Author

by Marta Pereira
on January 26, 2017

Our obsessive curiosity with bots

 

seebotschat

Earlier this month over 3.5 million people around the world stopped what they were doing to see two bots chat. They headed to live streaming platform Twitch where Vladimir and Estragon, two modified Google Home speakers running the Cleverbot AI software, were exchanging mesmerizing yet somehow disappointing arguments about life, god’s existence and their appreciation of waffles.

Why do we have such a morbid curiosity about AI? Is it because of sci-fi? It has been raising our expectations about AI capabilities for decades – even if often with dystopian outcomes. Because we expect so much, it’s always ironic when we find out that our bots have the same flaws as their creators. They can get racist (like Tay the Twitter bot) or engage in petty arguments.

 

Getting familiar with AI

 

No matter how flawed and funny these bot experiments are, they – along with the generalization of AI personal assistants such as Siri, Alexa and Cortana in our devices – tell us that AI is now much closer to us, the common consumer. And it’s evolving fast: just last week the MIT reported that Google has AI software that knows how to create more AI software – truly impressive (and very meta) stuff.

As consumers, we are starting to get familiar with both capabilities and limitations of AI, and we adjust our expectations according to that. What does this mean for companies and brands?

 

My customer service is a bot

 

optimized3

Some believe that bots will completely replace humans for customer support activities. We don’t know what the future holds, but judging from Vladimir and Estragon, doing that right now would be a tremendous risk. However, we cannot stay indifferent to these two facts:

  • the rise of mobile brought about the always-on consumer who expects immediate responsiveness from brands
  • the boom in messaging apps and these latest advances in artificial intelligence have enabled a new generation of chat bots which can help communicate with customers at a reduced cost

Chat bots seem to be a natural response to this always-on consumer: they are always available and the best among them can remember interactions and customize our experience (a good example is the Spring bot on Facebook Messenger). Bots allow customers to move seamlessly from point A to point B on their journey. They make it possible for companies to answer more requests for information than did ever before using traditional customer service.

The best approach, we tend to think, is one where you use AI as the starting point of a conversation with a customer. Your chat bot should be your conversational front-end, your first-line support agent. It should be able to solve simple queries on its own – but it must also be able to understand in which situations it’s better to ask for some friendly human help. This could mean passing the conversation onto a real, human chat agent, to ask if you want to get a call from someone in the company or to send an email to a technical support on the client’s behalf.

This way your chat bot will work as a “lead qualifier”. Bots aren’t great at selling stuff slightly more complex than a pair of boots just yet – but they can make sure that your sales team is not wasting time with irrelevant leads. At the same time, if you have a customer who just wants to find out where to download an invoice, he won’t need to wait in an endless contact center queue. It truly is a win-win situation.

 

BySide B3 – your handy, always-on problem solver

 

B3 Bot
Because we think this is the right approach, we built a chat bot that is able to do all of this.

The reality is that most chat bots you see today don’t offer much more than a sophisticated way to access FAQ content. This is not what we are aiming for. B3 uses Natural Language Understanding (NLU) and machine learning, meaning you can ask it whatever you want.

B3 analyzes the context and understands the intent of the conversation. Whenever the prospect is seen as a qualified lead, B3 can redirect the conversation to a real sales representative either by switching chat agents or by generating a phone call. This can have a great impact on the contact center operation: less non-commercial calls, more qualified prospects.

The bot’s knowledge base is created by automatically importing content from existing digital assets such as your website, FAQs, customer knowledge bases, and any external knowledge sources. Also, companies can grow the knowledge base and train the bot by asking questions and correcting its answers in the application’s backoffice.

Right now, besides having B3 on your website, you can integrate it in Facebook Messenger and Skype. We will soon have it running on Slack as well.

Give us a shout if you want to know more!

 

About us

Want to know more about BySide?


Check our website here