I have found myself on a bit of a chatbot curiosity bent this past week after reading the news about Sprout Social’s new Bot Builder for Twitter. Of course, you know how it goes, once you see one thing about a topic, then all of a sudden you see articles and research on it everywhere without intentionally looking (which makes my job easier anyway).
What is the big deal about chatbots and virtual assistants? Is the future of the bot arriving faster than we are ready for?
Chatbots are here, but their value isn’t yet clear
“Chatbots are still nascent. They are dependent on a host of interconnected and emerging technologies, many of which rely on machine learning and require massive amounts of data. Their value is just beginning to be tested.” This what Susan Etlinger, Analyst at Altimeter says in her report, “The Conversational Business: How Chatbots Will Reshape Digital Experiences.”
You can use chatbots in pretty much every conceivable way in your business from experience and services to commerce, sales and marketing, and much more. But it’s how you use them that’s important. As Etlinger makes clear, to be successful with bots you need move beyond commands to conversations. I think that’s where the real work is happening now.
There are companies taking advantage of bots in ways that are useful for customers, ways that don’t drop human interaction completely. As Sprout Social points out, automation should augment, not replace humans. Bots can do things like perform repetitive tasks and get context before passing a customer off to a customer service agent. Social customer care is a fitting example of where bots can bring a lot of value to both a company and its customers.
The role of a bot in social customer care
Social customer care is tricky. It relies on companies having dedicated support people always paying attention to social channels. And not just special channels created for customer support because customers don’t always go to a specific channel. Take Twitter as an example.
If someone has a problem with a company, they will often express their frustrations on Twitter. Having the ability to connect directly on Twitter via Direct Message is very useful because the customer knows their message is getting across. The question is whether it’s actually responded to when there is a limited group of support people working specific hours.
Enter Botty (my name for the bot today). If Botty responds to me right away, no matter the time of day, then I may get my questions answered faster, making me a more satisfied customer. We can use Sprout Social’s Bot Builder as an example here.
Bot Builder enables non-technical people to set up a bot on the company Twitter account. Don’t create a separate account for the bot says Sprout Social, and be clear to the customer they are talking to a bot (you can create a custom Avatar for the bot). You create a welcome message that greets the customer when they enter the Direct Message window. Offer a set of pre-set replies the customer can select that are essentially common questions many ask, driving a specific auto response that should move the customer along to getting their question answered.
Sprout Social’s Bot Builder provides full transparency. Agents can see each conversation in real-time in something called a Smart Box and can step in any time they want. There’s also no loss of user history, so if you have a conversation with a bot, leave and come back, your full message history is still there.
I tried this with one of Sprout Social’s customers (@BloomsburyBooks), and the chatbot was very nice to me, giving me the information I asked for. I did leave Twitter and come back later, went into my DM with them and my history was there. But I wanted to know something else, so I asked another question. No response. I tried the same thing with @EvernoteHelps – another customer and a similar result. It works pretty nice, and it helps get answers quickly, but it’s clear there’s still some work to be done:
To be fair to Sprout Social, they aren’t trying to implement some heavy-duty AI chatbot for Twitter. They are starting with the basics to give customers a way to engage with their customers quickly. And it is a great start for those who want to leverage Twitter for social customer care.
Note: @evernotehelps did respond back to my second question after awhile, so I’ll give them an E for effort:
Going beyond the basics with your chatbot
Starting with the basics is the right approach. It allows you to test without the potential to do serious damage (like mess with privacy). Etlinger offered six rules for building a better bot as you plan your chatbot strategy:
- Focus on a specific use case and purpose (at least to start) and be clear about what that is to the customer from the beginning.
- Depending on the objective of the conversation – the interaction model may change (speech, text, keyboard support, cards) and in many cases, you may use several models at the same time.
- Some use cases may result in multiple contexts (where you are when you ask a question is one example), so you need to plan carefully.
- Interactions aren’t always one-offs where a use case is performed and done. There will be instances where you’ll need to maintain context over a period to create the right experience.
- Emotions are just as important as intelligence, so you need to figure out how to understand and deal with sentiment in your bot’s interactions.
- There is only a small opportunity for branding, but when it arises, it has great potential.
“Implementing bots requires the right technology, the right data, the right use case, the right design, and the right cultural mindset.” (Etlinger)
My take – I’ll take Botty any day
Personally, I’m okay with bots, especially for customer care. The 9-5 work day that many companies provide for support doesn’t work for me. I want answers when I want answers, and if I can go on Twitter (or Facebook, or some other network) and have a chatbot help me resolve my issue, I’m fine with that.
But I do want to know it’s Botty. Otherwise, I might get a little cranky when with him when he doesn’t understand my question or gives me a bad response. I agree with Etlinger when she says there are three key things to keep in mind when you work with bots: they must have high value and be easy to use, they must be transparent and provide full disclosure, and they must respect my data and my privacy.
Bots will not replace people, but they will enable people to spend more time on more important issues or strategies. And that’s never a bad thing.
Image credit - Robot and human dialog © Mandrixta - Fotolia.com