How I Optimized Interview Note-taking: Guide included!

⁠

⁠

Intro: As a Product Manager, I chat with users every day. But taking notes while talking? Not easy. I used to record and then spend hours writing down the key points. Now, with GPT-4 and Google Sheets, I've found a faster way. Let me share how.

Content breakdown:

Problem:

How I used to do it:

Solution:

how you can try it yourself:

Benefits i see:

Conclusion:

Images/Gif/videos

Problem

As a PM I want to talk to users on daily basis. User talks are super insightful. Because am doing my interviews alone I am finding it hard to be present in the discussion and take notes at the same time.

Until recently my approach was to record my interviews. Then I synthesise the information afterwards. This normally takes about 2H (1h interview + 1h summarising, finding insights, tagging and sharing). If you talk to 5 users a week that is 10h.

Solution

I still spend 1h talking to my AEC professionals, totally worth it. I found a way to do the synthesis part.

All you need is → G sheets + GPT for sheets + GPT 4 subscription = Synthesis time/100 (below I show you how to do it)

Value

Spending less time on synthesis comes with the following benefits:

Instead of having to focus on documenting everything all the time. It free’s up mental power to zoom out and get the creative juices going. Less of a “secretary” and more an insightful leader

I dont stress about not having time to address other things.

Instead of looking at taking to users as a chore I look forward to it. (Not 100% sure if its because I like looking at how close GPT is getting to summaries the real points in the conversation, or because i actually like talking to site managers xD)

How to replicated it for yourself:

Get your transcript

Connect you Gsheets to GPT for sheets

Make an account for GPT 4 and get your API

Prompt engineering in Gsheets to get the close to the results you want (you can get the G sheet template, just comment “BSP!” and I will send it to you)

Copy paste transcript

Copy paste results into different tools as needed

Conclusion and reflections:

The approach above helps a lot, especially on the very busy weeks. But don’t expect that this will help you become a better PM. You will still have to put in the hard work of connecting the dots (as the most recently commercially available model is 16k tokens) and making sure everyone is aligned and “rowing in the some direction. As good as GPT can be, dont blindly trust it, make sure you read all the generated content and correct any halucinatinaitons might appear. Especially around things that require precision like deadlines and actions. I found that plugging in the transcript right after the interview helps a lot, since the conversation is still fresh in my mind and i can easily pick up on any mistakes. Overall it performs at around 80%-90% and this is with 3.5 turbo 16k. Can’t wait until the GPT 4 16k becomes commercially available.

Lastly I want to encourage you to try it and interact with the post so we can leverage not only AI tools but also community knowledge ;)

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.