Sweet Surveillance

There’s a famous quote often attributed to Ford: “If I had asked people what they wanted, they would have said faster horses.” There’s little evidence he said it. But there’s extensive evidence that it’s true. People don’t know what they want. And currently, almost all literate people on our planet are busy articulating what they want. To an AI of their choice. And very few of us are actually good at it. Because most of us lack the knowledge of what it can do, and also because our imagination skills for what it can do are still in the learning phase. And some of us aren’t good purely because of limited articulation abilities or tools.

Entrepreneurs like Ford succeeded because their observation skills, technological understanding, and pattern recognition were more sophisticated than everyone else’s. AI has all three at scales humans can’t match. So, quite simply, let it listen to everything. Enabling controlled access works better than amateur prompts. What I mean is: your internal meeting transcripts are the most valuable data sources for your AI agents. Your sales agent, content agent, design agent, innovation agent, should all attend every meeting.

Respect for otter and spinach. But meeting transcripts aren’t just note-taking material. There’s a lot more value in that data. Every decision, every misalignment, every cultural gap, every operational friction, it’s all there.

Here’s what listening can do. Sharing two small examples, quite different contexts, but the same principle:


Transformation Agent
Companies are investing heavily in AI transformation. I believe the bottleneck isn’t technology. It’s lethargy. People don’t adopt tools because they’re forced to. They adopt when they’re genuinely curious and excited. The transformation that sticks is the one where every person involved wants it.

That’s what a transformation agent can achieve. How? Most meetings end with individual tasks. Person A owns customer feedback analysis. Person B owns competitor tracking. Person C owns draft copywriting. The transformation agent listens to each task. For each person, it builds one skill. Person A gets a tool to automate feedback sorting. Person B gets a competitive intelligence scraper. Person C gets a copy variation generator. Not replacing them. Enabling them. Each person finds a tool on their desk the next morning that makes their specific job easier. That’s when people will actually understand what AI can do.



Culture Care Agent
A company’s culture is one of the most important factors for long-term growth, and, surprisingly, there are still no analytics for monitoring culture. It’s a weird corporate problem. It’s hard to manage, not because solutions don’t exist, but because non-adherence goes unnoticed until attrition reflects it. But for all the weird and complex problems, we now have AI.

I feel every company must have their own culture agent that should attend every meeting and point out gaps as per their guidelines. For instance, I still come across managers who manufacture fake urgency in every response (critical, top priority, need by EOD) and misuse borrowed authority (boss asked, founder said, leadership wants, etc.) as an everyday strategy. If you’re trying to build an innovation-friendly culture, it’s important to note such gaps. Things like these, and a lot more, can be easily tracked via a culture care agent.



Note: I have used the example of meeting transcripts because the tools to do so are easily available. And the confidentiality is manageable: make the agent join only the meetings you want it to. But if you’re running a big organisation, go all in, with proper tools to manage confidentiality. By all in, I mean: Outlook email data, Claude organisation level data, Slack chats, etc.