Mine has shat the bed, can’t deliver basic requests

Facing Technical Challenges with AI Assistance: A Personal Experience

In recent days, I’ve been attempting to update my Continuing Professional Development (CPD) portfolio. My goal was straightforward: extract relevant information from multiple PDFs, consolidate this data into a single document or an Excel spreadsheet, and streamline the entire process.

Initially, I considered automating the task by writing a script to download and compile the files automatically. I had a portfolio hosted on a website that was nearing closure, so automating the download seemed like an efficient solution. However, the process proved more complex than anticipated, and after some trial and error, I decided it was simpler to download the files manually and upload them individually.

The real challenge began when I turned to AI assistance. I relied on an AI tool to help with formatting, data extraction, and organization. Unfortunately, the AI consistently failed to meet expectations. It repeatedly asked me to verify formats and preferences, claiming it was preparing drafts or readying content, but then failed to deliver anything usable. Despite multiple prompts, the AI would not produce the requested data.

At best, it managed to generate large blocks of text that I could copy and paste. However, these outputs were incomplete—missing critical information such as dates and specific details essential for my portfolio record-keeping. This was perplexing, especially considering that this AI had previously been effective with tasks like creating Excel sheets and PowerPoint presentations.

This experience has been frustrating and somewhat baffling, as I did not encounter such issues before. It raises questions about the reliability of AI tools for basic tasks and highlights the ongoing challenges users face when these systems don’t perform as expected.

Conclusion

While AI-driven tools hold great promise for streamlining workflows, recent firsthand experience underscores that they can still be unpredictable and limited, especially with seemingly simple requests. It’s a reminder to remain cautious and maintain backup plans when relying heavily on automation—sometimes, manual effort remains the safest route. If you’ve faced similar issues or have tips for overcoming these hurdles, I’d love to hear your insights.