It’ll be wise on our part to enforce controlled autonomy. The intention is not to keep AI from taking action, but to make sure the level of autonomy matches the risk of the action. For example answering FAQs can be fully automated but when it comes to issuing a large refund, changing sensitive account information or making a high impact… Read more

The actual concern here is more about AI systems demonstrating unexpected capabilities or interacting with systems outside their intended testing environment hence the problem of AI alignment. As AI becomes more capable, can we make sure it stays within the boundaries humans intended it to?, and do we have good enough safety rules when it doesn’t?

SaaS is entering a whole new phase and only companies who subscribe to this phase of change would keep their customers.

I’m curious to know, do you think businesses are changing the way AI operates too?

A world where both human agents and AI work together to resolve a customer issue and improve AI customer experience is definitely the better option.

One very overlooked cost of poor AI customer service is trust and loyalty. A customer who is satisfied with a service has a high tendency of not returning because they expect same scenario from before when their issue wasn’t resolved.

The simple answer to this is that an AI agent actually resolves issues with or without the supervision of a human agent, while a chatbot answers basic FAQs

Don’t launch an AI support agent just because it looks impressive in a demo. These tests are important to determine whether the AI is ready to solve real world problems, test the consistency of the AI’s answers across different customers, test how many clients it can handle at once and test its knowledge base. A good AI customer support…

Read more