hesitation and add AI into the articles I write for the Buffer blog. The hope was that by doing so, I could speed up my writing process.
So, despite my initial resistance, I went ahead and used AI when outlining and drafting blog posts. Here’s what two weeks of incorporating AI looked like and how my thoughts on AI writing tools have changed.
When it came to choosing which AI tools to use, I opted for convenience. I’m constantly working in Notion and use it for outlining blog posts, so it was very easy for me to incorporate their recently launched tool, Notion AI, into my planning.
Since we recently added Buffer’s AI assistant into our product, I’ve also been experimenting more with this feature and typically have the tool open during the work day so I can just easily switch over to our ideas section and write. I didn’t use other more popular tools like ChatGPT because rather than learning how to use a new product, I wanted something that I could seamlessly embed into my existing arsenal of tools.
I was only using these tools for outlining and brainstorming purposes. I never prompted the AI to write out a full blog post for me, but instead just asked it questions to help me generate ideas and help build out the article. Throughout the course of two weeks, I ended up using AI for:
- Interview questions
- Suggestions for titles
- Brainstorming introductions
Right off the bat, I could quickly outline various topics quickly with AI. While I was still writing my own outlines, I was able to fill them out much faster than usual. To clarify, I never prompted the AI to “write an outline.” Instead, I just asked questions about my topic to see the key points it would generate. This really helped because sometimes when starting out a new post, I can have a bout of writer’s block and using AI essentially fixed that.
For example, when it came to my article on LinkedIn analytics, I used AI to help with the order of the post as well as some of the main points in each section. I did the same thing when refreshing one of our older posts on Instagram Story templates, as I updated the introduction and used AI to run through different versions.
I also found that these writing tools were a great way to boost my productivity when I was behind on work. One day in particular, I was feeling under the weather and took a half-day. The next morning, I was scheduled to have an interview with a TikToker about brand deals, and was able to generate a list of ten thoughtful questions in seconds with AI. While I did edit and tailor the questions to better fit my interviewee, this was a huge help at a moment when I really needed it.
Overall, it felt like brainstorming was easier thanks to AI. Writing is often a solitary act for me, so it was nice to have a sounding board to bounce ideas off of.
As I started using AI writing tools more, I began to concede that AI actually can produce pretty decent content. However, the one main issue I still have is that the writing isn’t always accurate. Like when it gave me the wrong list of steps after I asked it how to turn on LinkedIn Creator mode.
Similarly, when I used it for my article on Instagrams collab posts feature, the AI kept confusing the feature with the actual act of collaboration. So, it was instead giving me answers about how people could collaborate with each other on Instagram, which wasn’t the actual topic of the piece. I tried to redirect it a few times with no luck. But I may need to work on my prompting skills. My colleagues Tami and Phil wrote about the importance of good AI prompts and why adding context is necessary when asking the AI to produce work.
Even when I got back the right results, I wasn’t always happy with the final copy. As a content writer myself, I personally felt like the writing quality from these tools was good, but not great, and that I needed to edit it or spruce it up.
Another issue I had was that oftentimes the copy felt quite repetitive. Whether it was the AI repeating the same word over again or just the sentiment, this was a reminder that AI is simply regurgitating everything that has already been written before. Although that can sometimes make for decent copy, it’s not original and groundbreaking work. This is why real writers and their work should still be valued.
Ultimately, these tools didn’t really save me that much time in the end. While it was super helpful before my interview, for the most part I found that either I would still have to correct the AI or rewrite the work.
With that being said, I still found the AI useful. Before these two weeks, I wasn’t very comfortable integrating these tools into my everyday work routine, but now I plan to use them from time to time. While I personally don’t feel like I need these tools on an everyday basis, I recognize the benefits of using them on days I feel especially blocked.
If you’re also interested in experimenting with AI, sign up for Buffer for free and try out our AI assistant today!