Imagine sitting in a room with representatives from McKinsey, BCG, Deloitte, KPMG, Bain, PwC and other leading consulting firms.
Imagine being able to ask them questions like:
- What actually makes AI transformation work?
- Why do so many companies struggle to get real value from AI?
- What should leaders be focusing on today?
It would be a fascinating conversation. Probably a very expensive one too.
So I tried something else.
While working on AILAS, I ran an experiment to see how close I could get to that conversation by using AI to interrogate their latest AI transformation reports.
Here’s what happened.
The experiment started with one question.
Do organisations struggle with AI because they lack AI tools? Or because they lack a structured way to connect organisational knowledge, governance, workflows, decision-making and measurable value?
It was a reasonable hypothesis. But I wanted to dig deeper and see whether the evidence pointed in the same direction.
My idea
The major consulting firms already publish detailed reports on AI transformation. Those reports represent years of research, observations and recommendations.
So I wondered: what if, instead of reading the reports one by one, I could interview them through NotebookLM? 📝 If you’re not familiar with NotebookLM: Unlike a general-purpose AI assistant, NotebookLM grounds every answer in the documents you provide, making its responses traceable back to the original sources.
I gathered thirteen of the most recent AI transformation reports from McKinsey, BCG, Deloitte, KPMG, Bain, PwC, EY and the WEF/Capgemini.
I uploaded the reports into NotebookLM, then interviewed each consultancy separately using the same set of questions.
1. Where is enterprise AI transformation failing?
2. What blocks measurable value?
3. Who should own AI transformation?
4. What changes with agentic AI?
1. Where is governance becoming essential?
Expectations vs reality
The goal wasn’t to ask AI for opinions. It was to interrogate existing research in a consistent, source-grounded way, then compare the answers to understand where the reports agreed—and, perhaps more interestingly, where they didn’t, and why. But honestly?
I was also curious to see how far I could push this style of interrogation before the whole experiment fell apart.
Here is what happened.
It immediately felt as if I was interviewing different people. McKinsey approached problems differently from Bain. Deloitte worried about different things than PwC. Some consistently returned to governance, while others couldn’t stop talking about operating models or leadership.
Unsurprisingly they agreed on most of the answers. But they emphasised different aspects of the same problem, almost as if each consultancy had developed its own personality.
What I learned
The experiment wasn’t really about proving a hypothesis I already believed. Of course, I wanted to dig deeper into these questions, understand the why, and perhaps uncover a few hidden gems that aren’t immediately obvious when reading the reports individually.
But my biggest discovery wasn’t one of the findings. It was the methodology itself.
I had stumbled upon a completely new way of doing research and learning. Instead of passively consuming information, I was actively interrogating it. The process was engaging, fun, and far more memorable than simply reading report after report. By the end of the experiment, I wasn’t holding a summary of thirteen reports. I had the first draft of an interview guide for real enterprise conversations.
Conclusion
Three things that stood out:
First, reports became conversations. The biggest shift wasn’t in the findings. It was in the experience. It completely changed the way I interacted and absorbed new knowledge. Second, I wasn’t looking for answers. I was looking for better questions.
At the end the biggest value wasn’t in confirming what I already suspected. It was in understanding why different consultancies reached similar conclusions, where they placed their emphasis, and which questions deserved to be explored further with real organisations.
Third, this approach didn’t replace actual research. Nothing replaces talking to real customers, consultants and experts But it prepared me to be able to ask much sharper questions when I will eventually walk into those real conversations.