Using ChatGPT for competitive analysis as a Product Manager - Mix of success and failure
- Harshal

- 4 days ago
- 2 min read
If You Can Copy-Paste, You Can Use GenAI
I had published Using ChatGPT for Competitive Analysis as a Product Manager – Mix of Success and Failure on YouTube, and wanted to also keep the insights from my testing available in blog form, too.
I spent 10 minutes writing this and some time making the video. You need 1 minute to read this, and 20 minutes to watch the video.
Related:
Back in 2023, competitive research meant reading through dozens of customer case studies to learn where competitors sell, who their buyers are, and what needs drive adoption. In the video, I built a framework to capture jobs-to-be-done, geography, industry, and product needs from over 40 case studies, using ChatGPT to speed up the process.
This approach is basic compared to 2025’s advanced methods (such as using n8n). But it shows that GenAI can speed up your work, even if you are limited to simple copy-paste operations.
Prompt Goals
I tested several prompting methods, both with and without web browsing. The goal was to get structured tables instead of long, unformatted summaries when analyzing case (sample linked) studies. Results were mixed. ChatGPT extracted useful details from some cases but often missed the deeper customer needs or failed to turn long text into clear, job-focused insights. My goal was to ensure that the AI doesn't read the solution from marketing language in the case study, but instead tries to dive into the underlying problem that the customer faced before they purchased this solution. After multiple prompt iterations, it improved at identifying jobs-to-be-done, but refining columns and clarifying vague solution descriptions still took manual work.
Prompt trials:
Takeaway
AI can make competitive analysis faster, but it still needs human judgment to extract meaningful business insights. If you want to see the full walkthrough—including what worked, what didn’t, and how I refined the prompts—watch the video: link.
Related:









