To say that AI can lead to great efficiencies, particularly when it works right, is uncontroversial.
“Ultimately, the rise of AI is raising the premium on tasks that only humans can do: it is freeing workers from drudgery and allowing them to spend time on more strategic and valuable business activities. Instead of forcing people to spend time and effort on tasks that we find hard but computers find easy, we will be rewarded for doing what humans do best — and artificial intelligence will help make us all more human.”
– Timo Elliot, Forbes
For the service desk, the nuance comes in understanding the difference between AI Automation and AI Augmentation.
Here by “AI Automation” I am talking about conversational chatbots. The function is conceptually simple, although it’s not simple to create one and make it work right. We have our own chatbot and we have great respect for our colleagues at various chatbot companies, and there are quite a few of them.
By AI Augmentation I mean using AI in partnership with a service agent, each doing what they’re best at.
According to Forrester’s “Chatbots for IT Operations, Q4 2020”, for the IT Operations use cases it’s more realistic to expect 40% deflection. Also, from the same report, “Making chatbots useful often requires significant development efforts”.
Some Chatbots vendors will claim resolution rates of 70-90%, but you need to be careful about what use case they’re talking about. If your Level 1 first call resolution rate is 70%, what are the chances a Chatbot will resolve all of Level 1 and 2/3 of Level 2 and Level 3 issues?
So let’s stick with Forrester’s number of 40% for now. But how far does 40% deflection even get you?
Let’s look at an illustrative example.
If you’re running a 10,000 ticket/month serviced desk operation, and your costs per ticket and resolution rates are industry average, let’s just take the numbers from SDIv9 and MetricNet, then this is more or less what you an expect by way of where tickets get resolved:
Now let’s look at how much it costs to resolve a ticket at each tier, according to MetricNet:
You put those together and you find out that this is how your annual staff budget is spent:
Whew, those Level 2 & 3 folks are expensive! They don’t do as many tickets, but you can see that the cost at the back end dominates the cost at the front end of the service desk.
Ok, now release the Chatbot!
Let’s say it no-touch resolves 40%, per Forrester. Now, you think your Chatbot is going to resolve Level 2 and Level 3 issues for you? We think not. In this model we’re going to say that it only resolves Level 1 issues, and so the picture now looks like this:
Note that the top bar is just another way of visualising the first pie chart.
The first thing to note is that more than half of the Level 1 workload (blue) has been shifted left to Level 0 (light green). However Levels 2 and 3, the majority of the cost in the previous scenario, are unchanged.
Nonetheless, an extremely useful capability. So far so good.
Let’s look at the resulting financial picture. MetricNet reckons it’s going to cost you $4 to resolve a ticket via a bot. If we go with that assumption, this is how your money is spent:
Looks like we saved ourselves $936,000 a year, or around 20% of our initial staff budget.
So that’s a good scenario. With AI Automation, in a perfect world , you might deflect maybe 60% and that would get you $1.2m in efficiencies. Pat yourself on the back, take the team out to dinner.
However if you really wanted to move the needle you’d need a solution that addressed not only the column on the left but also the column on the right.
When you add augmentation into the picture, the “shift left” starts to happen across all tiers of the service desk.
Notice that between the “Automate” and the “Automate and Augment” scenarios there is a subtle shift left from Level 3 through to Level 1. It’s not huge, but the effect is disproportionate due to the high cost of those resources.
Comparing the numbers,
Dividing by the number of tickets per year, this equates to around $19.11 of savings per ticket. So if you’re doing 1,000 tickets per month you’d be saving $229,000 annually by implementing an Intelligent Service Desk.
Pretty significant, right? How many monthly tickets does your service desk process?
Finally, if you prefer to redeploy the service desk staff freed up by the increased efficiency, this is much their time is worth:
It’s pretty compelling stuff, but the real question is whether an intelligent service desk can deliver those sorts of gains. Contact us for a demo and find out how we do it.