Introduction to RPA & RCA

In all business sectors and industries, we are noticing an increase in repetitive and administrative tasks. Processes are becoming more complex, including many tools and systems that are often disconnected from each other. Heavy regulatory and compliance constraints require the implementation of cross-functional controls. This development clearly degrades quality, efficiency and proper use of resources. This leads to additional workload burdens for teams and departments, increases in life cycle duration impacting production, and additional costs and real risks for companies.

Robotic Process Automation (RPA) provides solutions for this situation. This technology is based on task automation by software robots, called "bots". These are programmed to reproduce and repeat the same actions as a user on a sequence of tasks framed by instructions and rules. With a fast, simple and non-invasive implementation, RPA is helping to create a new framework that updates the links between existing technologies and those who use them. 

In order to automate tasks requiring the simulation of human cognitive abilities such as reasoning and learning, some bots can use Artificial Intelligence. Image recognition, machine learning or even natural language processing gives access to data left out by conventional tools. This is Cognitive Automation (CA).

In recent years, technologies based on automation and artificial intelligence have shown that they enable significant time savings, quality increases, cost reduction and resource valorization.  

In Life Sciences and Health Industries, all departments have processes eligible for automation. From clinical research to product safety, quality assurance to supply chains, commercial operations and IT. Some players have already experienced it for many office tasks and also within critical workflows.

The transition to automation is a challenge that requires identification of benefits provided by these tools along with their uniqueness. ADN is currently working in partnership with Automation Anywhere, one of the world’s leaders in the development of these digital tools, on case studies in order to define new good practices to adopt them effectively and validate them with expertise.

Indeed, one of the key points of this transition is the establishment of a good balance between these new “digital employees” and human intervention within the workframes. It is therefore necessary to clearly identify the involved processes, to explore the opportunities and to produce proofs of concept.

In addition, like all custom software and applications, the implementation of RPA/CA in a GxP environment calls for a methodology based on risk management and robust controls. This is why the qualification of bots leans on the lifecycle of the automation scripts, with risk mitigation tests covering the system requirements, error prevention, data integrity, security and change management.

The validation of those tools distinguishes rule-based deterministic functionalities and learning-based intelligent functionalities. For cognitive bots one has to consider and supervise the life cycle of an algorithm that requires an initial learning but also has the ability to improve itself with experience once deployed. Particular attention is then paid to the quality and relevance of reference datasets, the statistical approach for the assessment of efficiency, the anticipation of changes for continuous learning systems or the monitoring of performance and deviations.

ADN provides clients with support for business cases, election of candidates, proofs of concept, validation, deployment and monitoring of automated processes.

51 vues0 commentaire

Posts récents

Voir tout