Take retailers, for example. The customer's intentions are pretty common. Order tracking, canceling an order, returns, exchanges, checking if a product is in stock, store location, payment options, you name it. Do these things sound familiar to you? These sound very familiar to customer service agents.
If there's one thing that we know after powering some of the world's largest contact centers with the ability to communicate with customers via messaging, it's this: customer service agents are always in danger of the psychological equivalent of repetitive stress syndrome.
They are doing many repetitive tasks every single day. Conversocial was really built on the premise that automation and AI are great for handling repetitive tasks and doing things that human agents really shouldn't have to do. The time and focus of live agents should really be reserved for more complex tasks where things like empathy and flexibility are really paramount to having a positive outcome with that customer.
So, when it comes to text-based channels, like messaging, of all the things that agents do day-in and day-out, replying is the most time-consuming. It's the thing they do all the time.
So we launched this feature called Autocomplete. And this is really an AI-driven feature that analyzes agents’ responses across the team, or even an entire company, and it looks for patterns around certain words or phrases.
Check this out: In a seven day period, Autocomplete provided agents with more than 40,000 suggestions that they accepted while they were typing, saving an estimated 48 hours of agent time.
We also have another feature called Clipboard. So, think about this as a mini knowledge base that allows agents to pull really common phrases and common policies or language they use often. Templates, if you will. They can search for them and drop them in and send them to someone on Messenger or WhatsApp or whatever channel they're using.
In a given day, over 18,000 searches happen within Clipboard to find content that agents want to send to a customer. This one thing, every day, saves an average of 37,000 minutes or over 600 hours of time saved over the course of a month. Think about that. That one thing that an agent does for maybe five seconds and 600 hours of time is saved. And that's just one thing that they're doing out of many.
But what if bots could do that for them? What if you were able to give your agents in the contact center an assistant. Executives have executive assistants. Why can't there just be an agent assistant?
What if your agents could say, “I'm not going to do this. I'm going to let the bot go do it." And they could dispatch a bot to go collect the information that they need or help that customer with something that's very common while they focus on other customers that have more complex issues.
Well, that's exactly what we recently launched and we call it Agent Assist: It brings bots and agents together on the same platform and allows the two to work hand-in-hand, giving customer service reps a personal assistant to make their job a little bit easier.
"Agent Assist is opening up new opportunities by saying “maybe there's something that the bot can do that the agent actually just cannot do, or maybe there's something that the bot can do more efficiently that the agent can't do that efficiently.” says Anish Bhatt, VP of Product at Conversocial. "The whole game for us is to figure out which entity is the real specialist. Because bots are specialists too, in some mechanisms, and agents are too when it comes to sentiment and emotion."
If you were just a bot company, you would say, "Oh no, we can do absolutely everything." But the truth is you can't. You can't really figure out the emotion. You can't figure out the sentiment.
And if you're an agent company, you would say, "No, no, no. You can't do anything more efficiently than us because we really understand everything. We're human, we have more dynamics or more variables in our head so we can make a better prediction of the right responses."
And the answer is: both are right, but both are wrong. But because we own both properties, we can uncover what's the true specialty of each entity.
So, for instance, imagine sending a payment request. Now, it's very easy for a bot to be able to do it because it will hit the APIs that will pull that item or service, and then they'll create a payment request and then they'll present it to the user.
For an agent to do that, they have to find the item or service manually, create the payment request and send it. And that is something that a bot can do very easily in less than a second. So why would we ever have an agent do it?
Instead, with Agent Assist, the agent can trigger a bot to handle the payment request. As soon as the bot does it and the user responds, it routes the conversation back to the agent. The agent can say, "Can I help you with anything else?"
Before it always felt like bot versus human. And then, over time, it was bots handing conversations to humans. But now they're on the same team. Bots and humans working hand in hand to resolve a single case. That's a little bit different than any other company has ever done.
"What if you were able to give your agents in the contact center an assistant? Executives have executive assistants. Why can't there be agent assistants?"