Using AI Cost-Effectively: Analysing Processes with Value Analysis

Artificial intelligence can speed up processes, prepare information, generate content and support decisions.

But a fast automated process is not automatically a good process.

If an organisation merely digitalises existing workflows or supplements them with AI, unnecessary steps, superfluous interfaces and rules that have grown up over time may remain in place.

The economically more interesting question is therefore often:

Which functions does the process actually have to perform – and which of them could be performed differently, more simply or automatically in future?

Don’t Confuse a Process with an Activity

A process consists of numerous activities. But not every activity is the actual purpose.

A function-oriented view abstracts from the way things are done today and examines, for example:

What information needs to be provided?
What decision needs to be made?
Which check is really necessary?
What documentation is required?
Which interface actually creates value?

This can reveal that an activity exists today only because an earlier system or an earlier organisational structure made it necessary.

AI as a Solution Principle, Not a Starting Point

If the analysis starts with the tool, the question quickly becomes:

“Where can we use AI?”

Methodologically, a stronger question is often:

“Which functions of the process must be performed – and which solution principles are suitable for them?”

AI is then one possible solution alongside standardisation, elimination, organisational change, conventional automation or better data integration.

Considering Effort and Benefit Together

Even with AI applications, technical feasibility is not enough.

Factors to consider include:

expected benefit,
process costs,
implementation effort,
data availability,
quality and error risks,
necessary human oversight,
integration effort,
ongoing operating and maintenance costs.

Value Analysis and Value Management provide a framework for considering benefit, functions and effort together.

A Possible Project Sequence

clarify the process objective and stakeholders
identify the customers or recipients of the process output
describe the process functions
analyse current effort and bottlenecks
critically challenge functions and activities
develop alternative solution principles
examine AI applications where they can make a clear contribution
evaluate concepts in economic and organisational terms
pilot and measure the impact

Value Analysis for Processes Too

Value Analysis is often associated with technical products. However, the function-oriented approach can also be applied to processes and services.

In AI projects in particular, this abstraction can help not just to speed up individual activities but to fundamentally redesign the process.

Using AI Cost-Effectively: Understand the Process First, Then Automate