IT Leadership Resources

How to Prevent Your Enterprise Chatbot Initiative from Failing?

Recently Facebook officially decided on stopping all work around its AI chatbot assistant on Messenger, called M. Launched in beta form in 2015, M was offered as a free service to 10,000 people in San Francisco. This was seen coming for some time now after reports went out early last year that the company was scaling back M which had hit 70% failure rate.

This is how Wired reported the shutdown of M: ‘Facebook’s Virtual Assistant M is Dead. So are chatbots’.

Though the headline may sound too ‘bold’ and ‘clickbait’, it is worth at least sparking a debate, considering the current hoard to get onboard the chatbot wave. It will not be an overstatement if I tell you that among the CIOs I have interviewed in the past few months most had either already set up a chatbot or were in the process of doing so. Undoubtedly, off late organizations are increasingly relying upon AI powered chatbots for serving their customers and interacting with them and even their internal employees.

However, the unfortunate fate of M leaves us with the moot question – how to ensure that the bubble doesn’t burst when the current euphoria around chatbots dies down.

The answer could well lie in the problem itself. According to the Wired report, M’s core problem was that Facebook put no bounds on what M could be asked to do. In other words, Facebook expected its chatbot to do virtually everything, which was a mistake.

According to a Forrester blog ‘Chatbots are transforming Marketing’, most chatbots fail because companies don’t clearly define their purpose and set goals that are far too ambitious.

Here is the relevant excerpt from the blog on why the chatbot initiatives fail.

  • Don’t clearly define their purpose. Companies tend to set a scope for their chatbots that is broad and generic, or they fail to clearly define their chatbot’s purpose and communicate it to users. A more focused approach helps meet customer expectations.
  • Set goals that are far too ambitious. Despite their promising future, chatbots are at a very early stage of development. Today’s successful chatbots are driven more by keywords than by machine learning. In the future, advances in AI will move chatbots’ potential from “question and answer” to “human-like.”

Thus, the basic principle is to keep the chatbot focused and grounded, something which Facebook failed to do with M. For my CIO friends that are either contemplating or are in the process of setting up chatbots this learning can come quite handy.

(Image Courtesy:

Leave a Comment

Your email address will not be published.

You may also like