If you’ve followed the previous steps, you should have now a good technical summary, at least a decent working draft.
You should have also saved a considerable amount of time, especially if you fed (i.e. trained) Chat-GPT with material from different documents, spreadsheets, and notes. I suspect doing this by hand would take considerably longer.
Of course, you still need to proof the material, check for accuracy, and confirm any sources it cited.
Interestingly, there are other ways you can get Chat-GPT to improve the quality of the summary.
As mentioned earlier, I use Claude for most of the drafting but Chat-GPT for others.
So, with that in mind, here are some prompts I’ve used in Chat-GPT to refine the text:
- Specify the purpose of the summary – is it for a research review, academic paper, business report, general audience? Tailor the style and tone accordingly.
- Provide any necessary background context beyond just the source material, so Chat-GPT understands how to focus the summary.
- Give word count or length guidelines if you need a very specific summary size.
- List any key points, figures, or findings you explicitly want included or highlighted.
- Note any topics, sections, or information to specifically exclude or avoid redundancy.
- Request the use of headings, numbered lists, or other formatting to organize complex summaries with multiple key points.
- Have Chat-GPT write neutral, objective summaries to avoid injecting opinions or bias.
- Ask for multiple phrasings of complex concepts to find most concise wording.
- Require paraphrasing of all content and cite sources to avoid plagiarism.
- Suggest starting with a one-sentence thesis statement to encapsulate the key message.
- Request a closing sentence that reinforces the main takeaway or impact of the findings.
The more you can specify about exactly what you want in the summary and how it should be formatted and written, the better results Chat-GPT will provide. The key is being very detailed and explicit in your instructions.
Finally, you can also feed the same material into both Claude and Chat-GPT, then compare the results.
I’d recommend this as both tools have their own respective strength and weaknesses.
Writing technical summaries is just one of the ways you can use Chat-GPT to save time, improve your workflow, and, in my experience, make you a better writer.
Let me know how you get on.
- How to use CHAT-GPT to Summarize Technical Documents
- How to Write Chat-GPT Prompts for Technical Summaries
- How to Refine CHAT-GPT Prompts for Technical Summaries