Who authorized that?
AI agents can now spend, send and delegate. Robots can now move, lift and open doors. Both act on behalf of people who cannot watch every step, and neither follows instructions precisely: they interpret in the moment. OTIS Labs runs working sessions on the questions that follow. How do you give an agent or a robot permission, how do you know it is acting within that permission, and who answers when it is not?
What did our AI Agent do?! The missing runtime authorization layer for agents and physical AI
A working session on a single set of questions that turns out to be the same for software agents and for machines. We start with why the problem is shared, why identity, security and guardrails each answer a neighboring question, and what a real answer looks like at the moment an agent or a robot does something that matters.
You will leave with a clearer understanding of why an authorization layer is a new foundational layer in the human-AI infrastructure stack, and what has to be in place before you would let an agent near money, records or a patient, or a robot near a customer.
Hosted by Marina Piller, founder of OTIS Labs. The Zoom link is sent to approved guests closer to the date.
- Founders and CTOs putting agents in front of customers in finance, healthcare, legal or insurance
- The people inside enterprises whose job is getting agents into production this year
- Robotics teams moving from demonstration to deployment
- Investors and infrastructure builders following the space
Runtime authorization for agents and physical AI, virtual
The same session, virtual, with more room for the robotics side of the question. Details are being confirmed.
RSVP hereGet session invitations
Event invitations, research releases and product notes from OTIS Labs.