Nowadays I feel every business function is in someway or the other using AI Agents, but how can one be 100% sure that the AI agents is working properly, giving correct, grounded, reliable answers, not drifting from what it is asked to do. How can one evaluate an agents performance
I work in BFSI industry in NA. After a lots of discussions, back and forth we are finally looking out for vendors in the Agentic AI space. Our use cases are pretty straightforward currently but eventually we want to expand the tech to our complex use-cases too. For example currently we want an agent to help us with data handling, summarization and all the other pretty basic use cases.
As an enterprise just getting started with Agentic AI tech, what all should be pay attention to in a vendor. We have shortlisted quite a few vendors, but wont be revealing them,
Which capabilities are non negotiable in an AI platform? Governance? Evaluation? What all?
Soo much buzz around Harness Engineering, AI agent Harness - Does this mean that the industry is moving away from the basic idea of Models doing everything or maybe let me put it this way - LLMs deciding what my agents are going to answer? I am genuinely interested in knowing more about this.
Soo much buzz around Harness Engineering, AI agent Harness - Does this mean that the industry is moving away from the basic idea of Models doing everything or maybe let me put it this way - LLMs deciding what my agents are going to answer? I am genuinely interested in knowing more about this.