Centralised focuses all RPA capabilities within a Centre of Excellence (CoE), and Decentralised places those capabilities within the individual Business Units, giving them complete control and autonomy.
In the middle of this spectrum sits the Federated structure, which also features a central team, but includes an RPA team within the Business Units. Under the Federated structure, the business unit RPA teams have many of the development and testing capabilities required to develop RPA solutions, with centralised governance, frameworks, and standards sitting with the RPA CoE.
While each model has the pros and cons typical to any type of CoE, the Federated structure allows more control in the business units where the risk predominantly sits – a relatively unique characteristic of RPA. This also ensures standards and shared accelerators are in place from the centralised team to ensure scalability.
All three operating models (or hybrids of) have value in different contexts and organisational RPA maturity levels. Organisations just starting out on their RPA journey may choose to first implement the centralised model in order to ensure a consistent foundation is put in place, while in the longer term, a shift to the federated model may be desired once operational maturity is reached so that the business units themselves can better maintain their virtual workforce. In any case, an assessment must be made to ensure the chosen structure is fit for purpose for the unique requirements of the organisation.