Make your Automation Case with Experience Data
Are you planning an Automation / AI initiative or thinking about Robotic Process Automation (RPA)? Listen to this episode to learn how employee experience data can help you find the right focus and prove the value it will bring to the business.
Pasi Nikkanen: Today we talk about automation, and I think the theme of the episode is employee experience in automation, how you can use these both together to actually justify your things. We will look at some examples, and we have picked up some customer data as well.
But let's start, Sami, if you first explain a little bit about the background, what we have learned from the Happiness Score™ based on the hundreds of thousands of feedbacks that we've been gathering. Why do you say that that already kind of justifies? Which reasons justify automation?
Sami Kallio: Many people are asking us that, when they are automating, the first thing that this is against, may be a risk to employee experience. But, if you think about it, in our measurement, in whatever we measure, in IT Incidents or Requests for example, the biggest so called factors or reason why I was happy, was speed of service. Or if I was unhappy, again, the speed of service or on the other side slow service.
75% of people who rate the services nine and ten are saying that the biggest reason for the happiness is speed of service. And now if you think about it, that's exactly what we can help with automation. And we can not only do it 8-to-5, we can do it 24/7, and that is also the growing thing with the Milleniums and so on. The other factor we measure, which is also a very very big thing in thinking about the negative part, is a factor called, "I had to explain my case several times."
If we are able to ask the right questions from the beginning and give the answers based on that, or help them based on that, you don't have to do that. And these are the two factors that we really believe that automation is a way to get people happier, if you do it right. And that is what we try to discuss today.
Pasi Nikkanen: Yeah. All right then, probably the next question is how to automate? And based on this data. So, I gathered a couple of examples with something that I've been using in webinars and some customers have been telling us when meeting them. So, one good example. We are always in IT maybe looking at things that are bad, but actually, when you think about automation, you might want to look at services where people are too happy. You have huge volumes. And they are only saying that "Maybe I only wasted ten minutes." But in reality, we have one example where there was a password reset, people were almost giving them +100, so almost full NPS score. But then they were saying that "I wasted ten minutes." And looking at the reasons saying that "I actually queued on the telephone for ten minutes."
So, if you would actually do this as a Self-Service, you would maybe reduce your overall happiness on IT. But on the other hand, you would free up the whole expensive channel of telephone for the important things. And also, all the manual things, even that few minutes that the agent needs to actually do the reset. So, looking at the huge volumes where people are really happy, is one good use case.
Sami Kallio: This password reset, I have to tell a story now. Now you said as well, that lose it from the overall score, but of course you can also measure automated services and that is a very interesting area where we are starting to do things to also get to understand that really, automation is not against good experience if we do it right, again.
Pasi Nikkanen: That's true.
Another thing could be for example, the use of chat bots, and maybe like, you know, you have Self-Services, you have knowledge articles. Earlier it was more like you feel that knowledge article, it takes a lot of time, you create, and people don't find them. But now actually when you are start introducing ChatBot, then the ChatBot can introduce actually help the user to find the Self-Service tool, or the self-healing tool. And that right knowledge article. So in a way, automating that first line, repetitive work of pointing people to the right places, is something that is just ridiculous, I think.
Sami Kallio: Do you think that there is certain kind of services that work better in ChatBot world or ...?
Pasi Nikkanen: Yeah yeah. Definitely. I mean, I would say that things where you actually ask things, you might not have a problem as such.
But it's something that I ask, like, "How do I do this? Where do I find that?" Approval type of things. "Who needs to approve this, or how do I get this approved?" Or something like that.
And also one example was that, how to use employee experience data to do automation? Or automate the priority of the tickets. One customer, Reckitt Benckiser or RB, they actually found that during the end of the month, there are certain services where people will get bad experience unless you do it really quickly. On the other time of the month, those are not high priority tickets. So, I think the high priority thing in a way that, in the right time for the correct service, that actually is the high priority. So that people get their work done, it's not like it should be based on a VIP person, or just say that this service is always the high priority.
So, I think that was clever use of automation. Basically, it's just automating the priority, not how they get solved or anything like that, but just making sure the agents actually automatically get the right tickets with the right priority.
Sami Kallio: Also, one big similar thing that I've been dreaming for now about a year, that I will want to find a customer that will do is to really help to use the automated feedback data, to add skills to agents.
Meaning that when you have an agent who has 1000 feedbacks from the last six months from end users, you can say where this agent is good at. What services he's good to solve. So why not to use that in your IT Service Management tool, to direct tickets to correct, right expert? So, automating a skill adding fragment also.
Pasi Nikkanen: Yeah, and then maybe on that kind of testing thing, having that person to actually teach the robot how to handle this case, because they are the expert on that. So, "Can you please teach this ChatBot how to answer all these things, and what are usually the culprits?" So, because then they won't need to do it every day.
Sami Kallio: Because for him or her that is a repeating job.
If they would after that, then our system would know that this person is the best in mobile devices, so they would be getting all the mobile device tickets, and that would start to be quite annoying from a learning perspective.
Pasi Nikkanen: Another topic which I think automation is even more relevant, is when we are talking about enterprise service management. So, stuff outside of IT, because automating the IT for IT is not really what people spend that much time on, their employees work on other stuff. But if you think about, like, we were discussing that employees ask things like, "How do I do this? How many of that do I have?" Then I think from there we find good examples.
And I think you just learnt something from Queen Mary University. Can you elaborate?
Sami Kallio: Yeah, I was in Leeds for an event. It was a lot of universities and Hendrick from Queen Mary University was talking about how they did an automation using Alexa. And the main learning from that it was started by interviewing students to find the real things they wanted to ask, where they needed help. Not that they would start to interview their own people and from their own perspective what Alexa should be answering.
So, Alexa is kind of a ChatBot here, not written but the other way. And they got really great results, and I think the learning here is whatever kind of automated service robot you are going to implement, start to understand the questions and problems more from the end users side before doing anything else.
If you do it kind of from an IT, HR, finance or your own perspective, you might formulate the questions with the language that you are using, but not the language that the end user would be using. So ... that was really interesting.
Pasi Nikkanen: Yeah. And that sort of looks at good way that you actually ... because sometimes you're wrong with the new technology, not figuring what you are going to use it for, but they kind of turn it around, so "Okay, what are their actual problems that we are solving?" From the users.
And when you talk with people from HR, and finance, for example, they are not that IT kind of focused, so they actually need to understand why we are doing this. How will they justify this cost of automating something, that their managers and so on. So, I actually went into our database, I looked at a few customers, people who are using finance services, measuring the employer experience on those. And specifically, I had a feeling that what if we look, like travel expenses? Something that I spend sometimes lot of time, doing my receipts, and stuff.
Sami Kallio: Sorry.
Pasi Nikkanen: Yes, no worries.
But for example, in this one company, they had for this year, 5800 feedbacks on travel reimbursement. The happiness score was +28, so really low. And the lowest time was 1 hour 30 minutes per ticket. So, if we take that ... use your model we have, like 50€ per hour for an employee, that's 75€ every time they waste time on ... they already spend time on filling the thing, but they are adding additional things, additional questions and so on.
And the main factor, the cause that our tool was saying was that, "I had to explain myself several times." Then I thought, "Okay, let's look what people are really saying." So, one person saying like 1 hour 10 minute wasted, to request information regarding data of hotels and travel receipts, was already given, and attached in the SAP Travel in question. This was unnecessary use of time. Another one, quite similar. "You already had that information."
Then I went to another customer. They actually had the average wasted time 3 hours 24 minutes for this year. And one person saying, "Wasted 2 hours 15 minutes, a lot of hassle due to missing numbers in SAP." So basically, when looking at where this request originated, both were using Self-Service portals, but the portals are just like dumb front ends for filling a form. And then it still manually handled, usually in a different time zone. So, you get a lot of this time wasted in that way. But then it's manually handled, and they don't maybe read all the information. So, you could pretty easily now automate the approval and the handling process. The machine could already fetch all the missing information, it could quickly ask with a notification that, "Hey please just give this information."
Sami Kallio: And ask it immediately when you are trying to send an inquest.
Pasi Nikkanen: Exactly, exactly. Immediately on that situation. So, I think that was like a ... And now when they have this data, if they do any automation, now they can compare it the old way, how it was. So, it used to be manual, people were complaining on these issues, this was the time wasted, this was the happiness. Now if we automate it, what improvement will we do?
Another one, basically another example, it was in episode 9, I was interviewing the Finnish Postal Company, Posti. And they had this RPA case, and the happiness for the RPA currently at that point was +100. So really happy. Because they were also asking, "How happy are you with this automated service?" And what they had done was the HR department, there was certain type of contracts, and it was just manually in service now. It was automatically handled by the robot getting all these contracts done, whatever needed to be done and there was no human touch.
So, I think that that's also really powerful.
Sami Kallio: Yeah, I think when doing the automations, there are the time saving, as we started this podcast is, it is the speed of service, that we have to think about it from the end users side, that now we have to handle this in less time consuming way when we can. Because this robot is there all the time. They can ask those questions, they can answer it immediately. They can at least get the right next step for that end user. Depend, it doesn't have to be, again, eight to five time. And that's helping you guys to speed the service up for your end users.
And the second thing I would really think about when you are doing these kinds of things, think about again, like in portals, think about the language you use. And don't try to do scripts that are talking your internal service language or talking about your departments. Your end users don't care. They have their own language, they want to get their things done. Use that language and just help them to get the information they need.
Pasi Nikkanen: True.
Sami Kallio: And I think in automation there would be that two key points I would point out, but-
Pasi Nikkanen: Yeah. And I would say that if the theme of the episode was like, "How does employee experience and automation work together," but that's really with the experience data, you could focus on what we should automate. When you have this as a data, it's again something that everyone can agree on, it's not a gut feeling from me, or as a vendor that maybe you should do this, or maybe you should roll out our new feature. But you actually identify a purpose. How does it change the work for the background process people? But also, for the employee. And then when you do it, you can justify it. So, you can do one case, see how the numbers do it. If you are successful you can justify it, and then kind of prove your next case.
But if you just blindly go and just go with the tech, you will most likely fail.
Sami Kallio: True.
Pasi Nikkanen: But hey, I think that almost sounded like a sad thing at the end. But still I think you know, we want to say that, "Hey, be brave, do it. But have the right measurement and employee experience is probably the one."
How do you make a business case for employee experience measurement?
Links from the episode, Deloitte, From employee experience to human experience: Putting meaning back into workRead more >