Frank Michael Kraft's Blog

Q: What is the relationship between ACM and BPM? Is ACM a kind of BPM, or disjoint from it, and why?

“BPM refers to a set of management disciplines that accelerate effective business process improvement by blending incremental and transformative methods. BPM’s management practices provide for governance of a business process environment towards the goal of improving agility and operational performance. BPM is a structured approach that employs methods, policies, metrics, management practices and software tools to manage and continuously optimize an organization’s activities and processes.” is one definition of BPM (Gartner: Cantara & Hill, 2008).

In the sense that BPM and ACM both have the goal to improve operational performance and to optimize the activities and processes within an organization, they are similar. Both offer process modeling, process governance and software tools.

But ACM is so different from BPM, that we, the protagonists of the book “Mastering the Unpredictable” say, that it is not BPM and that BPM is not ACM.

The ultimate goal of BPM is to define repeatable processes in a way that they can be managed, measured and optimized. This is a good goal, if the process is repeatable. Not repeatable processes are not the primary focus of BPM. ACM targets mainly not repeatable processes. Furthermore the approach of BPM typically is Top-Down whereas the approach of ACM typically is Bottom-Up.

Q: What is the key difference between BPM and ACM

The key difference is this:

In BPM the process must be defined completely before it can be managed, in IT terms it must be modeled completely before it can be executed within a Business Process Management System. In ACM a process emerges while it is being executed, and at the same time is managed.

Q: Describe how process modeling for ACM might differ from BPM. Will BPMN be useful for ACM?

I would not necessarily talk about process modeling within ACM. I would use the term process definition.

Process modeling for most people implies a graphical process model, just like with BPMN. But in ACM there is not necessarily a graphical process model. There may be for some parts or snippets of a process. Some graphical process renderings may be read-only, while the process definition is done in the form of something similar like a task-list. Some ACM process models may be graphical and editable. There certainly will be process analytics. But the emphasis in ACM is on process execution, not on process modeling. However different forms of process models will most probably emerge, that differ from BPMN.

I believe the way how ACM processes are defined should be designed independent of BPMN. Modeling skills should not be necessary to use an ACM system. It must be very intuitive. It must and will be as simple as writing down a task list or writing bullet points in a slide presentation or filling a spreadsheet.

However I think it will still be possible to export ACM processes to BPMN. If it will be possible to import BPMN to ACM I do not know. I have doubts. It is not the best use case in my opinion, because immature or rarely repeatable processes will first be defined within ACM and not as a BPMN model, and for me it does not make much sense to execute a BPMN process within an ACM system. But in the long run I may be surprised.

Q: What is an emergent process and what effect is there on the system?

An emergent process is a process whose form unfolds/emerges while it is being executed. It is the opposite of a designed process, because the form of a designed process is defined before it is executed. An emergent process starts with some definitions of goals and some first workitems. Then the process is started to be executed. Later in the process execution, after the first results have been achieved, the process is changed, so that additional/more detailed goals and/or workitems are defined, removed, prioritized higher or lower and/or planned, assign a different process owner or process members, new deadlines. Furthermore emergent processes may be split/merged while they are already in execution. Still the workitems are not independent from each other, because they are connected within the emergent process and contribute towards a common goal. So taking all measures to ensure that a process is still able to reach its goal even under changed circumstances is what is necessary here. We call this to adapt the process or the case.

The effect on the system is that there must be no distinction between a process design time and a process runtime. The design time is the runtime and the runtime is the design time. Any process system that distinguishes process design time and process runtime is not fit for this task. The process system must be designed from the beginning to be adaptive. Furthermore it is not sufficient to plan with tasks and relationship between tasks, because this process language is not flexible enough.

Q. How can work be unpredictable? How do you handle that?

Work can be unpredictable, because it is not possible to predict which results the first steps of the work will yield and succeeding steps depend on the outcome of preceding steps. Theoretically one could model all possible outcomes and all possible reactions to all outcomes – at least try to. This is modeling a scenario like for example a military battle scenario. However this investment is not economical in most real life situations and processes. Therefore it is often better to define a goal on a high level and leave the detailed planning of the steps that lead to the high level goal until later and/or to different experts. Modeling a complete scenario like modeling a military battle scenario still has many unpredictable components like unexpected events, ambushes, different than expected landscape, enemy troop moves etc. Generally spoken the sources of unpredictability in my opinion are these:

• Decisions of individuals that cannot be predicted
• Complexity of a situation, that cannot be analyzed completely a priori and
• Unpredicted events like an economic of financial crisis, rapid change in politics and other similar unpredicted events – even natural events.
• Limitations in individual knowledge.

The handling of the situation must be to perform a new evaluation of currently planned measures in the light of their contribution to a previously defined goal. This means to adapt the process as already mentioned above.

In my experience limitations of individual knowledge is the most common reason for unpredictability. In practical terms work is unpredictable, if there is noone, that knows all the details, even if they could – theoretically – be known. And as we all know this is how it is. We all know only parts. If we want to make progress quite naturally we want and need to explore previously inkown areas. To us it makes no difference if these areas are only unkown to ourselves or generally unkown. The effect is the same: We cannot plan the details ahead.

For example I want to write a software program and I want to integrate a certain functionality. I may have 2-3 modules, that could do the job, but I do not know which fits best, until I just try it. After I have tried, I know. Then I can decide and plan all the succeeding steps.

I have got some questions that are often asked about Adaptive Case Management. I want to give my opinion about them.

Q: Why is it such a hot topic right now to discuss alternatives to BPM?

It is the right time to discuss, if classical BPM has delivered on the promises. BPM is now around for some while and has been applied in many projects. BPM technology has been used in practice. It is a natural process to ask: what have we learned from it? Where did BPM work and where did it not work as expected? What are the reasons? Which basic BPM assumptions are there, which are limitations that should be questioned?

It turns out, that one basic BPM assumption was that work is to some degree repetitive. And some work is. BPM proved successful within processes which are repetitive and the profit of BPM is the economy of scale – i.e. the more often a BPM process is executed, the more profit comes from it. But on the other hand it becomes obvious, that a growing percentage of the work in today’s industry is not repetitive – especially knowledge work is not. Furthermore unpredicted events lead to discontinuation of previously repetitive work and the need to manage the unpredictable. Therefore alternatives to BPM should be discussed right now.

Q: What is knowledge work, and why does it require a new approach?

A knowledge worker is a professional – that is my personal definition – whose flow of work depends on decisions based on knowledge, which he or persons of his professional network create, collect and distribute. Therefore by nature it is not predictable a priori, because succeeding work depends on the outcome of preceding work. It is a chain of processing knowledge, decide based on the acquired knowledge and process knowledge again. The work is explorative, not prescribed. A new approach is needed, because classical BPM assumes, that a flow of work can be designed a priori and executed later.

Q: How important is knowledge work?

The number of knowledge workers is constantly increasing. According to Thomas H. Davenport (Thinking for a Living) it was about a quarter to half of the workforce of the U.S. in 2005. In my opinion it is decisive for the developed countries, to make knowledge work more effective in order to stay competitive.

In my personal opinion it is no question at all, that there is a overwhelming amount of knowledge work which can and must be made more effective. In my opinion the problem so far was and still is that the technology needed to make it more effective was and is not available yet. As soon as the technology is available, it will certainly be needed.

Q: What is a specific example of the kind of knowledge work that might be supported?

A specific example is described in my chapter “Improving Knowledge Work” in the book “Mastering the Unpredictable”. There is Leona who works for a telecommunications company as an engineer and she needs to do phone support. The work she does in the support area is described with examples, as customer complaints need to be solved. Some tests need to be executed and some countermeasures need to be taken. The work is unpredictable, because the tests and the countermeasures depend on the situation. However the work can still be supported with Adaptive Case Management.

Other exampled mentioned by Davenport are:

• Management
• Business and financial operations
• Computer and mathematical
• Architecture and engineering
• Life, physical and social scientists
• Legal
• Healthcare practitioners
• Community and social services
• Education and Training

I will discuss more questions in other posts.