Actions
Forum Thread Details
  • Replies 18 replies
  • Subscribers 94 subscribers
  • Views 467 views
  • Users 0 members are here
Related

Which AI tool do you use for documentation?

prajaktaa

I am curious to know that which AI tool the community members are using to create or improve technical documentation.

Whether it is for the project documentation, API documentation any troubleshooting guides, or internal knowledge base articles, AI has become a helpful assistant for many professionals.

Some popular options include ChatGPT, Claude, Gemini, GitHub Copilot, Microsoft Copilot, and Perplexity, but each has its own strengths and limitations.

My questions are:

  • Which AI tool is easy to use for documentation?
  • What type of documentation do you create with it?
  • What do you like most about the tool?
  • Have you found any limitations or challenges?
  • Do you rely on AI-generated content as-is, or do you always review and edit it before publishing?

I'd love to hear about your workflow, practical experiences, and any tips or best practices you've discovered.

Parents

  • I don't use AI for documentation.

    For serious projects I will usually run separate hardware and software documentation  streams.

    The documentation starts before and CAD or IDE are used.

    The idea is to describe things before the tools start to get in the way. As the design progresses you develop the documentation and then the code/schematic etc.

    For example - if you will use a micro-controller you need to decide which one. To make a good choice you need to know what it is going to do - and the design notes (as I call them) are used to develop ideas and define them in increasing detail.

    It is generally accepted that very early design decisions often define most of the project - and so they are very important. If you use AI to do this you will have delegated most of the cost of the development in that first one page description you give the AI.

    Of course documentation written after the event is crazy - if you have to describe the design when its complete it should prompt you to at least wonder how you got there !

    I can't see where AI comes into a good design process - except as maybe a spell checker.

    I suspect that I am already in a minority - welcome to the 20230s - where AIs will use terawatt hours  of energy to produce peta Pages of output which other AIs will summarise down to .......

    MK

  • in reply to michaelkellett

    I really like your point of view. I agree that documentation should start from the beginning of a project instead of being written only after everything is finished. It also helps explain why certain design decisions were made, not just what the final result is.

    I don't think AI should make important engineering decisions either. For me, AI is more useful after the planning stage. It can help organize notes, improve writing, create summaries, or make documentation easier to read, while the engineer still makes all the important decisions.

    I think AI works best as a helper, not as a replacement for the engineer.

Reply
  • in reply to michaelkellett

    I really like your point of view. I agree that documentation should start from the beginning of a project instead of being written only after everything is finished. It also helps explain why certain design decisions were made, not just what the final result is.

    I don't think AI should make important engineering decisions either. For me, AI is more useful after the planning stage. It can help organize notes, improve writing, create summaries, or make documentation easier to read, while the engineer still makes all the important decisions.

    I think AI works best as a helper, not as a replacement for the engineer.

Children
No Data