1. Peter Schooff
  2. Sherlock Holmes
  3. BPM Discussions
  4. Tuesday, 13 December 2016
  5.  Subscribe via email
Looking back at the year almost past, which tech development from the past year do you think will have the biggest impact on BPM in the years ahead?
Accepted Answer Pending Moderation
8K video cameras, which will make it easier for robots and people to "see", allowing 'users' to do more work "at a distance".

The result - an increase in automation and a greater need for interoperability in/out of environments where background BPM provides orchestration/governance

Big implications for facial recognition and pattern recognition.

  1. http://www.kwkeirstead.wordpress.com
I actually have a pattern recognition problem I need help with

1. easy scenario - truck rolls up to some venue, a camera is trained on near side, the opposite door is opened (cam is not able to see this), a package is deposited, the truck goes away - we can detect the package by comparing before/after.

2 difficult scenario - same, up to "door is opened", this time a package is deposited into an existing trash bin (taking care not to move/rotate the bin) - we cannot detect the package because nothing has changed in the scene and no, we don't have an X-ray device that can see into the trash bin.
a/ train your camera on "truck here" being the before/after comparison (unless you have trucks coming in that do not actually deliver something)
b/ get a smart bin, with some sensors (proximity, weight) that say "there's something inside"

(sigh). wish it were that simple. There are 500,000 sites that need monitoring. This is a large project.
  1. Emiel Kelly
  2. 1 year ago
  3. #2991
Let's apply some Poka Yoke and get rid of the trash bins on the delivery spot.
  1. more than a month ago
  2. BPM Discussions
  3. # 1
Accepted Answer Pending Moderation
AI - but I am not sure quite how it will play out - the obvious examples are voice activation.
And, what is behind it...
  1. more than a month ago
  2. BPM Discussions
  3. # 2
Accepted Answer Pending Moderation
As you may know, I work on the Systems Approach and its application to the system domains such as IoT, Smart-Home, Active-Assisted-Living and Smart-Cities. On my site you may see a few reference architectures ([ref1 and ref2]) for those system domains. I found that explicit and machine-executable business processes (as we understand them in modern BPM) are one of the essential system-forming factors for those domains. BPM helps with:
- digital contracts (see [ref3]) which are critical for the high level of privacy and security
- coordination between various actors
- management of various assets and resources

Thus I can say that IoT (joining physical and digital words) and blockchain (as the ultimate records management) have the biggest impact on BPM.

  1. http://improving-bpm-systems.blogspot.ch/2016/12/smart-home-as-system-of-systems.html
  2. http://improving-bpm-systems.blogspot.ch/2016/11/thing-as-system-reference-architecture.html
  3. http://improving-bpm-systems.blogspot.ch/2016/08/iot-as-system-of-digital-contracts.html
My whole-hearted vote goes to Deep Learning. Cognitive Computing/Machine Learning in specific areas that affect, improve, and accelerate business processes.
The more we focus on the Decision Management in our businesses, the more we improve outcomes.
Cognitive applications (like NLP & Voice to start) take codified rules into the gray area of value.
It is still all about value. I still maintain that these areas will bring more value per implementation, but unfortunately, most BPM folks are still focused on the tiniest of muda.
The number of Use Cases discussed for Cognitive Business Operations amazes me.
Happy Holidays, All!
Hmm... Your favorites are about improving the EXISTING way of BPMing. What's about innovations?
Hi, Alexamder - SWIFT may be a good candidate for blockchain
Exclusive: SWIFT confirms new cyber thefts, hacking tactics
@Karl, At first, SWIFT should be a BPM application.
@Alexamder, I definitely vote for your opinion. IoT incredibly extends demand for BPM and increases its scope of applications.
  1. more than a month ago
  2. BPM Discussions
  3. # 3
Accepted Answer Pending Moderation
Hmmm, what IT tech development will have the biggest impact on BPM in the years ahead? (We can technically define this as the net present value of the area under a long tail of a BPM benefits curve that could be traced to a specific enabling technology -- yikes that sounds hard! So, clearly a judgement call!)

And any judgement call is a call not only on potential impact (what technology we'd like to succeed) but also a call on what technology is likely to persist and not drop off the hype curve. Speaking of a hype curve, Gartner's "Hype Cycle for BPM 2015" (most recent version you can find without a subscription) lists about 40 technologies impacting on BPM! And what about technologies not listed on the that chart? Process mining is not listed, for some definitional reason I guess. Even more candidates.

So my vote is for business rules and the DMN spec, officially launched only in September 2015, and really gaining momentum in 2016 with several products well into market. Business rules of one kind or another have been around for ever, but DMN is a big step up. Business rules and DMN are especially attractive because business rules deployed as complementary to selected business processes can enable enormous process simplification.

There's one more criteria for choosing a technology - the effect can't already have been achieved. This is the case with business rules and DMN; wide-spread adoption has not been achieved beyond certain defined use cases.

Here are all the reasons why DMN and business rules might generate a significant on-going future impact on BPM technology adoption and usage.

Business rules technology . . .

1) enables huge benefits both technically and business-wise (especially against complexity),
2) is in-market and in-channel,
3) is as easy as BPM to learn and use (and maybe even easier!),
4) is understandable to management,
5) is found as a use case in many places not yet served by explicit business rules technology,
6) is a perfect complement for BPM as a core business semantic technology,
7) benefits from complementary business demand for analytics and big data,
8) as a technology has momentum and is continuing to improve, including around UX.

Place your bets. Given the use cases characteristic of your business model, a business rule play is one to consider for 2017.
The BR/BD technology is about improving the EXISTING way of BPMing.

My struggles with "rules" seems to be proceeding on an entirely different track. . . . . .

From 3 years ago . . .

Check your Business Rules on the way in and out
I have always regarded rules as important component of BPM.....business is full of rules....? Any BPM support platform must surely support rules and of course this opens the capability to start delivery of intelligent processes to help make the right decisions...?
  1. John Morris
  2. 1 year ago
  3. #2983
@Alexander: You are certainly correct that BRM relates to existing BPM; I suggest though that that doesn't mean there isn't a big win waiting to be claimed.
@Walter: Your link provides a great example of the benefits of abstracting out rules for better management overview.
@ David: It's true that many (or all?) BPM platforms support rules. But proprietary implementations (even hacks) are not the same thing as a community standard.
The example I like to cite for the benefits of rules with BPM is a simple escalation procedure for shop floor control. A BPMN-defined escalation procedure process diagram looks like spaghetti. The right approach is to abstract out any significant collection of rules into a rules application (e.g. here the escalation rules). Easy for everyone to inspect and use. And especially important, the BPMN diagram is then enormously simplified. It's surprising how often this sort of thing isn't done, thus the opportunity. The arrival of DMN and DMN-supporting products (commercial or open source) means that rules technology can more easily migrate from IT speciality to business toolkit.
@John, I m big fan of DMN myself and tabular BPM representation in general. Diagrams are not always the best way to expose business logic. An optimal combination of diagrams linked with tabular representation of processes and metadata give an optimal vision and balance. DMN is an excellent step in this direction.
  1. more than a month ago
  2. BPM Discussions
  3. # 4
Accepted Answer Pending Moderation
Definitely smart watches. Every process will be managed from the wrist in 2017. Exciting times ahead.
Sharing my adventures in Process World via Procesje.nl
Hmm.... "Shipments of smartwatches declined over 50% in the third quarter." http://fortune.com/2016/10/24/smartwatch-market-crashing-apple/
I think Emiel is contemplating a commercial application for smart watches that would be pretty distinct from the consumer market at whom those products are currently targeted. And we do already see industrial applications for wearable devices in healthcare, etc.
  1. Emiel Kelly
  2. 1 year ago
  3. #2967
Really? Oh no....
  1. Emiel Kelly
  2. 1 year ago
  3. #2968
No scott, actually I was triggered by the news Alexander referred to. My answerwas meant ironically (not so good I have to explain it ;-(( but to me smartwatches are a brilliant example of a basic lesson in bpm.

Understand first what problem the result/product of a process should solve before you start implementing and executing the processes. I have no clue what problems a smartwatch could solve.
But I have to admit, I never liked jewelry
RE "commercial application for smart watches " so, do I understand correctly that you don't expect customers be participants of business processes?
  1. more than a month ago
  2. BPM Discussions
  3. # 5
Accepted Answer Pending Moderation
Agree with those above (and presumably below) who are suggesting that AI will have the greatest impact. Should be pretty cool.

Consumer tip: remember to verify that your BPM product is "Three Laws Safe".
From the moment we put the fox in charge of the hen house, we can expect trouble.
  1. more than a month ago
  2. BPM Discussions
  3. # 6
Accepted Answer Pending Moderation
I'll go ahead and agree with @John: DMN looks like a modest contribution but, if used right, it should significantly change the way we model business processes. Another good candidate with significant impact (although we haven't yet figured out the problem space) is IoT (as per @Alex's take).

As for AI, blockchain - I'm not holding my breath.

Was just reading this yesterday: http://www.slate.com/articles/technology/future_tense/2016/12/software_problems_are_leading_to_wrongful_arrests.html
One of the ending paragraphs is particularly suited for AI as well:

These problems will likely worsen as software increasingly becomes embedded in everything we do. Odyssey clearly has its flaws, but at least court employees can identify a problem like a recalled arrest warrant, even if it’s too late to stop a wrongful arrest. With other types of software, however, errors may be difficult to detect. Algorithms designed to help judges decide bail, or to help the police identify suspicious behavior, may be hard for nonexperts to understand, let alone critique. The private companies that design and sell these products may also be reluctant to share their proprietary information.
CEO, Co-founder, Profluo
  1. John Morris
  2. 1 year ago
  3. #2984
Really important insight @Bogdan on AI. And a scary example of the dangers of AI.
Before people classify me as an AI-sceptic - I'm not at all. I have been through the AI field and I can vouch on its power to solve things human minds can't. But right now AI is plastered on every single thing just to make that thing sound way cooler than it is.
AI is excellent for certain problem spaces around objective function optimization. It thrives on massive data and is able to discern highly non-linear models where they exist. So discovery of subtle relationships and convoluted models is amazing. But you have to feed a lot of data to the beast. Not all business problem spaces will have that amount of data and so many critical relationships to unearth. Sure, feel free to stick AI on top of that and sell away to unsuspecting customers. But it's not going to make your BPM product better, anytime soon. Unless you generate 2 billion USD in BPM revenue.
  1. more than a month ago
  2. BPM Discussions
  3. # 7
Accepted Answer Pending Moderation
We have already discussed blockchain but could be much bigger as global events unfold where there is a move to "remove" cash.... yes really is the vision and you will have seen many countries call in notes as the start of a long new journey. Traceability will become dominant and of course is about people and their activity. This could be big opportunity for the BPM thinking .....I always have been ahead of the game but I think this journey may start in 2017....?
I think, this is a huge opportunity for BPM. Have a look at this "blokchain" demo https://www.linkedin.com/nhome/updates?activity=6214614635522322432 - it is a normal BPM application (being positioned as B2C or B2B provider).
  1. more than a month ago
  2. BPM Discussions
  3. # 8
Accepted Answer Pending Moderation
Several BPM.com contributors have voted for AI. And the arguments and judgements are compelling. Here however is a counter-argument (with tip of the hat to @Bogdan, per his note above). AI adoption and corollary benefits may be slower to materialize, for the following reasons:

1) DOMAIN SEMANTICS COST -- AI depends on domain semantic models. As we learned in the 80's and expert systems, domain semantics (or "expertise" ) is very hard to find and codify.

2) MATURITY OF TECHNOLOGY TIMING -- Machine learning is a possible solution to the cost of building domain semantic models. ML-driven AI is still aways away from mass adoption.

3) EARLY ADOPTION RISK -- The AI-based business errors we are already seeing are (see @Bogdan above) to be expected at this early stage.

4) DECISION RISK AND LIABILITY -- Humans can make judgements in which we can have a measure of confidence. The distribution of AI-based errors is unknown. This means great risk -- and liability.

5) TRANSPARENCY RISK -- We are seeing more and more concern about the lack of transparency for the decision outputs of AI. Business executives don't like not knowing why decision are made.

6) JUDGEMENT ATROPHY RISK -- No organization will want to de-skill core functions to AI; to do so would be to guarantee the loss of human judgement capital. This is the surest route to commoditization.

7) AI GOVERNANCE COST -- Automated decisioning is a new field and requires new governance (.e.g how to manage errors such as, per above, a "false arrest warrant problem" ). Putting in place new governance for AI as part of general management practices won't happen overnight.

AI is the goal, on top of domain semantic models, sensing, powerful analytics and machine learning (ML). And as the BPM future benefits tail stretches off into the future, certainly we can expect -- and look forward to -- amazing benefits. But that future has so many uncertainties as to be in the realm of science fiction. AI is technically difficult and is a governance challenge. Those future benefits therefore must be severely discounted for any assessment of value today.

We used mdbs' GURU in the 1980's for building hotels.

Without the AI person sitting in on every meeting, every discussion, we could never explain to the stakeholders why the model was saying this or that

This morning, I am trying to communicate a 'simple' model for picking strategic initiatives - I have 8 projects and 8 filter criteria and we, of course, have cells with calculations that cross link the 64 options.

One stakeholder refuses to look at the spreadsheet, another finds it complex - I did a GoToMeeting with my tech and they say it all makes perfect sense.

I1 lifecycle short/long
I2 grant success difficult/easy
I3 revenue potential low/high
I4 devel challenge high/low
I5 sales challenge high/low
I6 labor intensive high/low
I7 server deployment vs cloud (0,1)
I8 crowded market high/low

Each criteria other than one has a rating of 1-> 10
Re " ML-driven AI is still aways away from mass adoption."

True, but we don't need to stand by whilst we wait for predictive analytics.


Process Improvement. Perform work along BPM pathways, log the data. Now overlay stats from many instances on your graph and observe such things as:
1. the users always skipped step 122 - remove the step
2. the users never engaged sub-path aa - remove the option
3. the users inserted an ad hoc step at 179 60% of the time - roll out a new path with the step embedded in the new path.
4. . . . .

Lessons from other domains . . .

CPM has always had predictive analytics, sure it's easier, given we are in a deterministic environment (i.e. most of the work always gets done), but the ES-EF-LS-LF calcs let you know today if you will arrive on time and within budget.

How's that for keeping a sharp focus on objectives?

Hard to do this in BPM when the time gaps between steps rival the actual time to perform and worse still, when you have process fragments and you don't know who is going to insert an ad hoc step at any stage of Case timeline.
@John, excellent point on semantics! It is a false but wide spread impression that AI is capable of creating knowledge. It is fundamental and dangerous delusion. Yes, AI is capable to solve narrow and well described technical tasks in well formalized domain, often better than humans. But AI is entirely incapable in creating new knowledge, specifically in form of semantics. In terms of BPM, AI is superior in aligning and executing established processes, but AI is principally limited in exploration and synthesis of new processes.
  1. John Morris
  2. 1 year ago
  3. #3002
@Walter, absolutely marvellous story about using GURU and always having to have a "human interpreter" present to explain things (and at what cost too . . . )
@Boris, excellent highlight on what AI is truly good for. The rest of the time, AI is a nice topic for science fiction or magical thinking. :)
"Process Improvement. Perform work along BPM pathways, log the data. Now overlay stats from many instances on your graph and observe such things"
That's not Machine Learning - that's standard Process Mining, and it's got very clear algorithms for that.
  1. more than a month ago
  2. BPM Discussions
  3. # 9
  • Page :
  • 1

There are no replies made for this post yet.
However, you are not allowed to reply to this post.