If this works, a lot of customer service agents will be out of jobs, which can certainly be unforunate. Customer service is estimated to be a ~$350B global industry that employs > 10M people.

However, it’s important to note that a lot of customer service actually relies on human-to-human contact that cannot be automated. For instance, when people go out to restaurants, they prefer being served by humans as opposed to robots for that social interaction which is a significant part of the dining experience. Same goes for receptionists, and stuff. It’s not purely functional but also — human.

At this stage, our product is experimental. It will undoubtedly make mistakes, but our hope is that it will learn from them and not make the mistake twice while reducing the overall space of potential mistakes that could occur.

On the bigger picture, this venture will prove or disprove things conclusively:

1. Whether people do not mind interacting with bots for functional purposes (e.g. getting a specific task done or answering a specific question) — our hypothesis is that the answer to this is yes.
2. Whether customer service agents that man help desks can be replaced — numerous attempts have been made to automate call centers and customer service departments. In fact, most of them are automated to varying degrees, particularly with the tooling that they use.
3. More generally, whether automation can replace humans, and what jobs are generated as a consequence of automating a prior role that was occupied by humans.