[Question] An increasing number of businesses are tapping into self-service. The initial hype on intelligence assistance is over. What’s your advice to businesses who are looking into adapting self-service in their organization?
[Answer] I think we’ve got to the point now where the initial excitement about Virtual Assistants is over and people are in the test and learn mode. So let’s talk about some of things people should be thinking about.
I think one of the first things to say is that it’s the kind of technology that you should approach and try not to aim for perfect. What you should do is take a relatively simple area and then allow the customers to drive what your roadmap is going to be and how you iterate and expand the product. It’s a bit like a letterbox in the sense that the customer is going to post you their questions and let you know what’s important to them. Some of that you will be able to handle in terms of the intents and the knowledge that you use to support your answers, and some you won’t. But nonetheless, if you build a team ready to catch that and respond in an agile way, you can quickly build out something that actually works for the customer.
The second thing to consider is conversational design. At the end of the day, it’s a technology but it also requires a psychology. With a conversational design, you’re going to be able to increase the % of customers who stay within that particular channel and get the outcome that they want.
The other thing is to think strategically in a continuous way about the next use case, the next opportunities. I personally think we’re going to end up in a world where no more than 25% of inbound contact needs to be dealt with by a human. In theory, the rest of that is subject to be an opportunity for self-serve. And if you think in those terms, we’ve got an enormous set of opportunities to do that. So, keeping an up to date way of identifying the use cases, doing test and learn, and quickly establishing what the potential ROI is, is another key thing to do. In other words, try to get there as quickly as possible because as soon as you’re doing that, the cost model is incredibly transformed when you’re driving it through an automated process of helping people.
So in summary, do it quickly, do it in a controlled way, don’t be too worried about perfection, but iterate and learn how to do it as fast as you can do. You will quickly see adoption when you’ve managed to nail the right use cases.
[Question] Messaging as a service channel is really popular now. What is the reality? Should businesses consider a messaging centric contact strategy?
[Answer] Well I think the first point to be made if you look at it from a customer’s point of view, clearly messaging is massively popular. All of us get involved with messaging through one form or another. But if you look at the commercial market, how many of us have adopted that as a channel to engage with customers? I think it depends on where you are in the world. In the UK for example, we have typically been very slow to respond to that. Some of us think, “webchat will be okay, what really distinguishes messaging?”.
Well, there are several elements here. One is the asynchronous nature, which is attractive to the customer. They can just park the problem and wait until someone comes back with an answer. It’s also persistent, it doesn’t end at the end of a session, which means it can generate an interaction history.
So, I think there’s a compelling reason to consider it, but you might have to consider the security of certain social platforms that offer messaging.
You can of course have your own version of messaging in-app, but I think in terms of the challenges of that, you’ve got to be thinking about what that will do to your workforce management calculation, your strategy etc.
And I think one of the learnings people are beginning to understand is that it’s probably necessary to run a 24/7 model if you’re going to be running asynchronous, because how do you tidy that process up at the end of a shift if you’re going to shut down at say 7, 8 or 9 o’clock? You probably therefore need to have a combined strategy of both chatbot automation plus live as far as messaging is concerned. If you blend those two things together and run that on a 24/7 basis, you’ve probably got an answer.
[Question] According to a Gartner report, only 9% of customers get their issues completely solved via self-service. Is that a problem? What can businesses with self-service do to increase this number?
[Answer] Well, I think there are several things in the middle of that, that need to be unpacked.
First of all, the customer often still needs to escalate to live because they weren’t able to complete what they wanted to do with the channel. Some of that will be answered through better integration with legacy systems and better integration into knowledge management to ensure the answer we give through self-service is rounder and more complete, and therefore more satisfactory. For example, you might be a fashion retailer and one of the key questions you get is, “where’s my order?”. If you can’t answer that, then I’m still going to get an escalation to live, but if you plug that into third party real time delivery systems which know exactly when the parcel is going to be delivered, well then you’ve dealt with the problem.
The other thing is, I don’t really believe that the future belongs to either to live contact or to self-serve, I think the future belongs to blended. And I think the way that we’re heading is that within a set journey, you’re probably going to be moving between self-service and live. So the statistic in that sense is slightly misleading.
I think there will be a certain number of journeys which can be done entirely without any human intervention, but even something like mobile banking, which is well established, if I see that my account is down by £5000, I’m not sure I want self-service at that point.
So, I think you’ve got to always be prepared to have escalation, but the issue is, is the escalation because of an immature self-service routine or simply because that type of customer in that situation wanted access to a human? I think more often than not, we are intuitively, as customers, going to want to flip flop between those 2 worlds.
So, I think a better question is, how many blended journeys are completed vs. those that are 100% delivered by a human.
Read more about Martin Hill-Wilson here