One of Monaco’s demand generation agents is sustaining 20-40% response rates on conversations that went cold and turning them back into real pipeline. Last week we made Monaco generally available. I wrote that Monaco represents a different way to think about revenue software: not just software that stores the work, but a system that can learn how sales organizations actually operate. We’re already seeing that work in production. The visible part is the agent. The harder part is everything underneath it. There are really two parts to making that work. Part 1 is the system itself. The important context has to be captured when the work happens, not reconstructed every time an agent runs. Who is this person? What relationship already exists? What did they say? What did we do? What happened afterward? If that history is durable, the next agent inherits what the company already knows instead of starting from zero. When it acts, the outcome gets written back. That’s how memory becomes intelligence. Over time, replies, meetings, stage movements, wins and losses become evidence about which signals matter, which messages work, and which actions actually change an outcome. The re-engagement agent is one example. The same underlying system powers agents across prospecting, outreach, follow-up and pipeline execution. Part 2 is a different kind of org design. Some of the most valuable judgment still lives in the heads of the best operators. A while ago, we deliberately moved one of our top sales reps into the engineering team for exactly this reason. His job is to turn what strong reps actually do into product logic, evals, cases, test loops, decision rubrics and outcome loops. I think this is fundamental to vertical AI. If you’re automating expert work, the people building the product can’t stay several abstractions removed from the people who are actually great at doing that work. You need the system learning from production and outcomes, while operators continuously encode what good judgment looks like into the product. That’s the broader idea behind Monaco Not a collection of independent agents, but a System of Intelligence for Revenue where every agent operates on the same organizational memory and contributes back to the same learning loop. I wrote a longer piece on the architecture and org design underneath this. Link in comments.