We had the privilege of speaking with Brian Cantor, Principal Analyst and Digital Director of Customer Contact Week, about trends in the industry including the best customer experience optimization solution. In this second of a three-part series, we discussed the do’s and don’t of employing chatbots. Take a look!
RB.AI: Any other ways that you see bots being utilized effectively or not, that you haven’t already discussed?
Brian Cantor: Say here is that the worst chatbots to me are the ones that make it obvious. And now I want to explain what I mean by that because I don’t mean they’re obviously computers. I hate that argument because I think we’re comfortable enough as a society to accept that we’re not always going to be talking to a real person. We use and trust Siri and Alexa and we’re going to use technology if it gets us what we want on our terms. So this idea that people stress over — do we tell people it’s a bot? Do we not tell them? — I don’t really think that’s a productive use of anyone’s time.
What makes it obvious is that when the business makes the intention so clear through the customer experience optimization solution, but those intentions are so clearly not customer-centric. When you see customer support chat shoehorned into experiences either because they feel like as a businesses are supposed to be using it because it’s a hot trend right now, or because the business only cares about reducing call volume, what happens is, in both cases with technology, as I mentioned, really becomes more of a static self service tool. It’s not really AI. It’s more about just kind of — whether it’s kind of preventing the customer from calling, it’s the company doing everything it can to keep you from calling. And I believe that’s the wrong attitude.
What it should be is — the right use of a bot is making it so customers who don’t want to call don’t have to call. And that’s reallywhere the value is, right? That’s when you think about delivery services or financial companies that are using AI correctly, they’re offering functionality that makes sense to offer in a self-service kind of a chatbot context. This is the goal of any customer experience optimization solution. Think of a restaurant, let’s say Saturday night, you order a pizza from the local pizzeria. If you were to call, you’re going to be yelling over the phone, they’re going to have to look up your order, there’s a loud restaurant behind them, there’s a good chance that fixing the order is only going to make it worse. You’re not going to get transparency about what’s been changed, how it impacts your bill, it’s kind of a difficult experience.
Whereas if you were to just text your seamless app or whatever app you’re using and basically say — hey, I ordered with pepperoni, but I actually want it to be with sausage, and that could change — and it says — well actually sausage is 25 cents more, so this is your new bill — that all happens right in front of you, it’s so much more valuable to happen in kind of a messaging bot context, that that’s an example of where it made sense to offer it. The customer doesn’t feel as if you are only doing it because you didn’t want to talk to them. It’s a customer experience optimization solution that knows that the specific customer doesn’t want to talk to the business, so why would the business force the customer to do so when it’s not going to be efficient for anyone?
One of my main missions as an advisor in this space is to shift the focus to areas where AI is superior rather than just acceptable. The best customer experience optimization solution doesn’t just find what tasks can be handled by chatbots, it finds out what tasks are better handled by a bot than an agent equipped with live chat agent training.
And the value aspect doesn’t just have to be about the front end either. I mean there are a lot of internal applications where there’s also a value hook, where it’s not just about cutting labor costs or introducing a fun new toy for your agents to play around with. I think of areas that would be valuable — you have things along the lines of mining interactions, using a customer experience optimization solution for analytics to gather actionable intelligence, you have providing agents with more meaningful screen pops when a customer does come in, you can automate lessons based on an agent’s performance so if you see that they go below a certain average handle time, you can kind of make sure that the system gets a lesson right in front of him, so they can fix that on the fly. You can look at things like routing customers based on personality type or expertise.
And you could also perfect agent responses, and that’s one that I really like, which is you can basically use bots to test different language out. Let’s say people chat about a cancellation policy and you find out that certain lines really make the customer happy. They resolve the situation. Whereas other lines tend to drive escalations. You can take that information and because it’s a bot, it’s all computerized, it’s keeping track of everything. You could take that information, pass that directly to the agents who utilize your customer experience optimization solution, and now they know what they should and shouldn’t say if the issue ever comes up.
The way I would compare it to is, I don’t know if you play fantasy football at all or any fantasy sport, but there are those drafts simulators, right, where the tool analyze your roster, analyze who’s left in the draft, look at things like bye weeks, you know, you don’t want to have two players on the same bye week, that sort of thing, and then tell you who you should draft. And that can be the future of agent interactions as well. It looks at who the customer is, how they’ve interacted, what they’re asking right now, and then it takes all that information and runs it against company policy, the agent’s history of handling these issues, other examples of that issue within the CRM, and based on all that, it’ll let the agent know with a degree of confidence — what say or at least where to look in the knowledge base. And if you set the rules to the confidence being high enough, if it’s in a chat or messaging context, it could even automate the response. And I think that really is an area where you’re using technology to empower agents and you’re doing it in a way that’s actually valuable for both the agent and the customer.Follow us: