Launching the agent-based model for teamwork
A Streamlit app to simulate and study emergent properties of teamwork (... and a strategy to increase the ROI by 80% at no extra cost!)
Today, we are releasing version 1.0 of an agent-based model for social teamwork. It is deployed as a Streamlit app to make it easy for you to run simulations and analyze data in your web browser (it works well on large screens and iPad, less so on a mobile phone).
So, what are agent-based models? According to Wikipedia: “An agent-based model (ABM) is a class of computational models for simulating the actions and interactions of autonomous agents (both individual or collective entities such as organizations or groups) with a view to assessing their effects on the system as a whole. It combines elements of game theory, complex systems, emergence, computational sociology, multi-agent systems, and evolutionary programming.”
We have built an ABM for teamwork. Workers are assembled in teams to work on projects based on their (hard and soft) skills and their availability. Five different organizational models (e.g., the overcommitted or the emergent organization) are “preset” and ready to be simulated. In addition, team allocation mechanisms (e.g., random vs. optimized), budgetary and timing flexibility, the number of projects per timestep, and skill decay (among others) can be configured for the simulation.
The model framework is used to test several hypotheses, understand emergent properties, develop optimal strategies for team assembly, cognitive diversity, dynamic resource allocation, training programs, retention, the right level of ‘organizational slack,’ and also to generate simulated training data for the Machine Learning (reinforcement learning) prototype.
Key messages include the following:
Today, projects are staffed from team members within the same organizational unit (e.g., a department or division). This happens not least because it is the straightforward thing to do as you know the people and you avoid difficult discussions with line managers from other units. However, expanding the search space and including all workers of the organization as potential candidates is very valuable, and it is today an untapped resource. Maximizing the search space results in higher skill levels and better-balanced team compositions (all else equal).
Today, team formation is not seen as an optimization problem with an analytic solution but rather as a mostly arbitrary composition of individuals. Biases undermine team selection and poorly understood assumptions. Team members, while individually impressive, perform poorly together. Assembling teams using an optimization algorithm increases the return on investment (ROI) by approximately 40% (compared to a random team formation).
Today, projects are heavily formalized and subject to rigid constraints regarding starting dates and budgets. This means that the project will not wait for the right talent to be available, and sometimes it can simply not afford the right talent (particularly for high-stake projects). Relaxing timeline and budgetary constraints increase the ROI by an additional 30% (compared to an optimal team assembly but with no flexibility in terms of timeline and budget; compared to a random team formation, the ROI increases by 80%).
Today, training is not selective or targeted. Training efforts focus on the individual and not on the (current and expected future) gaps at the organizational level. Training of selected in-demand skills to close gaps will help the organization learn and mitigate the risk of fast skill decay.
Today, the level of ‘organizational slack’ is often high, expensive to maintain, and not managed well (or not at all). While some excess capacity is vital for the organization to remain agile and adaptive, most organizations can add the same value at one-third of their current capacity levels. Finding the ‘optimal slack’ level and implementing rigorous dynamic resource allocation adds value for the organization. (In an extreme case, Slack can also be “outsourced” to the ‘human cloud’, allowing organizations to become ‘talent-light’.)
In summary, and very much in the spirit of SuperScript’s value proposition:
Great teams are not born, great teams are made.
SuperScript allows organizations to continuously learn and gain a competitive advantage through more creative and innovative solutions to their challenging problems (not least via ‘cognitive diversity by design’). Organizations understand their workforce’s aggregate skills and gaps towards the desired state. As part of a fluid resource allocation, organizations can move the right talent to high-priority projects and nudge workers with adjacent skills to right-skill towards ‘in-demand’ roles. Thereby those organizations solve the logistics problem of human labor and dynamically deploy talent to where value is. They carry less excess capacity at any time and also actively manage the ‘organizational slack’ to be more resistant (or antifragile, even) to exogenous shocks. The approach promoted by SuperScript empowers an organizational transformation that emerges from the bottom up and, therefore, a change that is widely shared, understood, and thus more sustainable. Engagement is created through better and more inclusive use and development of in-house talent and higher employee satisfaction.
So, go ahead and run the agent-based model and play with different parameter configurations. What do you find interesting or surprising? Your feedback is much appreciated.