The Organization Design Forum offered an insightful session on the of role AI in Organization Design last week.
 
The use of AI has become a central question in boardrooms and executive meetings. Evidently, companies must find ways of positioning themselves around AI as a new opportunity to innovate services and increase effectiveness. Why should the field of Organization Design, which typically assists companies in such efforts, be any different? 
In last week's session, guest speaker Dr. Phanish Puranam (INSEAD) introduced Eunomicon, a Large Language Model he and his colleagues developed under the project leadership of  Dr. Nghi Truong at Sasin Business School in Thailand. The tool has been trained to offer thoughts and suggestions for organization design questions. It builds on Dr. Puranam’s and colleagues’ teaching, particularly their microstructural approach.
 
The microstructural approach to organization design captures common patterns of interaction between members of an organization. That output can then be used to structure along simpler elements that foster and support desired interactions and member efforts.
 
While this approach seeks to lead away from a focus on complex organizations, designing by member interactions and key challenges is by far not a cookie-cutter approach. In contrast, it relies on the reduction of blind spots and a common understanding of a desired future state.
 
That’s where a tool such as Eunomicon can be helpful. The language system was designed to act as an organization design co-pilot, which can be fed with cases or questions. It offers suggestions, explains its reasons, and perhaps more importantly, can also critique your organizational designs.
Just in case you were wondering: No, the tool won’t produce org charts for you! It doesn’t have graphic capabilities yet – and it is also not designed to support the idea that organizational design equals moving boxes on an org chart around.

The most important factor is that you can converse with Eunomicon in the same way you would with a human colleague, via text chat. In other words, it can add new perspectives just like an additional team member would.
When asked how we can prevent the tool’s views from being seen as the ultimate answer, Dr. Puranam stated that they are thinking of giving the tool a slightly annoying personality so that users will want to disagree and argue with it. And that, of course, would make it a refreshing addition to dialogical organization design and development approaches.