Few topics in technology have captured more mindshare in the past few years than the Internet of Things (IoT). Despite all this discussion, however, one important point never seems to get the emphasis it should: For IoT to deliver on its promise, system designers must treat it not just as a data management challenge but as an opportunity to apply workflow logic and process intelligence.
IoT discussions tend to focus on data – specifically, on the ability of IoT sensors to generate vast amounts of data about the objects and systems they are monitoring and on the need to detect patterns in all that data. But if we focus only on the data and patterns, we focus too little on the need to easily create business processes that respond to the patterns we have found.
Think about a scenario involving lighting systems for public spaces – streets, sidewalks, parking lots, parks, and the like. The spaces themselves can be vast, so lighting systems are spread over wide areas. Lights fail regularly; it may be months before anyone notices, and months more before a repair crew is dispatched. Throughout that time, lighting is sub-optimal and the city pays a price in public safety.
Clearly this is an opportunity for IoT, and indeed lighting manufacturers are seizing it. They are equipping lights with sensors that can send a signal when the light malfunctions. The signal can be routed to a central system that dispatches repair crews, and voilà – the light is repaired.
It sounds simple, and it is — at least in concept. In reality, even this simple scenario sets up a number of challenges best addressed in terms of workflow automation and process intelligence.
The first step toward an effective solution is to think of the system’s purpose, which is not to manage outages but to optimize lighting. That requires a consensus on an agreed level of lighting in a defined service area, which can probably be captured in a service-level agreement (SLA). Then the challenge is to manage to the SLA, within the universe of municipal considerations of costs, budgets, and competing priorities.
The most useful way to frame this challenge is in terms of an automated process or workflow. These workflows cannot rely on paper; they must be digital. They must minimize manual data entry. And they must be created without coding. Approaches that require coding often are not used because of the time and expense they entail.
It’s true that many cities and other agencies do not have these systems in place today, and that they face various budgetary, political, and technical challenges in implementing them. But it’s also important to recognize that without such systems, the goal of optimal lighting management will be nearly unattainable.
In its most basic form, the system must do three things. It must be able to capture sensor data regarding outages. It must connect with an inventory management system that ensures parts (i.e. replacement bulbs) are available and replenishes them as they are consumed. And it must connect with a system for scheduling and dispatching repair crews. But these capabilities are just a starting point.
Lighting management occurs within a broader context of city operations, so an effective system will connect to the city’s systems for human resources management and payroll. Costs must be managed within the city’s overall budget process, so lighting management must integrate with broader budget and cost management systems.
Even with these capabilities, this system still falls within an “outage management” concept rather than “lighting optimization.” To achieve the latter, process intelligence is required. For example, the system should be able to track and analyze frequency and location of repairs to provide insight on problem areas.
If a particular area consistently requires more repairs, administrators can explore why and respond appropriately — for example, by implementing a preventive maintenance program. The problem might be vandalism; it might be corrosive salt air; it might be any number of things. But an effective management system must be “insight-based” – that is, intelligent — not just anecdotal. An intelligent management system must enable automatic modification of a workflow based on pre-defined rules.
It is also critical that IoT workflows not be hard-coded; they must be flexible and easy to adapt, either through human design or automatically, based on data. If it is complex, slow, or expensive to modify a workflow, the system will be of limited value.
Ideally, workflows should be defined by the people who use them. From an initial iteration, people will think of additional connections and functions over time, and they should be able to tailor the solution as these ideas arise. We might say that people innately want to “work better,” and their workflow and process intelligence tools should give them that capability. Designers need the ability to work within their systems of choice and to build processes that interact and integrate with disparate systems of record.
How does this relate back to IoT, workflows, and process intelligence? To put it simply, broken light bulbs don’t exist in isolation. They occur within broader systems that must be connected and adaptable to deliver value to the organization.
So yes, granular data from IoT sensors is essential. But to focus too much on the data risks missing the forest for the trees. The real value of IoT comes when we envision an IoT system as a workflow that connects a broad range of enterprise systems, and endows the system with process intelligence that shows us how to optimize its value.
This article originally appeared in the May 2017 issue of Workflow.
Latest posts by Alain Gentilhomme (see all)
- In Workflows, the Point of Automation is Optimization - October 24, 2017
- Automation and Innovation: Choosing the Right Tools - September 1, 2017
- Process Intelligence is Key to Value in IoT systems - May 1, 2017