Designing thoughtful chatbots deserves as much if not more forethought than other elements of your online community experience.
A strong chatbot strategy can scale the capacity of your community management team as effectively as implementing a community ambassador or super-user program.
Super-user programs help a community manager to bring a community to life. These individuals become greeters, introducers, collaborators, and ambassadors of the vision for the online meeting place. Super-users who are active help to humanize the community.
Well-designed chatbots help the community manager to avoid spending time with repetitive or rote tasks that could be better spent on efforts to build deeper relationships with members. The chatbot, if well designed, can be more responsive, more consistent, and more thorough than the community manager. Well-deployed chatbots increase the capacity of community managers.
Chatbots simulate human conversation and can work in a variety of modes, including chat windows (as in Facebook Messenger or Slack bots), text messages, and voice (as in Google Assistant or Alexa or Siri). Some of the major enterprise community platforms like Khoros (formerly Lithium) and Salesforce (Einstein Agents) have the capacity to include chatbots. Discourse has had a chatbot called Discobot since 2017.
Although chatbots are definitely available, my personal experience has been that they are not being used to any great degree in communities so far, with the possible exception of customer service communities dedicated to product support.
And, unfortunately, most of us experience chatbots that interrupt us with questions as soon as we visit a website for the first time, before you can even get oriented to formulate a question. I don’t know about you, but I usually try to turn these rude creations off as soon as I see them.
Nevertheless, I feel like they have incredible potential. They represent a next evolution in community building, but like anything they require planning.
My personal history with chatbots
My own experiences with building chatbots was from an online game I ran bace in the early 2000’s. The developer I worked with had created a rules-based chatbot engine that worked in the game that I used to create interactive characters players could have conversations with. These little chatbots were seeded with a number of conversation starters and were primed to listen for anyone talking to them who used certain keywords as triggers to additional responses.
In this way, I could create a chatbot that seemed to be having a conversation. To prevent too much repetition, especially if someone used the same listened-for phrase a number of times, I was able to create alternative responses that were randomly selected.
Modern chatbots can still work this way although they have gotten even better at processing natural language and even learning responses. Unfortunately, I think that most of the bot experiences we know, like Siri, are still pretty terrible at understanding us. However, that’s more due to how they have been designed and not an inherent fault with the technology.
Will chatbots replace humans?
Even chatbots designed with great forethought to demonstrate empathy cannot compete with human interaction, at least not yet. They cannot interpret moods and alter their responses and, as Alexandra Radu mentions in the video interview above, they can easily and innocently offend people who are looking for help.
Therefore, chatbots are no substitute for humans, especially in online communities where the establishing trust is critical.
My partner, the chatbot
Think instead of chatbots as a partner for human support of your community. They can provide help faster, more courteously, and more consistently than a human community manager in several situations:
- Welcoming newcomers
- Recommending topics to follow based on a newly registered profile
- Answering questions about how to use the community
- Providing links to top or trending topics
- Suggesting connections with people who share common interests
- Escalating issues to a customer service representative, human resources, or another specialist
- Setting reminders
These use cases can go deeper. For instance, in welcoming a new member to the community, the chatbot could suggest several other newcomers with similar interests or suggest a pairing with a mentor.
A chatbot could recommend trending topics that are based on a given community member’s stated interests or on past browsing habits.