IOT is an environment of total dynamics with constant change of structure and interaction. It is filled with complex events and patterns that will defy human expert analysis and prediction. Therefore BPM has no relevance in this arena. But yes, those who lack an understanding of complexity and want to throw everything into BPM plus the kitchen sink will do so. IOT will defy predefined processes created by analysis, design, implementation, modeling, deployment, monitoring and improvement. There is no time to employ data scientists and process experts in this world.
IOT will use complex event processing and machine learning and also that is not BPM, despite other claims. There won’t be any process mining happening because of the diversity of data and the constant change.
IOT will push towards new ways such as machine learning to create the interaction with the human. But this is not your big data or predictive analytics humbug, but humans will teach software what to do when similar patterns appear.
Process will be absolutely fundamental to the IoT. After all there’s no point capturing data if it isn’t analysed and if necessary used to trigger a follow up action.
The industrial IoT has been alive and well for years providing critical services in high-stakes industries. They key point is that Industrial IoT is really not about devices but instead it’s about providing essential services. At the moment the consumer IoT has been focused on devices rather than services. This device focus is inevitable as sensor and Wi-Fi technology wait for other essential parts of the consumer IoT ecosystem to catch up. One of those missing parts is process. We are looking at a process future beyond the BPMS, to the consumerisation of processes, to BPaaS, or high volume, on demand or embedded processes that work with sensors and IoT devices that either eliminate mundane or repetitive processes or provide critical services.
BPM should help simplify the complex so that everyone is working consistently. IOT has made business infinitely more complex. BPM has a critical role to play in helping businesses understnad how to exploit and manage the explosion of data from the IOT.
This is a very wide subject. And the Internet of Things (IoT) is the beauty lady: everyone talks about her, everyone claims having some kind of relationship with her, but nobody really knows how to really marry her.
I have to disagree with Mr. Max J. Pucher. BPM has a lot of value to add to IoT, as several other disciplines and technologies do. The point is ‘where’. I mean, there’s no return on applying BPM technology to a fridge or a wash machine JUST for home use. But this is not the whole IoT world.
Think for a while in an industrial process. A real one. With costly machines producing thousands of parts to repair fridges each hour. Thousands of parts stored until someone buys them for a broken fridge. A huge stock of parts. A lot of sleeping money. Why don’t produce the part once the fridge is broken, or about to broke, but NOT before? The fridge could alert about it (that is IoT too). And once produced, send the part to the house immediately. Isn’t that a business process? Should a good BPMS help there? Of course YES. But that is just one, simple, basic, example.
There are milliards of applications of IoT in which the integration with a BPMS will be mandatory for success.
In the "Internet of Things", some "things" will be primarily for reference and others for action. I might want my fridge to order more milk when it notices I'm low (while, or course, taking into account current vs. predicted price for optimal order timing). Great opportunity for BPM there. More likely, though, when I'm at the market I'm just going to check in with my fridge using my mobile phone to see if I need milk; hence, no process.
Not everything, and not every Internet-connected thing, needs process.
Processes are very important to the IoT (Bosch, which ought to know something about "things", is betting on process technology as central of IoT programs; they are not alone).
1. WHAT IS THE IOT? To answer the original question, let's agree on what the IoT "is", other than a topic at the top of the hype curve for journalists to write about. : )
One could say that the IoT is just the latest generation of computing and software. We've had RFID and sensors and machine control and messaging and integration and analytics and dashboards for a long time. And now engineering and economics have pushed down the cost and size of sensing and computing devices by orders of magnitude -- so that we can now do computing on a scale never previously possible and integrated with human, social and technical systems as never before. The potentially large scale of IoT projects demands new software capabilities, including decisioning (DMN comes along at just the right moment), machine learning, distributed analytics and control, mesh networking yadda, yadda, yadda.
2. WHAT IS THE CORE IOT CHALLENGE? I've written elsewhere that the challenge of business semantics is the core challenge of the ioT (see my LinkedIn post "Business is the Only Thing: Business Semantics and the Internet of Things"). If you just wire up sensors to dashboards, without any intelligence in between, you've just built a fancy telephone switch. Which isn't much of a business model. All the investment in IoT only makes sense if in between "edge" and "management", there is software to help humans do more of the work for which an IoT project is deployed.
3. THE IOT IS ABOUT WORK -- FIELD SERVICE EXAMPLE. Let's say you want to help field service technicians responsible for refrigerators installed at retailers. You can imagine deploying sensors to alert dispatch when a compressor is giving indications of failure. And you can use rules to help do that "predictive maintenance". But notice that initial deployment, condition sensing, dispatch, route planning, truck stocking, ticket generating etc. involve multiple algorithms and rule decks, deployed via various machine and human interfaces -- and all tied together by process. And the use of business process technology makes evolving that field service programme much easier -- which is really important if you want to also evolve your business offering, meet the SLA, and win more business!
4. BPM IS THE TECHNOLOGY OF WORK -- BPM is the machine-friendly language of work. Maybe BPMN isn't perfect yet, but there is no other technology that makes "first class citizens of the concepts of work". (One could make an argument that the identification of BPM as the technology of work is almost tautological.)
The ultimate intention of BPM is that managers and executives can work directly in the abstractions of business process technology to faster organize the work for which these executives are responsible. Decisioning technology and algorithms are almost as important; other technologies, including SOA and integration technologies, play supporting roles in the domain of IT.
What's the alternative? Business has to express business semantics in software. Software is the force multiplier that makes our modern world possible. So to acquire the software-embodied business semantics, one either picks "code", and laboriously builds up frameworks which ultimately amount to the business-semantics-of-work-by-proxy, or embraces the power of BPM directly. It's ultimately much easier to choose BPM technology. And then get on with the job.
IoT when deployed in a business environment needs process to join up thinking and action. Good example in home health environment. Sure IoT gadgets used by individuals can help families but when it becomes linked to the institutionalize agency then process can improve outcomes and efficiency. As the gadgets create info so the relevant process can be triggered with real time notification of the right required support and escalation if required.
There will be many other examples and that may includes "apps" where process creates that joined up knowledge leading to effective action.
In my view the question should be focused on the business outcome: How important is it to sense key business events from machine-born data (devices, sensors and IoT), decide if action is needed and then respond in an appropriate manner.
Deconstructing the question like this leads to 3 key elements for consideration.
Firstly, organizations need to gather Operational Intelligence using event stream processing and complex event processing as Max pointed out. To do this the OI platform needs some form of adapters or connectors to easily “Ingest” information into the OI platform. The streaming event data (from IoT or other sources) may need “Conditioning” before it is “Integrated” with contextual business data before “Rules” are applied to identify key business moments or trends.
The second element is to provide decision support. It can be as simple as a business rule or it can be more involved with social collaboration (asking others for their opinion) or provide BI style analytical support. Machine Learning can applied at this point for predictive analysis if the use case supports it. The objective of the decision support element is to decide if an action is required. Not all IoT data needs to be acted on.
The third element requires an “appropriate” response based on application. Sometimes is can as simple as a well-defined workflow for a planned response to an anticipated exception. BPM adds value in this scenario. More often than not it required an unplanned response to unanticipated event. In this case it may be as simple as sending an SMS alert or, as we often suggest to customers, an Adaptive Case Management (ACM) solution to orchestrate the response on a case by case basis. The appropriate response often requires automation of actions in another business system such as the Supply Chain Management module of an ERP solution. In that case BPM does not play a role as an OI solution with agent behavior can accomplish this.
There is no simple answer for dealing the sea of data from IoT other than creating some Operational Intelligence from it, deciding if any action should be taken, and then use the appropriate technology to manage (machine or human) the response. We call it Sense > Decide > Act
BTW my simple definition on the difference between BI and OI is that BI requires storing data and the creating some retrospective intelligence over it (what happened) where OI first interrogate event data (what is happening) before deciding if it should be stored.
The essential question is how “Intellect-of-Things” (IoT) will contribute to this civilisation. Potential on synergy are huge thus very critical (thanks to Patrick).
Considering that things will participate more and more in our lives, those things must be intellectual enough to don’t annoy us but help us. The latter is feasible if those things a) understand our lives/situations as processes, b) accordingly adapt their own behaviour to fit better those processes and c) join their efforts if necessary. This is very similar to implementing customer-experience-as-a-process (see http://improving-bpm-systems.blogspot.ch/2013/06/practical-process-patterns-cxaap.html and http://improving-bpm-systems.blogspot.ch/2015/01/bpm-for-cx-customer-experience-example.html ).
We just need to teach IoT how to use BPM correctly. For example, use process-mining to understand implicit processes and make them explicit, tailor standard process patterns, quickly develop a process and share it with others to collaborate in a particular case, etc.
What are your data points in your current BPM process?
A person fills out a form, clicks submit and your process kicks into action.
The form is filled, the action recorded and analysis can then be done.
But what if the person isn't there? Can we make the form fill itself out?
In many cases, through the internet of things we can.
For example, take an order delivery process. Normally a form filled out by human operatives at key stages.
Stages which can be automated. When the item is picked, IoT sensors can record that fact. Same when it goes on the truck, enters the warehouse, goes on the van, arrives with the customer.
Interestingly, humans being what they are, the error rate through these automated touch points is usually lower than the human "mistype, misroute, forget to record, put it down somewhere and forget where" process.
In future most BPM processes will have IoT inputs as prime data.
That answers the question "How important is the Internet of Things to Process".
But turn it round.
All those IoT sensors. All that data. Looking for a killer app. Something to make them make sense.
Process is that sense. Turning random data into something with meaning, making sense of sensors on a variety of seemingly unconnected products. Making each sensor a vital step in a logical, connectedProcess.
Process is what IoT lives for. And IoT is the killer app Process has been waiting for too.
Now there’s an interesting question. The answer is that processes could be very important to IoT, but it won’t necessarily turn out that way. Before I get into that, I want to look at the other question.
Emiel’s counter-question is interesting too, and not only because we always seem to want to change the original question before we answer it. IoT is hugely important for processes, because so many IoT applications are about sensors, i.e. process triggers. Internet connected sensors create many more opportunities for triggering processes that otherwise be outside the reach of BPM simply because the manual alternative would be too laborious, too unreliable and too late.
For example, you’re far more likely to follow-up fire alarms in a consistent way if the alarm itself can trigger a process engine, or check refrigeration equipment if a chilled area’s temperature rises above normal limits. You can probably think of dozens of examples yourself, and this might be what Gary meant.
Returning to the original question, I can see why this importance isn't symmetrical. Despite enabling BPM in interesting ways, IoT doesn't need BPM. After all, how many IoT devices - things like Nest - integrate with BPM platforms? I haven't counted but I’m guessing none, because IoT is mainstream - for everyone - while BPM is not. That BPM is a far more specialised area than IoT is the point that Peter Whibley misses, I think: although there will certainly be data capture and automation, it won't necessarily use a BPM approach.
BPM could be useful to IoT applications, though: if a BPM platform offered simple enough cloud-based integration and a usable user-interface, then it might actually be appealing to IoT applications to apply a process management approach. Wouldn't that be something?