A closer look at how our Life-Sciences and Healthcare clients are enhancing efficiency and performance for greater profitability and ROI by deploying IPA
RPA vs. IPA
Robotic Process Automation (RPA) has been the talk of the town for some time. It has helped companies eliminate outdated manual steps in a process and improve accuracy for underlying business functions.
Most of the benefits of RPA have been in the context of savings in resource cost by eliminating one or more steps within the workflow of the entire business process. Moreover, automation thus far has taken a single dimensional approach – An action post decision often involving mundane and repetitive tasks can be automated as a script for a machine to do more efficiently.
This approach, whilst having proven successful in short-term gains, has sometimes had adverse impacts on process improvement in a post digital enterprise. An analogy would be to automate a flawed step in a business process with a mix of legacy systems and newer applications that do not integrate well, which would make the entire process more hazardous rather than derive desired benefits. Hence, it is imperative to not look at automating a process piecemeal just to expel repetitive scripted human effort, but more-so, take a holistic approach to Intelligent Process Automation (IPA) via underlying digital worker concept to unlock significant value to an organization.
IPA in a digital enterprise, on the other hand, uses an “artificial intelligence-first” approach to provide decisions and recommendations – by using intelligence to automate all or most of the process steps within a business process and using human intervention only for review or exception handing. Feedback from such interventions is henceforth used to improve decision capability by the Intelligent Automation Platform.
RPA to IPA
RPA has been leading the front in terms of partial automation of actions or script-based services with respect to full-time employee savings and quicker turn-around time. However, its limitations to scale are more widely encountered in the inability to process unstructured data, cognitive and diagnostic capability, and all-in-all end-to-end automation of processes that require some form of interpretation, contextual extraction, and enrichment and cognitive decision making.
One of our Life-Sciences customers embarked on a RPA journey a couple of years ago to automate the Accounts Payable process by building Bots to take necessary actions such as updating ERP, initiating ACH for payments, and sending out invoice processing notifications once the manual validation and decision is performed on each invoice. This has since helped them achieve an average cost savings of ~ 3 FTE per month and ~ 30% reduction in TAT. Whilst the business celebrated immediate gains, they realized potential limitations to scale this automation across the globe with continuous increase in volume of vendors, associated invoice formats, languages, and accuracy to work order.
The client engaged Marlabs in AP process prioritization and assessment feasibility, post which we along with our IA partner recommended an intelligent processing platform. The platform enables smart extraction and enrichment of line items, invoice amounts, and work order and applies contextual reference and inference to empower decisions about invoices such as pay, reject, or request more information. The downstream workflow continues to be part of the RPA BOT journey as earlier. Review and exception handling by human workers are used as feedback to improve abstraction and quality of the platform. The initial implementation has confirmed to not only 4X benefits but also achieve 100% compliance and accuracy.
AI 1st IPA
Large pharmaceutical companies today deploy almost 10% of their R&D budgets on patient safety and a large piece of it is allocated to case processing.
Even with an exponential increase in case volumes YOY, pharma companies are finding it challenging to leverage newer technologies for automation.
This is partly due to low standardization across processes around case intake and translation, stringent regulatory requirements in case processing, reporting and risk management, and maturity in frameworks to comprehend cognitive models to work with RPA, OCR, ML, NLP and other technologies.
Along with our IA partner, we are working with a Fortune 50 pharma company to digitally reframe their end-to-end patient safety process in R&D using the AI-first approach. Self-learning orchestrated cognitive models are used to classify various sources, contextually extract semi and unstructured data, and semantically translate it to enable automatic discovery, identification, and confirmation of relationships between various medical and clinical data entities. These entities include drugs, procedures, symptoms, and ailments for case processing and reporting.
We are now helping the client operationalize AI to drastically accelerate decision intelligence and automation via in-house framework and pre-built tools, partner platforms, and other COTS products to scale multi-dimensional enterprise grade AI business solution. The near-term outcomes are estimated at $5 million annualized savings and over 50% reduction in processing time.
Maturing the IPA Curve
IPA requires business alignment and top-down commitment to have the right combination of business, IT, and operations seamlessly function as one team. We not only help our customers evangelize the functional blueprint and roadmap, but also provide an outcome-based operating model that focuses on right-fit governance and change management for man and machine to work together.
Our IPA framework helps customers evolve and move up the automation maturity model ladder – from script-based task automation to autonomous processes for amplified outcomes, as measurable improvements in operational productivity using cognitive and augmented intelligence for continuous learning.