There’s a new buzzword in IT this year – chatbots, aka bots. Of course the term isn’t that new or unusual – IT terms seem to appear and become trendy almost as fast as Pokémon characters. And a “bot” isn’t just a misspelling of bought, boot or boat. In Internet-speak, chatbots are conversational interfaces for messaging platforms.
Other names include virtual agents, virtual assistants, messenger bots, voice-controlled assistance, and no doubt others.
Chatbots must be important – I discovered there’s a magazine dedicated to them. The beginners guide, which provides examples of why chatbots are emerging as a new “ecosystem,” is worth reading. There are also various informative websites, such as chatbots.org which claims to list all existing bots.
What is a chatbot?
A chatbot, according to Wikipedia, is “a computer program that converses in natural language.”
The newest twist on the technology, however, is to combine a chatbot with a text messaging platform to form a new business to consumer (B2C) channel. If text messaging has become a primary user interface for smartphone users, then why not adapt it for commerce as well?
At least this is what I understand to be the perceived new opportunity.
Without a doubt, the definition and scope for chatbots will evolve over time as the technology moves forward. Perhaps chatbots could even be viewed as “human APIs.”
Chatbots are not really that new – for example, the telephone IVR (interactive voice response) systems that we all love to hate can interact with people in a limited way.
A strawman for a chatbot maturity model could be:
Level 1 – Initial – Basic text and/or voice response; little or no conversation (e.g., Telephone IVR);
Level 2 – Repeatable – Voice and/or text recognition and response; rules-based interaction (e.g., Google Voice Search, Siri);
Level 3 – Defined – Basic topic-oriented conversations; may include multimedia inputs and responses; basic intelligence and agility;
Level 4 – Managed – AI-based natural language conversations; multimedia responses; context and basic learning including access to related information;
Level 5 – Optimizing – Omnichannel AI-based conversational interactions with learning.
One example of an advanced chatbot system may be VIV, although its capabilities are yet to be seen.
While text messaging will not be the only delivery channel for chatbot applications, it could be an important area of early development.
A chatbot architecture and ecosystem needs to be defined!
What are chatbots for?
Perhaps an even more challenging question is not what they are but what they can do.
The promise is that chatbots will be able to substitute for human interaction in a wide variety of situations, from games to purchases to simple social conversations.
Here are some examples, taken from the beginner’s guide referenced earlier:
- Weather bot – ask for the weather using text messages;
- News bot – ask CNN to provide news stories;
- Personal finance bot – help manage money better;
- Scheduling bot – get a meeting with someone; and
- A bot that’s your friend. – Xiaoice, built by Microsoft, that over 20 million people talk to in China.
Facebook recently provided developer interfaces to use Messenger for 3rd party chatbots. They claim there are over 11,000 bots have become available in less than 6 months.
So the answer to the question is that the sky is pretty much the limit and chatbot innovation reigns supreme at the moment.
The underlying caveat, of course, is that different messaging platforms don’t (and probably won’t) be interpretable across providers.
Where do chatbots fit?
I think we’re just seeing the beginning of an era in which chatbots become an important part of the consumer computing ecosystem.
Some use cases would include:
- Notifications: you could receive many different forms of notification from many different applications that include messaging interfaces – your home bot could advise you if your house is burning down or your health bot could provide medical alerts;
- Permissions: systems could ask permission for things such as actions to be taken, events to be tracked, people to be advised, etc.;
- Shopping: people could talk to purchasing applications with text messages or human voices that simulate a store experience;
- Control: the user could send instructions via a chatbot with assurance that the chatbot will track progress;
- Social: chatbots could become chatterboxes for general companionship; and
- Assistance: the chatbot could serve as an assistant (e.g., as a calendar manager) or advisor.
The million-dollar question is, what would you pay for chatbot services?
What I think is that I’m not sure what I think about paying for chatbots, especially if they are coupled with advertising. Where will chatbots fit into your life?
