
Context
SitePoint has approximately 15,000 articles on its blog, many of which are not optimised for search. Updating the content manually would be incredibly time consuming, and not particularly cost effective for many of the older articles. As part of efforts to improve overall technical SEO for the site, I proposed using OpenAI’s API to augment the existing content and make it more discoverable in search.
We decided to add two sections to suitable articles: FAQs (presented using FAQ structured data) and Key Takeaways. After prototyping the process manually, we then rolled out tests on small subsections of the site to test effectiveness and check for hallucinations and other errors, before rolling out much more widely on the site.

My input
- Proposing the idea and getting stakeholder buy-in
- Working with the SEO team to prototype the initial idea of selected articles, then define an automatable process
- Helping to test and refine the process to add improvements
Outcomes
- Succesful rollout that required only a small amount of manual editing to fix errors/hallucinations.
- Traffic improved by around 15% in the first six months of 2024 following these changes.
Leave a Reply