How do you think about context when really designing these agentic experiences?

Having as much as data from the customer's end and an understanding of the intent is super critical for an agentic system to succeed.

If an agent knows what the customer's profile is about, what they have been calling or messaging about, their sentiments in the past, what kind of frictions they have had, and if that is provided as a proper context.

which also has been evaluated in terms of different workflows that the agent tech solution is going to handle.

It provides the agent and capability where it can take right decisions depending on if there is a need to escalate to a human agent or if there is a need to solve that problem then and there.

As an example, if I have to tell you, if a customer is patient enough with most of their Jira tickets that had been opened with the company, that means the agent tech system has much more control over solving those problems for the customers rather than immediately bringing in the human agent.

But let's say as an example, if a customer had been provided a very different billing for a certain SKU from the company's end, and if there is a new product that has been launched, now the agentic solution has a proper context that's available from that customer end.

Why do so many AI initiatives look great in demo but really struggle to become products that customers consistently use and trust?