Over the weekend I built a tool to begin to automate the SEO efforts for my companies - I strongly believe that SEO is a great growth channel, but most people are not patient enough to see great outcomes.
With Stairway, we were able to be the top search result on Google for many of our target keywords (typically high volume terms relating to topics within the Mathematics, Physics and Biology syllabi - certainly helped by being featured by BBC, among other publications).
In line with the aim of building Positive Sum to be run by as few people as possible, I am looking at ways to automate this. My target flow looks something like this:
- Input target keywords or topics to generate post ideas
- Write a first draft of post ideas
- Manually edit and improve the post, to reach a final version (this will mainly be to improve the tone of voice of the post, as well as to fact-check anything included. I believe that it will be possible to automate both of these in the future)
- Publish the blog post, via the API of a headless CMS being used by the website (including automatically populating metadata like tags)
The step I took over the weekend tackled the second step - here's how it works:
The code takes in a headline for a post, via Interval (so that I didn't need to waste time to manage a front-end application for myself, and in the future my team). The API takes this, sends it as a prompt to the OpenAI API (along with some additional information to improve the quality of the output), and then displays it (in formatted Markdown), as well as sharing it in a Slack channel so I have a saved copy for later.
N.B. I will never use this to write on my personal blog here.
The whole thing took less than 30 minutes to build. Here is the code required to build it: