ChatGPT turned one year old recently and so it’s time to reflect on what has happened in 2023 as a result of what was first viewed as disruptive, almost dangerous, technology. I was an early adopter. Once I realised that AI could make me faster, more efficient, and I could outsource some of my thinking to it I was sold. Since then I’ve become the agency AI bore and will shoehorn it into any conversation I can.

This year in SEO it has been impossible to hide from AI and the majority of news, conference talks, and angry tweets have referenced it in some way. To try and keep this article from spinning out of control and becoming a treatise on the fundamental benefit of technology adoption we’ll focus on a core theme. The focus will be the things I’ve done to speed up the simple stuff, enhance my capabilities, and how the “new and scary” has quickly become mundane and pedestrian.

So, what does all this mean for those of us on the ground doing the day-to-day optimisation work? Fortunately, for everyone reading this I can’t speak for all but what I can do is give a rundown on what I have been using generative AI for. 

My favourite things to do

1. Creating Fun Tools
2. Mastering Excel/Google Sheets Without Learning How
3. Writing Javascript and Python

Creating Fun Tools

One of the biggest doors that has been opened by Generative AI for me is turning ideas into reality. The best bit about this is the only thing you need is a good understanding of your idea and a bit of patience. The process goes like this:

Step 1) I have an idea, I would like it if I could get AI to read through the Google quality rater guidelines and then rate some text based on those guidelines. 

Step 2) Hey, Chat GPT here’s a description of my idea as a step-by-step prompt (read the guidelines, read the content, tell me what you think of it and how to improve it).

Step 3) Take the output from Chat GPT (in this case it was a Google Sheets app script) and run it.

Step 4) Watch it fail or give weird, unexpected results.

Step 5) Engage unlimited patience and iterate. Go back to ChatGPT and tell it what went wrong and why it didn’t match your expectations. 

Step 6) Keep going back and forth until it works!

Using this foolproof 6-step method I’ve managed to make some tools that, unsurprisingly, loads of other people came up with too. Along with the content quality grader, I’ve also successfully made a product description writer, an FAQ summariser (turn text into FAQs), a structured data generator, a redirect mapper (only works if the URL has proper words in), and a meta description writer. I’ve had a few failures along the way too, most notably an image ALT text generator that I’m still hoping to get working eventually.

In all of this, the most pleasing thing was to see the same ideas played out and released as actual tools by big software vendors. Conductor and Keywords Everywhere have both released a content quality grader that appears to work in the same way as mine though the Conductor version is executed in a much more professional way!

Mastering Excel/Google Sheets Without Learning How

If you’re reading this you probably have to use Excel to do more than the simple stuff like use a trusty LEN to make sure your page titles and meta descriptions are on point. But, if you’re like me, you probably don’t know how to get the most out of your spreadsheets without leaning heavily on Google. For anything more than a VLOOKUP I usually spend far too much time going through result after result trying to find what I want. More often than not the solution is close to what I need and I have to spend a while adapting it or putting two separate solutions together to make it work.

Now, this is where ChatGPT, for me, really shines. As long as I know what I want to do and I can express it in a step-by-step way then generally speaking I get to an answer that works in a few minutes. This means I can quickly do bulk actions that, for someone good at Excel, would be pretty straightforward. For example:

“Please write a google sheets formula that will match the string in cell b2 of sheet3 if it appears as part of the string in any of the cells in column D in sheet1 and return the full value of the cell in which it is partially matched”

And I get back:

=ARRAYFORMULA(INDEX(Sheet1!D:D, MATCH(TRUE,IFERROR(SEARCH(Sheet3!B2, Sheet1!D:D) > 0, FALSE), 0)))

And it works! Perfectly, first time. There is no way I could come up with that myself and it would have taken a long time to piece it together from Google searches. 

I now have a long list of formulas like this that I have been compiling over this year into a mastersheet that I, in true old-school SEO fashion, am pretty much keeping to myself… Not really, but I am keen to help people work out how to use ChatGPT to create their solutions rather than just giving them out.

Writing Javascript and Python

Once again this one has to do with creating a thing I couldn’t do without a serious amount of upskilling. Sometimes, the ideas I have can’t be done in a spreadsheet even with a custom app script or I want to use Chrome Inspector/Dev Tools so that’s where the ability to write Javascript or Python comes in handy. The only problem there is, that I’ve tried to learn both of these skills before but struggled sufficiently to put me off dedicating enough time to get good.

So instead, I get ChatGPT to do it for me! 

Some of these things are pretty simple like, I want to see all of the render blocking resources for a page or all of the external links. Well, ChatGPT can write a Javascript snippet that will do things like that when you run it in the DevTools console. Instead of using my favourite tool (Screaming Frog) to crawl the page I just have a load of scripts saved that I can run whenever I want. This makes getting to the data I want quick and easy.

The biggest thing I’ve worked on in this way is a link redirect automation tool. This took a few hours of back and forth with ChatGPT and Google Colab to come up with a script. The script that I ended up with uses a bunch of Python libraries to do some weighted semantic analysis and matching based on the URL to find the best match for a URL in column A in a list of URLs in column B. 

So Why Does Any of this Matter?

For me, the use of generative AI in SEO is not about getting a machine to write content, create wonky page meta titles, or produce images so I don’t have to use Shutterstock or a professional designer. It’s about the democratisation of technology, empowerment and, ultimately, being able to do more with less. The more work I can fit into every hour gives my clients greater value and enables our team to punch well above their weight. It comes down to using generative AI to enhance what we do, not replace it because, let’s face it, the technology just isn’t good enough to do the stuff that takes creativity. Writing engaging content, creating stunning images, and writing complex software from the ground up all take real skill that is acquired over time and with experience.

The one thing is that this technology is not going away. Learning to live with it is fine but, embracing it and leveraging the ways it can optimise your workflow is like having a shortcut to new abilities. It can be argued that if you lean too heavily on it it may stop you from acquiring new skills the right way but we also used to have to remember phone numbers and carry maps in the car to drive places.

These simple things have been outsourced to the magic rectangle in your pocket already so why not take the plunge and make 2024 the year you make the most of this simple, cheap way to do more. Download the complete This Year in SEO 2023 report.