Lightcast WordPress API Plugin Case Study
Full Narration Script (3 Minutes)
Slide 1 — Opening and Framing
This project did not result in AI building a plugin for me.
It resulted in me being able to build something I had no prior experience building, under a deadline, because AI shortened the learning curve and reduced dead ends.
That is the real value of AI here. It shortens the learning curve and makes unfamiliar technical work achievable.
This presentation is a short case study of how that played out while building a WordPress API plugin for Lightcast.
Slide 2 — The Existing Problem
We originally relied on a Lightcast iFrame integration. It was very easy to implement, but over time it became limiting. It was clunky, difficult to adapt, and often did not fit partner needs. We had very little control over how the data was used or presented.
Slide 3 — The Risky Decision
Because of that, I suggested switching from an iFrame based approach to a direct API integration and managing the presentation ourselves. That decision came with real risk. I had no prior experience building an API driven WordPress plugin, and we already had a deadline in place.
Slide 4 — The Technical Strategy
To reduce that risk, I made one clear decision early on. The plugin would be bare bones. No UI, no styling, and no presentation logic. Its only responsibility was to fetch data from the Lightcast API, store it correctly in WordPress, and expose that data through shortcodes.
This approach allowed us to focus entirely on building a reliable data pipeline. Once the data existed in WordPress and could be output consistently, design and presentation decisions could happen later without blocking the technical work.
Slide 5 — How AI Helped and the Outcome
This is where ChatGPT became extremely useful, not as a coding agent, but as an in browser technical guide.
The first phase was planning. I used ChatGPT to break the work into clear steps: understand the Lightcast API, identify the required data, decide how it should be stored in WordPress, and expose it through shortcodes. That structure prevented the work from turning into unfocused trial and error.
The next phase was implementation. ChatGPT helped interpret the Lightcast documentation and understand how GET and POST requests needed to be formatted, what payloads were required, how responses were structured, and how errors should be interpreted. This significantly reduced guesswork and made progress more predictable.
Then came debugging. Requests failed, data did not persist, and responses were not always what I expected. Instead of guessing, ChatGPT helped narrow down what to check next, whether the issue was request formatting, response handling, or WordPress behavior. It did hallucinate at times, which was frustrating, but treating its output as guidance and verifying everything kept the process moving.
The outcome was not that AI built the plugin.
The outcome was that I was able to build something I had no prior experience building, under a deadline, with fewer dead ends.
That is the real value of AI here. It shortens the learning curve and makes unfamiliar technical work achievable.
Slides (Minimal Content)
Slide 1 – Opening
- AI as a learning accelerator
- Not automation or replacement
Slide 2 – The Problem
- Lightcast iFrame integration
- Limited flexibility
- Hard to adapt to partner needs
Slide 3 – The Decision
- Switch from iFrame to API
- Deadline already in place
- No prior API plugin experience
Slide 4 – The Approach
- Bare bones WordPress plugin
- Data pipeline only
- Shortcodes for output
Slide 5 – The Outcome
- Planning unfamiliar work
- Understanding API documentation
- Faster debugging and iteration