Organisations as information processing systems

Hierarchical information structures are an essential concept of complex life. But (except for some exceptions) it doesn't work by granting power over others to some parts of the system. Hierarchy typically happens when agents in an underlying layer get incorporated into an emergent higher layer in the system. What happens is that living agents are assembled in a hierarchical information processing system.

To do this, life uses signal/boundary systems. In it, a semi-permeable boundary is used to selectively amplify meaningful signals and dampen meaningless signals, shuttling meaningful messages to the appropriate subsystems. This mechanism helps organisms to make sense of their environment, and to perpetuate the conditions for the continued survival of the system as a whole.

A key innovation in human society, was the development of the management information processing system. With it organisations were able to grow much larger through the implementation of hierarchical information structures implemented as management pyramids.

Command and control hierarchies and later management pyramids were a key innovation, that allowed organisations to evolve into modern organisations. But even as the management scope changed, the essence of the information hierarchy remained the same with humans acting as information relay, information amplifier, and information processor nodes. On an organisational level the believes, knowledge, roles, and objectives of these people and the groups of people that they manage, and how these groups are connected with each other, act as a network that is trained to process certain types of information.

The rise of information technology is now changing this system in multiple ways:

• Increases in speed allowed larger organisations to remain coherent over larger territories.
• Decreases in message costs have increased information flows.
• The development of a global near instantaneous information network has shifted the bottleneck in information processing from gaining access to meaning seeking.
• Artificial intelligence can be trained to recognise patterns in data streams.

Together these new trends are deconstructing the role of the manager in the organisational information processing system:

• Information delivery: Managers are no longer needed to get messages delivered, technology enables direct communication across the organisation.
• Information amplification: Managers often still need to amplify important information that otherwise would not get communicated (e.g. job satisfaction of team members), but several technologies have created instrumentation that allows people at different layers of the hierarchical system to directly measure key performance metrics. The role of managers has shifted to the selective amplification of meaningful unstructured information.
• Information processing: The recognition of new types of information that the system's current processing architecture is not able to recognise is probably the most important role of management today. This role does however not need to be reserved for managers, and could be distributed throughout the organisation.

An organisation's information hierarchy is architected to amplify the kind of information that was historically necessary for its survival. Teams, departments, companies, industries and even the domain language used by the professionals working in them are all constraints that are organised around the ideas, and activities that historically maximised the organisation's ability to perform meaningful work. These constraints however evolve only slowly, and often require an existential crisis for them to shift.

While information processing through analysis and meaningfulness testing still requires human intelligence, there is an opportunity to delegate this task to information systems that can bypass the existing institutional boundaries, as the existing boundaries might work against the amplification of types of information that don't fit the current amplification and processing architecture of the system.

Artificial intelligence could play a role as an information amplification aid, presenting data to deliberately chosen people in the network who can determine if that information is meaningful. This could help organisations to sidestep existing hierarchical information processing structures, to amplify meaningful information that does not fit the existing information processing architecture. Which is the first step towards organisational change when it is needed.

This article was inspired by the following books and some amazing conversations:

1. "Dynamics in action, Intentional behavior as a Complex System" by Alicia Juarrero
2. "The Silo effect" by Gillian Tett
3. Signals and Boundaries: Building Blocks for Complex Adaptive Systems by John H. Holland
4. Several insightful conversations with Marc Burgauer