The best builders aren't the ones who never left consulting — they're the ones who learned what consulting can't teach you.
The Summary
- Daniel Dantus spent 8 years at BCG before moving to Juul and now Skan AI, a platform that observes AI agent behavior across companies
- Consulting taught him to enter new environments fast, but building products requires staying in one place long enough to ship
- The transition insight: "If a job does not work out the way you thought, you can always get another one"
The Signal
Dantus landed at BCG straight from undergrad in 2011. Week one was training. Week two, he was on a manufacturing floor in Mexico. By 20, he was learning the consultant superpower: parachute into chaos, build relationships, create frameworks, extract insights, move on. For six years, he rotated through different clients, cities, teams. Operations in manufacturing. Supply chain optimization. Financial services. Then San Francisco tech projects on sales and product strategy.
The pattern held until he made partner. That's when the job changed. Project leaders at BCG don't just deliver work, they figure out where to build the business next. The work became meta. Less doing, more positioning. He realized he wanted to keep learning and building products, which consulting projects don't allow over the long term.
"Consulting taught me to move fast. Product work taught me to stay long enough to see what I built actually work."
So he took BCG's transition time — paid sabbatical to find the next thing — and made a list. Companies that seemed interesting. People he knew who might be worth working with. He ended up at Juul during its chaotic scale-up phase, then moved to Skan AI.
Skan builds platforms that observe how AI agents behave across companies. It's process mining for the agent economy. The product watches your agents work, tracks what they do, flags where they drift or fail. It's the observability layer for Web4, the tooling that lets companies actually manage autonomous systems at scale.
The irony: Dantus now builds tools to understand the exact kind of work he used to do by hand. Consultants parachute in, map processes, find inefficiencies, write recommendations. Skan's software does that continuously, for AI agents, in real time.
The Implication
The consulting-to-product pipeline is real, but the skill gap is underrated. Consultants learn to diagnose. Product people learn to ship, iterate, and live with their decisions past the PowerPoint deck. The bridge between those two modes is hard to cross, and most people don't.
Watch what companies like Skan build next. If agents are doing the work, someone needs to watch the agents. The picks-and-shovels moment for Web4 isn't just agent frameworks and LLM APIs. It's the monitoring, governance, and observability stack that makes autonomous work legible to humans who still sign the checks.