BPM.com
  1. Peter Schooff
  2. BPM Discussions
  3. Thursday, April 13 2017, 09:51 AM
  4.  Subscribe via email
From Karl Walter Keirstead: “Case” appears to be generally accepted as a run-time environment for hosting BPM process template instances and process fragment instances. What would you say are the “must have” features for a Case Management System software suite?
John Reynolds Accepted Answer
Case Management products must be able to support highly structured Business Processes :p

I'm actually serious. Businesses aren't served well by having separate platforms for case and bpm.
Comment
+1. Hell yes.
  1. E Scott Menter
  2. 2 weeks ago
Oh, the sacrilege! :-)
  1. Bogdan Nafornita
  2. 1 week ago
+1 . . . Users of a CMS should, at any point in the lifecycle of a Case, to be able to

1. engage processing of a process template instance,
2. engage processing of a process fragment template instance,
3. insert an ad hoc step

Users also need to be able to suspend or terminate a pathway that a Case record may be on, skip steps, and record data at steps not yet current.

In respect of already committed steps, (any of the above), users should be able to re-visit steps and record data.

Already committed data must NEVER be overwritten. so data at a step that was committed on 2017_04_15 at hhmmss must remain in the Case History (Hx), as is, with new/overriding data appearing in the Hx at, say, 2017_04_17 hhmmss..

Refinements to this protocol include a strikethrough capability at any "yyyymmdd hhmmss" Hx record to reduce the chances of a user going to the Hx, relying on data values at that date/time thereby carrying forward an error. (i..e in medicine, recording of BP as 190/90 when 130/90 should have been recorded).

For this reason, the order of sequencing of date/time entries in the Hx needs to be reverse chronological (newest first) order.

A more versatile solution would be a facility to 'attach' a "sticky note" to to the original recording at the Case, clearly indicating to a viewing user that one or more data entries at the original recording have been superseded by later postings to the Case.

The BP data recording example reinforces the need for "smart" processing.

In the example, the bad BP reading could have resulted in issuance of a prescription giving rise to a subsequent need to

a) contact the patient to advise them to stop taking the medication then
b) contacting the pharmacy to reduce the number of refills to 0 then
c) writing up a mandated medical error report.

Clinicians reasonably need to be able to view all pathways a patient is on with the equivalent of felt pen progress markups as part of "fixing" the problem(s) caused by certain errors so that appropriate action(s) can be taken.

Easy to see that a consolidated "workspace" is essential for "Case".
Keith Swenson Accepted Answer
Oh, so many Peter!

The thing about case management is that knowledge workers have quite varied needs, so case management systems tend to be quite varied to meet this. There are some very common requirements though:


  • Documents. All types. Searchable. Case management centers on a folder.
  • Commitment Tracking: what needs to be done, when, what have been done, by whom.
  • Dynamic tasking: the ability to ask anyone to do anything at any time and the system facilitates
  • Flexible Communications: comments, chat, notes, etc. Kinda like a social platform but exclusively to the case.


The above form the core of any case management platform. Beyond that you will find specialization in more structured documents. If you are dealing with legal cases, there are going to be specific databases, with fields, and forms to view and enter the data values that are specific to that domain. This is where the customization becomes important. Along with ability to access external data sources and sinks. Like to look up a criminal background, or trace the title on a house.

P.S. we will be having the AdaptiveCM Workshop again this year, this time Canada in October.
References
  1. https://social-biz.org/2014/06/16/open-source-adaptive-case-management/
  2. http://acm2017.blogs.dsv.su.se/
Comment
Great points, Keith. Also, I was trying to check out your OS ACM project - but it seems to be offline. Please, let me know, when you have it back, online.
Thanks!!
kay
  1. Kay Winkler
  2. 2 weeks ago
Fascinating reference Keith to "commitment tracking".
As an entertaining detour on the road to better software, it's worth noting the history of the idea of "commitment tracking" as a first-class concept in process and case automation. (I'm sure you're familiar with this history.) One of the earliest examples of a focus on the idea of "commitment instantiated in software" was in the 1980's. Based on the work of Terry Winograd and Fernando Flores, Action Technologies' product "the Coordinator" was an early pioneer in workflow -- and specifically very much based around the idea of commitment. (The backstory of the Coordinator is fascinating, and even touches on Soviet cybernetics.)
Please allow me to underline the importance of the items in your list above; these items are not just "nice to have features", but fundamental concepts that enable humans to perform real work more easily. And especially the idea of commitment is a foundational.
  1. John Morris
  2. 2 weeks ago
Here's S+B in 2009 "Fernando Flores Wants To Make You An Offer"
https://www.strategy-business.com/article/09406?gko=ce081
"Spend any time with Fernando Flores and he will assess you. He may make an offer, which you are free to accept or decline. If you accept, he will make a commitment to fulfill his promise."
  1. John Morris
  2. 2 weeks ago
Great list. +1.
  1. E Scott Menter
  2. 2 weeks ago
"Documents" by far is the most difficult

Unless you convert all document types to a few, browsing Case Histories requires that you load the environment (MS Word, PDF, etc.) and have the document open in its native environment, displaying the "hits".

Clearly, it's less than desirable to have to "point" to a document so the likely solution is your document converted to text in a BLOB with the original in another BLOB. Your search algorithm goes to the text version and then you highlight the hits in the native version.

if you point to a document you risk the author deleting or changing the document or changing it's location.
Here is a mockup of the posts at this Forum, illustrating the complexity of doing searches across multiple document types.

The video only runs 1.5 minutes so don't be afraid to click.

https://youtu.be/ynZSMuSN8qU

@Keith - "Commitment Tracking"/"Dynamic Tasking" does a good job of highlighting the need for 'workload' management'.

Not to be confused with Case-level "workflow management', workload management operates across multiple cases (i..e however many an individual researcher may be working on and an augmented list for a manager who may be the ultimate decision maker for a large number of cases).

I can't see many options here.

The 3-tier RALB (Resource Allocation Leveling and Balancing) model we use sees
1) tasks posted to User InTrays from BPM engines
2) tasks posted by the user himself/herself,
3) other tasks posted by supervisors, by incoming event transactions that launch process fragments, by software in places where operations run 24 x7 and handoffs need to take place at changes of shift.

The way the model works is

I. tasks post,
II. users then micro-schedule these tasks
III. supervisors balance the workload across workers across cases
@John

I read the article "Fernando Flores wants to make you an offer".

I can see his ".... commitment between individuals consists not of pure information, but of prompts for actions" as a sub-culture to any corporate culture that reasonably operates strictly on a "no verbal orders" basis.

"No verbal orders" needs to be widened to include e-mails, any commitments, in fact, that are not contractual in nature (notably contractual penalty clauses).

So, Flores' methodology could work in a flat organization, better in an organization where all of the members are shareholders.

My point may be made clear with a story from my days with Bechtel where meeting contractual commitments had a singular focus.

One day I was sitting in the GM's office when a project engineer, who had been summoned to appear, arrived. The GM reminded this person that the upcoming Friday was a project milestone - the engineer said, "I am sorry but it is going to take a bit longer"

No worries, replied the GM, "take as long as you like, just finish by Friday" and on that note, the engineer was escorted out of the office.
@Karl -- gr8 Bechtel story! Keeping to one's commitments is certainly a good thing; the idea can be taken too far or used as a tool of bullying, the better to hide organizational failure. The whole Winograd and Flores thing, with it's emphasis on commitment, was picked up in a cultish way by the "EST" movement (this is mentioned too in the S+B article". It's a case of "workflow meets California cult". Who knew you could mix the two? : )
  1. John Morris
  2. 3 days ago
Kay Winkler Accepted Answer
Based up on the premise that Adaptive Case Management is an evolutionary step from BPM, naturally everything that makes up a competitive BPM Suite must be present in ACM as well, out of the box.
What has been defining ACM as its very own category for us the most, were the following 2 additional components: a strong Business Rules Engine (structured and ad-hoc, allowing for the "A" in CM) as well as a tightly integrated Enterprise Content Management (ECM) platform, on top of a robust BPMS.
If anyone is interested, we laid out the different ACM components for FATCA (Foreign Account Tax Compliance Act) solutions in the chapter of the book, linked below.
References
  1. https://bpm-books.com/products/best-practices-to-support-knowledge-workers-print
NSI Soluciones - ABPMP PTY
Comment
Business Rules Engine & ECM - I agree both are essential to any Case environment that is hosting background BPM.

Rules:

It's not too difficult to ensure that there is at least one "pending" task at a Case - since tasks only post to the attention of eligible performing resources, the only person likely to have full oversight is the Case Manager.

This is especially true for User "system" (whose tasks do not post to any human InTrays) so oversight by Case Managers involves determining whether a task is pending because of an imposed delay or whether it is sitting in a user InTray (human or system) waiting to be accessed by that user OR blocked from access because of a pre-processing rule that will not fire.

The only truly practical rules we have been able to invent in an ACM environment are pre-processing rules. Once an organization starts to accommodate ACM, that organization can no longer rely on data being available as a result of execution of a set of tasks in a set order.

Accordingly, whenever a user tries to convert a pending task to a current task, we typically need to ensure that pre-requisites have been completed. i.e. if protocol is that no shipment takes place without a QA release, any "ship" task should not engage if the data does not have QA release data value.

ECM:

See my comment under Keith's post - we are not as advanced in the area of free-form searches as we would like to be in our Knowledgebase envronments (free-form searches across Cases of Cases) so things are no better for individual Cases.

In theory, there could be data of major importance on any Case form or in any document attached to a Case.

Despite rigorous effort to automate writeout of a link in the Case History for any document attached to the Case, it seems prudent to allow a free-form search across all of the data and across all of the documents at a Case.

In our Kbase environments where users can see summary data from all Cases and engage free-form searches across all Cases, you can see from the short video that it is tedious to have to convert any document to plain text AND then link to the source.

We need to do a lot of work in the area of accommodating "((a+b) or (c)) and NOT (e)", not to mention "close to" and "sounds like" etc. I don't see any solution other than indexing every word except "a, the, it, etc."

For images, video/audio, the only practical approach to search/find is to use key words. Unfortunately, user mindsets change from the time they invent keywords to later trying to use these, so keyword searches are not very practical in real life situations.
E Scott Menter Accepted Answer
Blog Writer
Expanding on what John said above, case management products should be built atop more traditional BPM platforms to ensure the maximum flexibility for implementers, and the availability of familiar tools.

Not all case management applications are created equal, of course, but often, a specific application is neither pure case management nor pure flow. In my experience, case management apps we build for clients are nearly universally a combination of both. So you really want to be able to share tools and techniques and, most of all, data, to the extent possible.
http://www.bplogix.com/images/icon-x-medium.png Scott
Comment
@Scott... Right .. and the variables for data sharing seem to be a) push vs pull, b) parsers/formatters and c) encryption.

Can you think of any others?

We favor generic data exchangers because they allow sharing of data on a need-to-know basis and they allow both publishers and subscribers to post/read data using their own native data element naming conventions. A side benefit of data element name mapping is a publisher cannot, on his/her own, publish a new data element that causes the receiver's import routine to crash.

The only thing a subscriber has to worry about is that the storage they set up is of a suitable size and type so they don't end up with truncation or rubbish.
1. Case-related artefacts (the whole life cycle) repository; social gadgets are included here.
2. Enterprise-related artefacts (the whole life cycle) repository; Karl’s fragments are here.
3. Work management engine - who did/do/will do what, when, where, with what means, with what results and how.
4. Pluggable to the work management engine coordination techniques (classic flow-charts, decisions, gantt, check-list, etc.); Karl will be happy.
5. Automation scripting (in many languages) outside the work management engine.
6. Pluggable UI

Thanks,
AS
Comment
@Alexander - nice list, thanks
John Morris Accepted Answer
Sales is a kind of case management; so I will add a "must-have" to the case management list. This is "narrative support", a.k.a. "note-taking on steroids". Narrative shows up already of course especially in health care (and some other case-oriented sectors including law). And beyond raw narrative ( which is also known as "story" ), we can add things like "text mining" and "annotation".

(Narrative itself is both human- and machine-generated, including machine-generated from process model instances and from document repositories. And useful narrative exists on multiple levels, including raw original, redacted, enhanced, summarized, annotated etc. etc. "Social" can be considered a subset of narrative; on it's own, and for multiple reasons, social has been to my mind less successful than hoped.)

In the world of sales, relationship narratives are the poor second cousins of "pipeline" and "funnel" and whatnot -- in some of the most popular CRM packages one is hard-pressed to pull together "the story". And yet when making another call, the most important thing you must do prior to that call is "boot the story into your brain" and figure out what the next chapter in the story should look like. (The overlooked business case is astounding; say five big calls a day and twenty minor calls a day and "narrative boot time" at twice the effectiveness in half the time will make a very significant difference in results, all things being equal.) The case for narrative is also strong in healthcare. One can read charts all day and lab results. But the narrative is also an essential part of support for deciding what to do next.

I've written about this elsewhere; human beings "live" in narrative. Where case and business process are concerned, narrative support is an essential part of the whole technology. The technology is rapidly developing; implementation lags far behind. Even note-taking in many products, across product categories, is often frustratingly fragmented. The vendors that get narrative right can really get a jump on the competition. Narrative as first-class citizen of case and BPM process software is one of the final frontiers of software technology. At least before the singularity . . .
Comment

@John. we have had moderate success with machine-generated narratives - one would think they would be great for referrals in health care?

i.e.{{patient}} is a {{years}} old {{gender}} who presented with {{symptoms_signs}}

The pickups know where to find the data elements so there is no need to consolidate all of these one one form to make things easier for the narrative report generator.
David Chassels Accepted Answer
In terms of a Platform it should be a single environment with appropriate architecture to handle the fundamentals of any Case Management (one could argue all business is about a "case";).

The following "engines" should be included;
Process engine orchestrating data from any source as required
Rules recognising business is all about " rules"....and of course compliance always have a say!
State giving real time reports etc of activity empowering people (keeping compliance happy.)
Calculations to manipulate data as required...might include algorithms for complex needs
Audit trail of all activity
UI creation with adaptive capability recognising users their authority and presenting data as required for that instance. The forms should collect data only once even on same form which should automatically populate elsewhere.

In terms of the features of the resultant application building a custom solution with flexibility for change by business is that Adaptive necessity.

Finally but most important before purchase of any such capability do research on just how this is achieved.....
Comment
@David ,,, "Audit Trail" clearly is #1 in the list of "must-haves" (i.e. Case History)

The attributes of the Hx are a) date b)time c) performing resource's "signature" then data, as it was, at the time it was collected, on the form versions that were in service at the time.

Why #1?

Because Case Managers (and delegates), rely (most of the time) on Case Histories to decide what the next intervention at the Case should be (i.e if you don't know where you are coming from, you are somewhat hobbled in terms of deciding forward direction).

Karl. Yes agree and with time should come auto escalation if time critical supporting the simple concept what gets measured gets done...or managers can reallocate work real time
  1. David Chassels
  2. 1 week ago
+1 @David: "Finally but most important before purchase of any such capability do research on just how this is achieved....." --> the implication is that the technology is not mature, and is not a commodity, but that "results may vary" and there is significant risk between rhetoric and performance. The buyer thus has responsibility. Not all buyers are ready for this, and the result of naive buyers and today's technology can be disappointment.
  1. John Morris
  2. 1 week ago
John. Sadly so true about naive buyers...not really their fault as enterprise software has thrived over many decades in complexity and tech speak.....Frankly little help from analysts who have failed end users by not doing real research to push vendors to deliver better value and understanding how business really works as BPM principles have promoted.
  1. David Chassels
  2. 1 week ago
@David ... "one could argue all business is about a "case" " is not a distortion.
Examples:
1. for a supplier, a case can be a customer where we find sub-cases per order placed on a supplier
2. for an insurance company, each claim is likely to be a case, given that the insurance company does not routinely expect multiple claims from the same insured
3. for a hospital, a patient is likely to be the case
4. for a researcher, publication of a scientific paper focusing on a particular area of research would typically be a case.
John.. "significant risk between rhetoric and performance".

I feel sorry for buyers - few are able to separate bafflegab from practical reality. Looking at a product's features is unrewarding. Looking for a product that satisfies current needs exposes the buyer to having to shift horses down the line. What they need is something that can address unanticipated future needs.

A reasonable compromise is to call the vendor ask for a demo and then at the demo, tell the vendor you want control over the mouse / keyboard and that you want to build a small yet representative workflow, compile it and run it, within a 30 min timeframe. The desired outcome is you have to ask a few questions but otherwise, you are able to do it.

We find these "cake bakes" give prospective customers a notion of confidence that they will be able to build, manage and own their own processes. We always raise the caveat that they WILL need help with rule sets (even the most trivial of these require being able to walk and chew gum). And we tell them that interoperability is a 2nd hurdle they are likely to need help with.

We did a profiling system for FBI agents to use in the management of critical hostage, barricade, and suicide incidents. I figured 2 hours for the training but with groups of 20 agents at each sitting, we found it took 2 days to get them to where they could run the app on their own.

I guess when a person figures they can do something, they think it is easy, but oftentimes it is not so easy for a new player.

Bogdan Nafornita Accepted Answer
I simply wish the concepts of highly structured and loosely structured flows (or case vs workflow) were buried forever in vendor speak cemetery.

I hope the business platforms of the future simply do not even mention these in their approaches to customers - because it simply does not matter to them. I have yet to encounter a customer (not yet intoxicated by vendor speak) that actually cares if they use case or process to solve their business problems.

(specifically @Keith: your list applies equally well to workflow systems too - why not?)
Managing Founder, profluo.com
Comment
@Bogdan - there is an easy way to ensure that no user ". . . . .actually cares if they use case or process to solve their business problems".

Just declare an ad hoc intervention at a Case to be a "process of one step", and ad hoc interventions become no different from "process fragments" or "end-to-end processes".

Then, implement all processes at a Services Menu as checkboxes and the only option any user has is to go to the Menu to engage some processing.

The menu can be made to post once when a Case record is set up, thereafter the menu can be re-called on demand at any time, and, if it is important to keep a Case record in the foreground, an option can be set such that when a Case has NO pending steps along any of its "pathways", we force the menu to post. (i.e. what do we do now?)

So long as you include one menu option called "none of the above" in the Menu, the Menu can handle 100% of any user's needs for recording activity / requesting activity at a Case.
@Karl - it's already been done, game over. BPMN allows for ad-hoc constructs and CMMN allows for structured sequence of steps.
My point is that, ultimately, the customer is not aware of the difference and he shouldn't.
My less sophisticated (but always very fresh and sharp) customers tend to explain the business problem completely ignoring vendor domains (thank God!) - any solution they require includes a bit of process / case / content / document / records management, UX and business intelligence. And we offer that without pointing to the customer how we integrate those domains in front of the customer - they simply don't care.
We sorta care, internally, because the smarter we are about the integrations, the more reuseable our platform is.
  1. Bogdan Nafornita
  2. 1 week ago
@Bogdan.. When you say "it's already been done" for ad hoc interventions in BPMN are you saying that within the runtime environment a user can without any reference to any of the library of processes, call any of 200 available forms, record data and then simply commit that ad hoc insertion?

Ordinary users would not fancy having to go to the mapping environment to add a new process of one step. That does qualify as ad hoc mapping but is quite different from run-time insertion of a task..

The happy scenario is no more complexity than having a to-do list on a piece of paper where, given a pencil/pen, you just add one more item to the bottom of the list.
@Bogdan..

And, the smarter a vendor is about integration, the less time required to put together a set of processes for any one customer.

We are able to get our customers to build their own workflows but NOT (with minor exception) to get them to build rule sets on their own. As Senor Wences said, "easy for me, difficult for you" - most customers have to get us or a consultant to build the rule sets they need.

In the attached, it would be hard to think of an implementation with a less complicated rule set but, as you can see, looking at the entire build-run-execute cycle, it is procedural and tedious.

https://youtu.be/2vMeZSq3cAI

Less tedious though, than at first sight, because once you have built a "needs more info/does not need more info" type of branching decision box, you can put this in your toolkit and avoid having ever to rebuild one of these from scratch.

Easy to see the value of a front end menu if I have say 1,000 processes as the unpleasant alternative is to expect the customer scroll the list to find the one they want/need to run. With a menu, the problem of finding a process template goes away completely.

Not in the demo, but important, are pre-processors at the 1st step in any process fragment - in the absence of pre-processors to block engagement of process fragments that need data from other process fragments, threading processes together becomes difficult.

It took me 15 minutes to build-compile and roll out the PTSD demo process.

The forms in the demo are bare bone forms and not representative of real forms that typically need 50 form controls per form, with labels, of course, requiring a half hour to one hour for layout.

Bottom line, this little e-learning app could take an entire morning to build but it could then be expanded in one day to deliver content relating to 1,000 topics.
  1. one week ago
  2. BPM Discussions
  3. # 8
Here's the deal - only one working day before this Forum moves on to a new discussion topic so it would be nice to get some more participants at this "must have" features for Case question.

I can't see any point everyone NOT moving to Case when it's crystal clear that none of the benefits of BPM are lost when you move to Case and that a ACM+BPM capability is better than either alone.

I have an upcoming interesting application in the area of infrastructure protection where the objective is to do NO work, other than take evasive action when a remote device (sonar, radar, vibration detection, face recognition) alarm sounds.

There ARE procedural steps that have to be taken when an alarm sounds but otherwise the entire system is totally automated and we clearly need to keep the Case open and active 24 x 7.

The "command center" needs to know what is going on so that rules can invoke specific processing but the thing is some of the signals take longer to process than others

The rules need to rely on what happened in strict time sequence, based on the start of detection (e.g. at time t00 a person walks up to a face detection camera and the FR software starts to look up the image - it can take 2 mins; at t05 we get a vibration detection alarm at a fence that comes in immediately.)

We don't want the rule set to assume that the possibly same perpetrator first tried the fence and then approached the camera because the exact opposite is how the Case is unfolding,

I am hoping we will learn something from this (seems to me like an automated process control application) and add a couple of features to our BPM implementation.

Any ideas?
Comment
Hi @Karl. I've share this item on alarm fatigue before:
https://www.linkedin.com/pulse/20140730013817-1524359-what-manufacturing-iot-project-leaders-can-learn-from-healthcare
But the literature on this (e.g. "alarm flood" and "event modeling" etc. etc.) seems to be completely applicable. Modeling events for any given environment is expensive; one has to fund the business analysis. And one has to fund the business analysis on an on-going basis. This is a business challenge. My rule is "data is cheap and business analysis is expensive", the result being that alarm overload is almost inevitable, without all-in on-going senior management commitment.
Example: For periodic field device maintenance, you want to be able to turn off alarms so as to prevent false positives. But then you also don't want to trigger cascading "device failure warnings" at the same time. Except there may be exceptions to this rule - lots of movie plots are predicated on tricking authorities into disabling alarms. Lots of complexity.
Does this make sense for your situation?
  1. John Morris
  2. 3 days ago
@John. I guess the only way to proceed is going to be to write up various scenarios all starting with 'Three robots walk into a bar . . . ."
David Chassels Accepted Answer
Karl. sounds like a configurator where rules automatically change as previous actions determine next steps.....just think options step by step and you should get there....
Comment

Hopefully, we won't have to change rules on the fly, the rule sets can be put through their full range of governance if we import a sufficient number of different scenarios.

Option lists, step by step can help but not in terms of identifying the last option one should have thought of.

Perhaps the right approach needs to be:

White Rabbit: "Where shall I begin, please your Majesty?"
King: "Begin at the beginning and go on till you come to the end"
Thanks to all contributors to this question.


My original list of "must haves" for Case was . .

1. an environment capable of hosting any number of Cases, with any mix, within any one Case, of ad hoc vs structured steps. (implies background orchestration capability: implies a cross-case workload management capability: implies background governance that can help to rein in, extreme, unwanted, deviation away from corporate policy/procedure/practice.)
2. a non-subjective means of assessing progress toward attaining Case goals/objectives.
3. interoperability across multiple local and remote systems, applications, devices.
4. a user interface that makes it easier for most users to use the environment as opposed to not use the environment.
References
  1. https://kwkeirstead.wordpress.com/
Comment
  • Page :
  • 1


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