BPM.com
  1. Peter Schooff
  2. BPM Discussions
  3. Thursday, 12 July 2018
  4.  Subscribe via email
To build on the excellent discussion last week on How Do You Make Sure Automating a Process Doesn't Result in a Worse Outcome? what would you say is the key to managing automated processes?
Max Young Accepted Answer Pending Moderation
Blog Writer
A willingness to be led by the evidence. Amazingly simple, but amazingly hard to execute. You have to see the effects of the changes, compare the effects of the changes against your values for the domain, and be honest with yourself about the implications of the results.
References
  1. http://www.capbpm.com
Comment
@Max your comment is also "amazingly succinct" -- but it points to a powerful trend to evidence-based management. "Evidence-based medicine" is a good example. And for sure managing from an evidence-based perspective can "hard to execute". The challenge is not only technical, but also social. Managers need the humility to sometimes give up subjective or pre-conceived ideas, in the face of evidence. (There are many examples in the world of business process -- one can think of process mining as "all about the evidence".)
  1. John Morris
  2. 2 months ago
  1. more than a month ago
  2. BPM Discussions
  3. # 1
Tim Stephenson Accepted Answer Pending Moderation
Instrumentation. Both for KPIs so you _know_ it's not getting worse and to provide sufficient transparency for when things go wrong and you need to diagnose them.
Comment
Get the automation right in the first place.
  1. Fred Nickols
  2. 2 months ago
  1. more than a month ago
  2. BPM Discussions
  3. # 2
Karl Walter Keirstead Accepted Answer Pending Moderation
What is the key to managing automated processes?

Response:

Precisely the same as the way you manage manual processes - via QA.

On engagement, to make sure you have what is required as inputs to the automated process, that the data is not inconsistent, not out of range. etc.

Then, on exit, did the process perform as expected?

In cases where the processing goes on and on, QA needs key process points within the automated process.

This is why pre-processors were invented and why post-processors were invented (Dr Bernard Meyer - 1986,"Design by Contract", the Eiffel programming language).

https://en.wikipedia.org/wiki/Design_by_contract

Eiffel had/has the most elegant implementation of pre/post-processors, in my view.
References
  1. http://www.kwkeirstead.wordpress.com
Comment
"Precisely the same as the way you manage manual processes" - much better than you manage manual processes, because automated processes are highly risky as they can create "bigger" damage much "faster".
  1. Dr Alexander Samarin
  2. 2 months ago

@Alexander - True, but all of this parallels to an extent the design approach used in the area of industrial automated processes - you have to assess the risk and put in place checks and balances to the point where your automated processes get suspended/shut down quickly if your process control points detect errors at steps.

Every initiative any of us gets involved with or considers getting involved with has risk and uncertainty
  1. Karl Walter Keirstead
  2. 2 months ago
  1. more than a month ago
  2. BPM Discussions
  3. # 3
Dr Alexander Samarin Accepted Answer Pending Moderation
To manage automated processes you have to know how many and how some manual activities must complement those automated processes.

Thanks,
AS
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 4
David Chassels Accepted Answer Pending Moderation
In terms of business processes there are a lot of routine tasks that can be automated such as creation of letters, emails, collection of data from legacy, reports on activity, escalations etc. Then there are what I would call a sub process which could be automated which link into the end to end process. Of course "management" of all activity is important and real time reporting will not only help to empower users but ensure "issues" are quickly identified and addressed including where automation is implemented. As ever just how these requirements can be achieved is important before embarking upon a project which should give business confidence they are still in control of such important end to end operational processes.
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 5
Kai Laamanen Accepted Answer Pending Moderation
Hardly ever it is possible to automate whole Business process... as David noted we may automate some routine tasks by robotics or support human activity by AI approach. As in normal IT-programming the design or management issue is how mitigate the risk of failure or misuse of the automated process. In human activity we often notice if things go badly wrong and make corrective action.... Automated processes often do not have such capability.

From improvement perspective automated processes may be easier to develop data driven. By nature automated systems gather data which can be analyzed and interpreted. And if the automated system is not too complex, it might be easy to make small scale experiments and do changes based on the effects. The change management is easy (compared human) the technical systems do not resist changes.

My last point is about to automate or not to automate.... all data driven processes can and will be automated sooner or later. If the process has tasks including value based assessment or decision making, I would not recommend to automate it.

br. Kai
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 6
Kay Winkler Accepted Answer Pending Moderation
The definition of process KPIs is one of the key ingredients for continued improvements and process management, which in itself usually poses a greater challenge than its measurement. Since the process contains data sets of both, business data (e.g. sales volumes) as well as process data (e.g. case worked time tracking), the definition of comprehensible indicators can turn complex very quickly.
The other key ingredient would be, in my opinion, the definition of ownership in order to make the statistical process output, in fact, actionable. More often than not, the lack of a formally established and enabled process CoE is what hinders an effective process management and not the in-existence of indicators.
NSI Soluciones - ABPMP PTY
Comment


I can see value of KPIs for end-to-end processes. i.e. taking an example from the industrial sector, how is my cement plant functioning ? (i.e. one automated process, direct impact on the organization).

If we try to consider KPIs at all possible levels within a business (not a totally automated process), the initial impression is we could have 3 sets of KPIs:

a) Strategy level across all initiatives (i.e how am I building, sustaining and augmenting competitive advantage?)
b) Case level at each Case (i.e. how is this Case progressing /how did it progress to its completion?)
c) Processes within Cases and across Cases ( i.e. how does this process contribute to the advancement of this Case?; how does this process contribute in general across Cases where the process was engaged?).

The problem with b) is that we have tracking/measurement in the form of goals/objectives. No need for KPIs. No need to graph the progress timewise either, we know it has the shape of an "S" curve.

The problem with c) is most Cases (other than end-to-end, comprise a mix of multiple process fragments and ad hoc interventions) - we can't easily measure the progress of one process toward meeting the objectives of a Case or multiple Cases.
  1. Karl Walter Keirstead
  2. 2 months ago
+1 "comprehensible indicators" and "actionable". This could be a whole dialogue on the topic of how business process management environments support decision making. Which is of course what management is "for". "Comprehensibility" exactly corresponds to "decidability" which corresponds to "actionable"...
  1. John Morris
  2. 2 months ago
@John. Sure, there is room for a "whole dialogue"

My take is "business process management environments" host two types of "Cases" 1) ongoing initiatives, an example of which is repeat orders from a customer within a customer entity (i.e. the Case stays open) and 2) initiatives where once Case goals/objectives have been met, the Case closes.

The "decision making" revolves around a) is the work contributing to advancement of the Case ? and b) are the goals/objectives set at the Case still valid in terms of their contribution to building, sustaining and augmenting competitive advantage?

At the strategy level, a KPI is typically a consolidation of many Cases (not to exclude a 1:1 relationship between an important initiative and a specific KPI), so, "comprehensible indicators" are often a problem.

The question becomes which of many initiatives/Cases are causing a KPI to not track in the desired direction?
  1. Karl Walter Keirstead
  2. 2 months ago
Great inputs, both. Not necessarily a solution - but maybe an approach, at least, could be taking a hard look at the so called "strategic KPI's" and make sure that they are in fact numeric, measurable expressions of something specific. (Example: We want to be the #1 in BPM has to be on a strategy level translatable to something concrete, from the get-go. For instance, we want to be the number 1 in BPM by being a provider to at least 80% of all banks in the US. until 2019)
  1. Kay Winkler
  2. 1 month ago
  1. more than a month ago
  2. BPM Discussions
  3. # 7
Boris Zinchenko Accepted Answer Pending Moderation
Due to explosive growth in robotics and AI automated processes are quickly becoming an essential part of business. Wide adoption of automated processes creates significant challenges for business management.

Automated processes are both simpler and more complex to manage, compared to manual processes. Simpler, because automation is void of any personal attitude or bias of human workers. More complex, exactly for the same reason of absolute impersonation and concentrated responsibility.

In case of an error in manual process there is always a space for an additional control and adjustment of task by a worker. In contrast, automated tasks run autonomously and on far higher speeds. It makes a mission of controlling automated tasks far more stressful and dangerous. Any occasional error on a single process step can randomly propagate the whole running process chain causing catastrophic damages for the whole business.

These elevated control requirements for automated processes make it impossible to run wide scale enterprise automation without a detailed business model in place. Professional BPM and consistent EA are cornerstones in the success of process automation in any organization.
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 8
Ian Gotts Accepted Answer Pending Moderation
The critical part of managing automation is good documentation. You can't change it with confidence if you don't understand what it does and for whom and why. This an increasing problem as low-code platforms make it easier and easier to create automation quickly.

I can't remember what I had for lunch last week, let alone how I configured the workflow last month. Documentation is business process diagrams (in whatever format - photo of whiteboard even), requirements, specifications, and training material. This needs to be stored/linked in the context of the automation. It needs to be attached at the time it is created. Not after the project is completed - because that never happens.

Only then can you evolve and change it with confidence.
Comment
Agree but even better when the process diagrams are actually the live application with transparency for all to see just how the end to end process including automation is actually working.
  1. David Chassels
  2. 2 months ago
  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.