Project Two: Thinking out loud

What is the best way to analyse unstructured data using AI? I don't know. So I am using Project Two to figure this out based on 157 sources and counting...

Joel@Discy

asked me a question this week: How are you analysing the responses to project 2?

It's a great question. AI tools like Notebook LM give us the chance to analyse large amounts of text/ unstructured data in ways we never have before. Getting the best out of such data will be a key skill and maybe even a source of real competitive advantage.

Right now, my answer is: I don't know.

This week has been the first time I have had a chance to start digging into all 157 submissions. Rather than plug away in the background, I thought it would be worth sharing my process. Not because I think it is especially clever. I think it is an opportunity for everyone to learn, especially me.

I am starting with three things:

1. Doing some basic statistical and demographic analysis.

2. Reading/ watching/ listening to every submission. For me this is an experiment so it is worth putting in the hours to understand how the AI compares with my manual effort.

3. I have generated two short podcasts using Notebook LM.

One is a general review of the Project two submissions.

The other is based on a different Notebook. Alongside Project Two I have been keeping a bunch of "expert" sources - PwC, McKinsey, Bain, some big VCs, some academics etc. My idea is to compare what people are really doing with the best practice/ thought leaders/ expert consensus view.

The podcasts themselves are an experiment. I don't consume information this way and I am deliberately trying not to steer these first versions or use my own thinking to influence them.

Love to know what you think about any of these questions:

• How well do the podcasts work? Should I do new versions as the analysis progresses?

• Do my first steps make sense? What ground work have I missed?

• How would you answer Joel's question? What is the best way to analyse such a great set of responses?

Thanks for being part of Project Two