A CIO with hundreds of bots already running doesn't want to hear "rip and replace." So what do you tell them about AI agents? I put that question to Vikram Kakumani, Deputy CTO at UiPath, at FUSION 2026 on The Ravit Show.
This one got deep into architecture:
- How to take an RPA estate with hundreds of bots into agentic automation
- What "controlled agency" really means, and where humans should stay in the loop
- The biggest blocker keeping AI agents out of production
- Orchestrating agents, robots and people together
- The architecture mistake companies keep making with agents
- How the Map of Work gives agents the enterprise context they're missing
- Where Maestro takes orchestration next
The part that stuck with me was controlled agency. Everyone wants agents that act on their own. Very few teams have decided where a person still needs to sign off. Vikram laid out a practical way to draw that line.
Where would you keep a human in the loop for your agents?
Podden och tillhörande omslagsbild på den här sidan tillhör
Ravit Jain. Innehållet i podden är skapat av Ravit Jain och inte av,
eller tillsammans med, Poddtoppen.
