Original subtitles

- Agent AI can generate design ideas,

test, and iterate towards optimization goals.

This can be in a factory setting or a car showroom.

Everything the agent AI does is to analyze,

get results, reason, and present output,

and iterate to arrive at the desired optimization.

The fastest growing use case for GenAI

is customer experience with intelligent chatbots,

but this involves human interaction to make decisions

when the chatbot makes mistakes

or when customer gives feedback

that can be applied to enhance their experience.

How can an agent AI break down this task

to enhance customer experience?

The agent AI should first engage to get customer input.

This could be a chat or a voice-based input.

It should then use tools to understand the customer response

and use reasoning to identify their complaint or feedback.

Then it should iterate based on the right response

to fix the customer's problem

and arrive at the best response,

which can be validated by a different agent

that collects customer feedback.

How exciting that we've arrived

at another example of a multi-agent AI.

I can't wait to show you how to build some real agent AI

using a hands-on exercise.