Reality check: is robotic process automation a panacea for transforming regulatory affairs?
Regulatory operations are burdened by resource-draining document and data processing tasks, but is robotic process automation the definitive solution? If not, where does it have greatest application and appeal – and how can life sciences firms exploit the full benefits? Agnes Cwienczek scrutinises the technology’s potential.
In regulatory affairs (RA) and regulatory operations in particular, the potential for transforming workload management with robotic process automation (RPA) is substantial, due to the sheer volume and intensity of administration involved in the function. It is hardly surprising that companies are evaluating the possibility for reliable, expedited help with this.
Even without any real applied intelligence, RPA systems can accelerate the completion of repetitive manual tasks – freeing up expensive talent to use their knowledge and skills more productively, while reliably processing work items that might invite error as the human brain switches off.
Simple versus less-structured RPA
For simple task execution, RPA is a great place to start for RA functions looking to apply their resources in smarter ways and reduce risk and time to market.
Useful cases where RPA could be applied include highly structured, ruled-defined processes and tasks that lend themselves to this form of automation: the kind of undertaking once routinely outsourced to third-party service providers for superior cost-efficiency. Such tasks might include automated data entry; extracting data from Excel sheets for uploading into databases, importing documents or archiving them; checking data or document quality; or parsing emails.
In clinical regulatory operations, meanwhile, where there are hundreds of reports coming in from contract research organisations (CROs), potentially running to thousands of pages of information. Here, an RPA tool can help take the strain away from teams who otherwise would have to input data manually from these documents according to a given checklist of entries. Creating or verifying hyperlinks between related data in documents is another aspect RPA tools can help with.
Checking documents for submission-readiness according to defined criteria is another strong candidate for RPA; is the PDF file the right version – does it have the right settings and bookmarks? Checking off documents prepared by scientific professionals against 10-20 criteria (which could mean the difference between acceptance or rejection), is a full-time job – occupying whole teams who have to manually process hundreds of documents each day.
Processing and parsing emails is a further prospect. Here, RPA could transform the extraction of standard data from routine documents such as standard agency approval letters. This is an example where more intelligent, artificial intelligence (AI)-enhanced RPA can boost a tool’s potential – and the payback.
AI-enabled RPA allows tools to cope with unstructured scenarios as well as fully standardised, highly structured contexts where the parameters remain consistent and predictable. As agency approval letters can vary by country, an RPA tool with some degree of machine intelligence could help by first determining which country the letter has come from and therefore where to look to extract the required information. It can then identify how to interpret it before uploading the results into a regulatory system.
…companies should be thinking about further standardising the way they capture, record and manage data”
As RPA use becomes more widespread, companies – especially the large players – are becoming more mature in their application of the technology as well as more ambitious in their aims. Once they have tried out RPA and seen for themselves the tangible benefits, these organisations are beginning to think laterally about where they could take the technology next.
A welcome shortcut to success
The use of RPA as an agile, interim solution to deliver quick wins in parallel to larger-scale transformations of regulatory information management (RIM) is becoming increasingly common. Where companies are impatient to deliver return on investment (ROI) and accelerate speed to market now, targeted RPA applications – turned around quickly and affordably – can readily demonstrate their worth and reaffirm the business case for regulatory digitalisation.
Targeted RPA applications also help to highlight what is possible, inspiring investigation into more advanced uses and reassuring teams that automation is not a threat to their jobs, but rather the key to making them more interesting. Building a good, solid regulatory intelligence database is a good example of a next-generation RPA use. A blended approach of RPA-extracted data and human insights can result in a powerful resource with wide-reaching benefits in accelerating and improving the quality and success rates of global submissions.
Targeted application drives up the results
Identifying strong, targeted uses will be important as companies become more serious about RPA. This will help to build credibility and confidence around the technology and break down fears about technology replacing people at work.
Validation of RPA technology could conceivably be a challenge, particularly where systems are continuing to evolve using AI and machine learning. Niche, application-specific ‘bots’ which execute repetitive tasks and are relatively restricted in their use and predictable in their performance, however, should not pose too much of a problem.
Who does what?
Next companies decide to what extent they will develop and run their own RPA capabilities or whether they will lean on third parties to create and operate the tools for them.
There is an expectation that some process automation tools will remove the need to heavily rely on external services to improve operational cost-efficiency. In other cases, use of advanced technology will become a pre-requisite when choosing service providers, to ensure maximum economic benefits when outsourcing.
Maximising RPA’s scope
To fully exploit RPA’s potential, companies should be thinking about further standardising the way they capture, record and manage data. RPA bots are relatively easy to code – the bigger challenge is harmonising processes and channels and shoring up data quality so that automation can be applied easily and reliably.
One tip is to identify where tasks are executed according to check lists or which are the most resource-draining or inefficient outsourcing relationships and use this as the steer for advancing with digitalisation.
Beyond the applications already suggested above, processing product data from contract manufacturers (CMOs) and compiling and publishing dossiers are workloads that offer considerable potential for digital transformation using automation. The wider benefits of all of this, of course, are that life sciences organisations start to build a much-needed capability around and appetite for digital process transformation, which could be extended far and wide across their operations as new opportunities for improvement are identified.
About the author
Agnes Cwienczek is Head of Product Management & Consulting at Amplexor with a remit that includes the provision of business process and data management expertise in the areas of Regulatory Information Management, Document Management and Submission Management. Prior to joining Amplexor, Agnes worked at Merck in its Global Regulatory and Quality Assurance department. She received her Master’s degree in Information Management from the University of Koblenz-Landau.