Ishaan

... be the roadmap owners.

They should be thinking like, uh, risk officers, like CIOs, CISOs, because when agents take over, the buffer or the room for humans to review and validate, everything becomes too thin because the velocity and the speed, the crazy speed that agents operate in, humans cannot match that.

So whenever you are in the boardroom, you are determining new business processes, or you are doing some, um- Technological transformation, I would say always, always make sure that when you deploy an AI agent to perform some reasoning, before that AI agent updates your CRM or your ERP or your, uh, or your systems o- of record, make sure you plug in an governance layer in between.

You define some thresholds, what-- You define some guardrails, you define some rules for the agent when it can take fully autonomous decisions, when it can pull in a human reviewer to validate, verify its output, and what items are critically blocked for the agents to operate on.

I think those three categories you should be defining before moving forward.

And then a-what I typically do, Jay, is I usually, um, have a risk register.

Could you talk to me first, what is the Token Optimizer's Manifesto and where, where has that born out of?