Adaptive Case Management (ACM) Adaptive business management

Adaptive Case Management (ACM) is the concept alternative to the concept of BPM. Alternative — it means that the same problem of the effective organization of business processes, but other means is solved. If it is absolutely short, then the ideological difference in approaches can be characterized as follows:

BPM goes from structure of process, structure of steps which need to be made to achieve the goal. Data can arise during process, being completely avisimy from it. ACM goes from information and business data arising in a work progress and necessary to consider result reached. Processes arise in the context of data, and not vice versa.

File:Адаптивный кейс-менеджмент.jpg

1. ACM place among other technologies

Let's begin with the fact that BPM, Project Management (PM), ACM and some other managerial technologies, are applied to management of organization activity.

To show how ACM is built in practical activities and what distinctive features of this technology in practical application, I will use very simple model of activity in which are selected:

• goal;
• the action directed to achievement of this purpose;
• the contractor making this action.

Such reduced model is necessary not to eclipse parts key properties of the considered technology.

Further, we will consider the following properties of each element of this model: recurrence (as possibility of repeated reuse of the same descriptions and models) and predictability (as possibility of preliminary planning).

Here by the signs '+' and '-' properties which would be preferable are mentioned or, respectively, are undesirable to this technology. He is familiar'?' properties are mentioned, existence or lack of which affects technology a little.

In this table there is one unevident moment. Almost in all sources the success of fight against unpredictability is considered the key advantage of ACM, but affirms as the table that the ACM technology is good only at the predictable purposes and moreover — when the purposes are rather permanent and repeat from a case to a case. I will prove the point of view by the proof from the return. For this purpose we will consider what would lead systematically unpredictable purposes to:

When the purposes change, also requirements to result of activity change. It means that statistics of activities of users which an ACM system could accumulate will not be able to prompt any more what actions approach this situation. Compare, for example, to project management where unpredictability of project objectives is compensated by relative predictability of the planned actions and contractors. Well, let automation in the conditions of complete unpredictability is at a loss. Then it would be possible to try to make something at the organizational level. But also here problems: in regulations it will not be possible to record either process, or its result, nor even its purposes. And regularly to write regulations under each again appearing purpose is the level of technologies at which in ACM do not even stammer. Further, we will put that it is possible and to refuse regulations, hoping only for an initiative of non-manual workers (knowledge workers). But who will be able to show at least one company numbering more than ten people where employees are given free rain in goal-setting, the choice of means and methods of work? End of the proof.

Summarizing the aforesaid: each of the listed technologies requires predictability though in something. For ACM — it is the case purpose.

However we will return to the 'total unpredictability' covering including, the purposes of business processes. What should be technology of management in such conditions? The answer to this question including justification to me occurred in the book Novikov A.M., Novikov D.A. 'methodology' — for the unpredictable purposes is most adequate project management style. But at the same time it will be required to prepare for reorganizations of the plan and change of a command already during the project.

One more foreshortening, to complete the picture. In terms of project management 'total unpredictability' is a sign of projects of high complexity for which key questions are not technologies of automation at all, and … leadership skills and will of the management. However, the question of projects of high complexity is well opened in a series of articles of Sergey Churyumov.

2. Extremely specific example

There is a process: processing of the requests arriving via the website. The first process step — to define for each request whether it is necessary to accept it or to reject. After all automatic checks, for control by the operator remains 'only' four conditions. Task: prepare the instruction for the operator on order taking.

It would seem, at a look on top, everything simply. But at immersion in a part 'troubles" begin:

• Sometimes the decision can be made after check of one-two conditions, and the others are not required to be checked any more.
• In different cases the different sequence of check of conditions is convenient, strictly it is impossible to fix the sequence.
• In certain cases a part of conditions can be checked only with the address to the expert, but it is desirable to avoid it needlessly.

As a result, the optimal algorithm (and, respectively, the optimal scheme of this process step) contains a lot of branchings and about ten different statuses. And all this only for one process step of processing of the request expected accomplishment within several minutes of real time!

Both the verbal algorithm, and the flowchart do not help the operator to understand plainly as after all it is necessary to distinguish correct requests from incorrect correctly. In the same way they do not help to control to the operator the work as do not contain visually formulated requirements to result.

Let's look what will change if not to try to describe process, and to try to describe result which is required to be received?

If to write out all possible combinations of results of check of the ill-fated four conditions, to mark each option with sign 'Accept' or 'Reject', tabulate options and sort them by the most often checked conditions, then 'check sheet', ready to the use will turn out … (conditions are depersonalized as it is excerpt from the document chipboard):

Such table needs only to be supplemented with requirements and recommendations which are formulated by the linear list and it is extremely simple as there is no need any more to describe branchings. Here, unlike tangled schemes, the operator will rather easily understand that from him is required, and itself the style of work will find the most convenient for itself. And what is important too, in complex cases the operator will always be able quickly to recheck himself according to such table.

Already on this micro example, it is well visible as far as the result changes (in this case — the organization of work of the operator) if to move attention focus from structure of process on structure of the processed information. Also on this example it is visible that application of the 'latest' concept, in practice can lead to already known solutions. As they say, 'all new — well forgotten old'.

Also here it is possible to mention one more example of ACM of the instruction constructed by the principles published by Anatoly Yumashev and also a remarkable post of Maxim Smirnov one implementation of business processes the check sheets. Though, in my opinion, the control list of actions is not only of possible implementations the check sheets, but it is described very well.

3. A look from the regulating documentation

Further, I suggest to rise from the microlevel of the organization of business processes by its top and to estimate what conclusions can be drawn, looking at all system of the regulating documentation through a prism of the ideas of ACM?

Naturally, the aspects connected with variability of business processes and variability of their components are most interesting here. On analysis results of the systems of corporate standards of a number of the companies, it was succeeded to select the following five aspects:

• Redistribution of functions and powers. When at rather stable processes in the essence contractors, controllers and also responsible for result change.
• Influence of data on the content of process. When the slightest changes in data or in requirements to their processing can change process to unrecognizability.
• Existence of end-to-end rules of global action for all processes — the politician. When change of one rule leads to change of a number (and even all) business processes.
• Variability of implementation of regulations in different divisions. When in branches or affiliated enterprises of holding the same procedure adapts to local conditions that leads to growth of complexity of implementation of changes.
• Dimensions and regularity of updates. When the dynamism and scales of business lead to the fact that even acquaintance with changes of all divisions is an uncommon task, without saying about supporting changes at the level of automation systems.

How to cope with the listed difficulties in automation systems, depends on subtleties of implementation of a specific system. And depends even more, than on to what class formally this system belongs. Therefore here I will stop only on some most general recommendations and in advance I will apologize if someone considers them insufficiently specific.

• Estimate the automation system — what predictability she expects a component of business processes.
• Estimate how automated business processes are steady against changes.
• Split up schemes of processes and aim at their reuse. The less scheme — the is less than chances that it will change at the next updating of regulations.
• Reveal a variable part of procedures and take out it to the area of setup which users, and not just business analysts and programmers could execute.
• Calculations of business rules should be expressed in business and significant terms to improve their reuse and a soprovozhdayemost.
• The exact specification of business and significant parameters of business processes with which should satisfy implementations of processes in a system. It simplifies processes of changes.
• There where it is justified, to formulate regulations not in the form of schemes of processes, and in a type the check sheets.

4. A look from the current activity of employees

The regulating documentation gives an idea of the accepted activity model in this organization. But the model of processes is not the same that becomes actually. I think, situations when regulations disperse from really operating practice are familiar to all or do not disclose its essential aspects. Therefore I suggest to look with the next step what valuable can be taken from 'digital traces' of activity — magazines of audit of the actual actions of users.

Some time ago in the post Achilles heel of process I brought up a question of search of alternatives to process management, so far as concerns unstable processes. And, as by request, soon there was ACM positioned as such alternative. The main idea then was — to reveal steady components in unstable processes and to find a method to automate a steady part of processes. Now it is possible to formulate this idea more specifically:

• reveal templates of custom actions and a condition at which these operations are performed.
• teach to support automation systems these templates.

Solving the first problem, it is possible to offer the following technology of the analysis of statistics of custom actions. Each recorded event of initiation of a task is followed by a set of attributes among which three important groups are selected:

• The initiator and his place in organizational structure of the enterprise. Gives the answer to a question: who needs a template?
• Object of a system and its property from which the task is initiated. Answers a question: in what place of a system initiation of a task of a template is necessary?
• Structure of a task. Answers a question: what situations can this template cover?

All array of information on events is analyzed in three listed cuts. At the same time the order in which there is an immersion in data affects the received results. I will give several examples:

• The structure a task/investment — where would be useful the button on fast start of a task?
• Structure tasks/investments/contractors — whether it is possible to fill a task route at once?
• The structure the task/investment/initiator — to whom to show this button?
• Initiator/structure of a task — to a lump we can help with elimination of a routine?
• Structure the task/initiator — to whom will be interesting this template?

Outputs which can be received using such analysis:

• Give reasonable offers on development of functionality and, especially, user interface
• Reveal the groups of persons interested in changes
• The similar analysis can be carried out both by vendors for standard solutions, and integrators for the purposes of specific projects at the customers.

Conclusion

I tried to present the subject ACM from four different points of view in article. It is natural that the results received in each case externally considerably differ from each other. It means that the concept of ACM is rather uncommon.

At first sight can seem that the given examples and outputs are far from the ideas and concepts stated in the marketing texts devoted to the subject ACM. But it is those results which turn out in practice at consecutive use of these ideas to already settled systems. It more than once occurred, and will probably happen to many good ideas still more than once.

• Source - Flexible processes