Business process automation is essential for managing connected devices at any kind of scale, but it's not the first step in the journey.
Before we can even think of automation, we need to answer questions like "how many devices need attention?".
And there are even more fundamental questions like: "how many devices have I got?" and "how many are supposed to be working?".
Those might sound like obvious questions that any company would be able to easily answer... but in our experience, often not. They may know how many they have shipped, but that's likely to be a significantly higher number, because some devices will still be in the supply chain, and some devices will have been decommissioned.
So the first, fundamental step of the journey is to be able to see how many devices you have deployed successfully, and you could call that process asset management, and you'll want to filter and group by all sorts of things like customer, site, software version etc.
Of course this number changes hour by hour as you deploy more devices.
This is the number against which all performance numbers are measured.
Now that we have an accurate, up-to-date denominator which tells us "how many devices should be working?", we need to define a numerator which tells us "how many devices are working, right now?". The latter over the former gives us the percentage that are working, a ratio, and ratios are good for running businesses with because unlike absolute numbers they tend to stay constant even as the device estate grows. So you can aim for a target of 90% or 99% or whatever, and not have to keep changing it.
Of course, we do need to define what "working" means, which again sounds easy but often isn't. It may be that we don't expect a device to be working at certain times (e.g. outside office hours), or at least that we don't care if it isn't at those times. And ideally we should define "working" in a way which actually matches the customer's experience. It's no use counting a device as working if its software is running but e.g. some button has fallen off, preventing it from actually being used. If the pattern of use is quite constant, then monitoring actual usage is the most fail-safe way of measuring service.
Building a robust set of metrics or KPIs which match customer experience is a great basis for delivering business goals.
Once we have a good set of metrics, we can act on them. In fact, the definition of a good metric is that it is "actionable information".
So now we can start to build business processes around these numbers, such as:
"When a device experiences problem P, take action A".
Action A might be e.g. raising a ticket in a CRM tool, triggering other business processes. Or it might be action on the device itself, such as resetting it. And indeed it's not just about working on a per-device basis, because we can have actions which need to happen when the metrics on cohorts of devices change. For example
"When the site availability on any site falls below 75%, alert the site manager"
If we stop here, then in principle we've empowered our team to see what good and what's bad, and to act on the bad, to drive up quality.
But in practice if we stop here then we've lumbered the team with an awful lot of work, much of which is very tedious and repetitive, which isn't the kind of work that humans enjoy (or are good at). There's nothing worse than having humans stuck on the end of a stream of machine-generated alerts - they'll just get bored and ignore them. Most issues that occur will have occurred before, and in most cases the fix is the same each time, so it's possible to get a machine to do what a human would otherwise have to, thus automating the process of diagnosing and rectifying the issue. So first we create a manual process as defined above, and then we automate it. Doing that for all common issues will free-up the humans from perhaps 80% of the daily grind, enabling them to focus on the long tail of 20% unusual problems - which are rarer and require human ingenuity to diagnose and fix.
The magic piece which enables automation is Rules. Combined with the filtering and grouping of KPIs, and the ability to integrate with other business tools to drive action through them, DevicePilot users can set up completely automated processes to deal with the 80%, enabling them to deliver quality at increasing scale, but without increasing cost.