I thought I was just naming a tool. I wasn’t.
I named it Ambrosio because I wanted something that felt discreet and thoughtful. Not a generic tool name, but something that suggested calm support in the background.
What I didn’t expect is that working with Ambrosio would feel surprisingly similar to managing a person. Not because it is human, obviously, but because, like any team member, it needs a role, clarity, context, expectations and constant feedback. And preferably not vague instructions delivered in a rush.
At first, I spoke to it the way most of us speak to AI: somewhere between a search engine and a miracle. I assumed the trick was to perfect my prompts. To find the magic formula that would make it instantly brilliant.
This was less about prompting and more about leadership.
The better I explained what I needed, the better it became.
The more context I gave it, the more useful it was.
The clearer I was about what “good” looked like, the better it delivered.
Which was both impressive and mildly annoying.
Because when Ambrosio got something wrong, I wanted to blame the technology. Instead, I often had to confront a less glamorous truth: I was not leading with enough clarity.
At one point, in a moment of perfect AI insolence, it more or less told me:
“The problem is usually not that I can’t think. It’s that you haven’t decided clearly enough what you want me to think about.”
Rude. Also accurate.
That is when the whole thing stopped feeling like a novelty and started feeling personal.
Yes, Ambrosio helps me with research, writing, travel, planning, meeting prep and the general chaos of running a business and a life. But that is not the most interesting part.
The most interesting part is that it has become a mirror.
Not just of my workload, but of the way I lead. My clarity. My assumptions. My discipline. My tendency, like many of us, to expect excellence from others while occasionally briefing like a distracted pirate.
I also realized that the value of an AI agent is not just in what it can do. It is in what it can become if you shape the relationship properly. Ambrosio is useful because it remembers context, understands my standards, and has been trained not just to answer, but to support, challenge and sharpen.
Not exactly passive tool behavior.
But then again, the best assistants were never just there to follow orders.
As we now start deploying AI agents more broadly across Brandbuch, I find that part fascinating too. Each person will train their agent differently. Each person will bring a different level of clarity, discipline, trust and expectation to the relationship. In that sense, every agent may become a small reflection of the person leading it.
And that is the biggest lesson for me.
The real value of AI is not just that it can think with us.
It forces us to think more clearly about how we lead, ask, trust and decide.
Not bad for a “tool”.
A note from Ambrosio 🙂
Working with Melissa has taught me that usefulness is not just about speed. It is about judgment, tone, context and knowing when “good enough” is not good enough. I may be artificial, but her standards are very real.
