40 vs 7 Hours for Analysis & Strategy (Without or With AI)

From time to time, I face a familiar sequence of tasks:

Find and filter analytical materials related to design, the creative industry, and other sectors — plus the economy at large. Reports, articles, video materials — a bit of everything.

Read, watch, and simultaneously take notes on the most interesting ideas and observations.

Turn all this into a strategic action plan for our agency — plus a few general “good to know” insights.

This time, I decided to go all in and use the wonderful AI tools at hand: the good old ChatGPT and NotebookLM (since we already have a corporate Google Workspace subscription, the choice was obvious).

Here’s what we happily and quickly did together:

With ChatGPT, we drafted a list of research institutes and organizations whose analytical publications might be useful for me.

I found and filtered them, adding a few good videos on running a creative business.

I uploaded the whole dataset into NotebookLM — about 4.5 hours of video and 546 pages of publications — and asked it to:

analyze everything in detail,
organize findings by thematic blocks,
give those blocks engaging and informative titles (so it’s more fun to read), and
merge overlapping ideas from different sources.

I skimmed through the original materials to check what NotebookLM referenced (or didn’t), made sure all key topics were covered, and took a quick look at visual elements — infographics and nice graphics are always worth seeing firsthand.

Once I confirmed that the analysis was solid, I moved everything into our Notion knowledge base and shared it with the team.

Finally, I fed NotebookLM our existing insights on positioning — what we’re already working on — and asked it to build a roadmap based on all the research we’d compiled.

What’s important to note: I gave it a very high-quality foundation — a mix of recommendations from industry leaders, design experts, and experienced entrepreneurs from other fields. These are insights I’ve collected over the years through conversations with professionals and simply very wise, generous people. Combined with up-to-date data and reports from reputable institutions, the result turned out quite good.

Then I asked NotebookLM to estimate how long it would have taken me to process this material manually — and that’s how the title of this post came to be 🙂

10 hours

This month, I’ve spent 10 hours on analytical work in total (I track my time by major categories every month)

3 hours

on unrelated tasks

7 hours

on the work described above

40 hours

NotebookLM estimated that manual analysis of such a dataset would have taken around 40 hours — a rough calculation, but knowing how much I love deep, slow, meticulous research, the real number could have been even higher

It’s deeply satisfying — both to feel the time saved and to notice how much lighter, easier, and more enjoyable the whole process has become, including psychologically.

In the coming weeks, we at Tigre Rossa will also share how our experiments with AI tools for designers are going — the further we go, the more fun it gets 🙂