As project managers, we’ve all faced the challenge of figuring out what went wrong on a project after it's finished—why it ran over time, blew past the budget, or failed to meet expectations. Recently, we tackled this issue head-on by comparing two versions of a project plan—an initial one and a second from about six months later. Using ChatGPT, we dove into the details of the project to uncover the real sources of cost overruns and time delays, providing crucial lessons that any project manager can apply.
The analysis started with comparing baseline and actual data for each task. We fed both project plans into ChatGPT and guided it through specific prompts to pinpoint where the project diverged from the original plan. For example, tasks like “Requirements Definition” took longer than expected, and development costs were significantly higher than estimated. By leveraging ChatGPT’s ability to process large amounts of data quickly, we identified the exact points where things went off track. This kind of insight is only possible when you have baselined project schedules that are regularly updated, something every project manager should maintain.
One of the key findings came from identifying new tasks in the later project plan—ones that hadn’t been accounted for initially. This led us to uncover scope changes, such as additional development work or change requests, which drove up costs and extended timelines. Using ChatGPT, we could filter out irrelevant tasks and focus on the most impactful areas. The right prompts, like “What tasks are new in the updated plan?” or “Which tasks show the greatest cost overruns?” helped zero in on the problem areas, making the analysis both efficient and thorough.
In addition to identifying overruns, we used ChatGPT to formulate questions that project managers can ask before a project begins. Prompts like “Have task durations been validated by the team?” or “What’s your process for managing scope changes?” can help uncover potential risks before they escalate. ChatGPT can also be a great tool for facilitating lessons learned sessions, where you can use specific questions based on real data to guide meaningful discussions about what worked, what didn’t, and how to improve next time.
Key Steps to Analyze Your Project and Uncover Lessons Learned:
If you’re interested in using ChatGPT to analyze your project and discover lessons learned, here are some key steps you can follow:
Gather Your Project Documents:
- Start by compiling your project schedules, baseline plans, and any updates that show actual progress. Be sure to include key metrics such as task durations, start and finish dates, baseline costs, and actual costs.
Cleanse Your Data:
- Make sure your project files are free of unnecessary or incomplete data. Remove tasks that are irrelevant to the analysis (e.g., placeholders or completed without impact) and ensure that baseline and actual metrics are aligned. Ensure tasks are clearly labeled to make comparison easier.
Identify Key Areas for Analysis:
- Use ChatGPT to assist in comparing baseline versus actual data. Start with prompts such as:
- “What are the differences in task durations between the two project plans?”
- “Which tasks exceeded their baseline costs the most?”
- “What tasks appear in the later version but not in the earlier one?” These questions can quickly highlight the tasks where things went wrong.
- Use ChatGPT to assist in comparing baseline versus actual data. Start with prompts such as:
Run Comparative Analysis:
- Analyze specific metrics such as cost overrun, delays in task completion, and scope changes. Use detailed prompts like:
- “Show me the tasks with the highest variance in planned and actual completion times.”
- “Which tasks were added after the initial plan, and how did they impact costs?”
- This will allow you to isolate the tasks driving overruns.
- Analyze specific metrics such as cost overrun, delays in task completion, and scope changes. Use detailed prompts like:
Turn Findings into Actionable Lessons:
- Once the analysis is complete, use ChatGPT to help craft questions for future lessons learned sessions. For example:
- “What would you change in task estimation to avoid overruns like those in Development?”
- “How could earlier identification of resource bottlenecks prevent delays?”
- “What processes need to be in place to control scope creep effectively?”
- Once the analysis is complete, use ChatGPT to help craft questions for future lessons learned sessions. For example:
Document and Share Lessons Learned:
- Summarize the key findings from your analysis into a structured document that identifies specific overruns and their causes. Include clear lessons and actions that can be applied to future projects, ensuring that the knowledge is shared across teams.
Prompts to Try in Your Own Analysis:
Here are some additional prompts you can use when diving into your own project data with ChatGPT:
- “Compare the baseline cost and actual cost for each task in my project plan.”
- “List the tasks that caused the most time delays and explain how they impacted the overall timeline.”
- “Identify the tasks where rework occurred, and what impact it had on project costs.”
- “What scope changes were introduced, and how did they affect both time and budget?”
- “How did resource allocation contribute to delays or overruns?”
By leveraging ChatGPT for these types of detailed project reviews, you can uncover insights that might otherwise be missed, turn project data into meaningful lessons learned, and prepare more effectively for your next project. Whether you’re identifying scope creep, resource bottlenecks, or task delays, this approach ensures a clearer understanding of where things went wrong—and how to avoid similar pitfalls in the future.
