A common error in chatbot building is a twist on the "build it, and they will come" theme. The idea being that chatbots are the future, so if we build one, then the future will come to meet us. Chatbots have strengths and weaknesses like any other application of technology.
Each generation of chatbot technology eliminates more of the weaknesses and boosts the strengths. Hence, a realistic approach to choosing a chatbot strategy is to assess the advantages you are hoping to gain with a chatbot is to look at whether the straights out-way the weaknesses:
SCALE: The chatbot can scale and deal with hundreds or thousands of users.
CONSISTENCY: The chatbot will provide the same, most up to date, answer every time. No variation, no spin. And provided the chatbot is up to speed, it will apply the policy, not' workarounds.'
ANALYZE AND ITERATE: the conversations the feedback and the interactions can be analyzed without the need for audio transcriptions and at scale from feedback customer feedback. Therefore, Chatbots are quicker and easier to iterate change and have that change applied consistently (see above).
SIMPLE: Regardless of the advancements in technology, the chatbot will do what out has been told to do. It is not learning like a new employee, nor will it sympathize with real-world scenarios (e.g., my credit card was eaten by my lawnmower = I have lost/damaged my credit card)
BEST GUESS: It will only be making best guesses at what someone has said, therefore it should handle the most common cases well, but the niche cases are still a struggle.
NEEDS TOOLS: If you can't enable the chatbot with the tools that can actually fix user problems (e.g., to change passwords, transfer funds, or raise a problem ticket), then the chatbot is impotent and will soon become a frustration.
Chatbot strategy is all about picking the use cases where chatbots' strengths apply.