Typical challenges that intensify failure modes leading to loss of bot productivity

To fix this, we leveraged web-services with load balancing features coupled with auto-sensing mechanism and prescriptive algorithms to produce a scalable orchestration platform for guided deployment of pre-determined number of bots as a scheduled task. The task can be scheduled to execute both attended and unattended bots.

Functional flow diagram  showing AI–OCR bot execution in sync with Master Orchestrator Console & Dashboard

The orchestration platform takes in account the priority of individual bots, with an overall objective of optimizing the cost of operations, reducing cycle time, improving productivity while monitoring real-time exceptions and reporting them, thereby maintaining high standards of accuracy and compliance for in scope process automations.

Key design principles to deliver experience with sustainable business impact

Experience Led Design

• Continuous prediction based on real-time median values of AHT and available inventory to arrive optimal number of bots for a scheduled task
• Continuous auto-sense of inputs - available bot virtual machines, inventory, and applications connected / disconnected states, bot run & exception states.
• Continuous monitoring of metrics and inputs to re-organize bot runs
• Continuous self-organization and allocation of freely available VMs at regular intervals for optimizing bot productivity 
• Self-recovery of exception transaction with maximum three re-attempts to run the exceptional transaction before moving to the next transaction 
• Self-triggering of warning messages and error message on console window, along with emails messages due to breaches of baseline bot metrics, application(s) disconnection, bot exception(s), bot(s) idling due to process completions.
• Batch processing by bots to minimize Database Input Output Operations and improve AHT.
• Bot allocation and run based on either user defined priority or inventory ageing or both
• Autonomous trigger to perform bot Schedule, bot initiation and bot kill at any time.

Digital Everything

• Dashboards showing real-time bot metrics with digital-VSM for end-to-end visibility
• Effective load balancing of controller service to scale bots operations across multitudes of virtual machines seamlessly and simultaneously
• AI-OCR bots for automating Citrix XenApp Applications 
• Next gen customer experience with multi-channel support
• Certifications from global renowned static and dynamic vulnerability testing software.
• 128 bit encryption for added security 
• Artificial intelligence/machine learning (AI/ML) for advanced analytics

Open Innovation

• Outcome focused commercial models for higher ROI
• Digitally enabled and upskilled resources to monitor bot runtimes and optimize outcomes

A predictable & scalable orchestration cum monitoring engine

Wipro's agile solution addresses challenges that impact customers by providing a predictable scalable orchestration cum monitoring engine with real time metric updates and supports bot runners with a 360 degree view of the automated process, while optimizing productivity and minimizing delays due to exceptions.

Functional flow diagram showing bot processing within healthcare claims ecosystems, by navigating through applications

Key benefits

• Open source tool usage led to low cost deployment within a span of 8 months while transforming end to end claim processing landscape of a large Healthcare organization in virtualized environment comprising of Linux Servers, hosting flask application with G-Unicorn for multithreading, coupled with Python Bots running in virtualized Windows environments, while connecting to Mongo DB for storage, and Nginx F5 for load balancing.
• Ability of master controller to span multiple bots across virtual machine for fully automated bot execution, using master salve run-time configuration.
• SSL encryption for all data exchange between bots and server DB to ensure security.
• Pass certification from Sonarqube for testing code quality in both static and dynamic manner, for all bots and services, thereby cleaning eliminating bugs, code smells, and security vulnerabilities.
• Self-guided orchestration platform with provision for manual intervention, achieved 40%+ STP using a pool of unassisted and assisted bots across claims landscape for a US healthcare company.
• ~15% to 25% reduction in exceptions resulting in reallocation of capacity to handle exceptions.
• ~10% to 15% incremental ageing inventory reduction.
• ~15% to 25% improvements in cost takeout due to reallocation of cost drivers.
• 100% adherence to accuracy and compliance norms.

• 25% of the incoming inventory is non workable and gets pended daily
• Manual handling of non-workable inventory was to the tune of 15 to 20 FTEs per month. 
• 25% of the pended inventory was resolved on any given day
• 75% pended inventory got re-pended that led to manual rework on next working day
• 200%+ growth in volumes lead to pend inventory ageing to the tune of 30+ working days.
• With learning curve of 6 months’ duration, existing staff could stretch to add 20% capacity.
• Restricted availability of bot virtual machines, due to VM cost escalation by 30%.

• Design led thinking led to creation of unassisted bots for inventory classification, pend scrubbing bot with single interface orchestrator service
• Scalable architecture with multithreading built with Nginx.
• Predictive analytics for guided decision making by the orchestrator console
• Non-invasive automation through assisted bots, for handling corrected claims, rekey of claims, encounter and MedSup claims.
• Mongo DB to handle new schema dynamically even when the process is re-designed multiple times.
• Culture sensitization, skill building for bot runners

Benefits

• Total business value provided to the client through orchestrator enabled assisted and unassisted inventory management solutions is $ 1.5 million in FY 2021-22
  
• 25% reduction in cost of new age inventory management
• 90% reduction in cost of non-workable inventory management
• 95% automation of inventory management

Awards

• Platinum Award in Watts Humphrey FY 20-21 
• Gold National Award in NQCI FY 20-21
• Platinum Award in CII for Low Cost Automation in FY 21-22

If you are interested in learning how Wipro is helping our clients achieve their vision of digital transformation, we should talk. Connect with us here.

About the Authors

Suprateem Moitra - Associate Vice President - Quality, Wipro

Suprateem has more than 18 years’ experience improving process delivery through Robotic Process Automation, Process Orchestration, Workflow Optimization, and Intelligent Automation for Healthcare, Banking, and Insurance domains.

Suprateem is an Automation Leader for Healthcare vertical and holds dual Lean Six Sigma Black Belt certification. He has a Masters in Management from the Indian Institute of Foreign Trade (IIFT) New Delhi and a Bachelors in Engineering from the Indian Institute of Technology Kharagpur.

Prasun Bhattacharyya – Automation Team Lead, Wipro

Prasun has more than 2 years’ experience improving process delivery through Process Orchestration, Workflow Optimization with Intelligent Automation for Healthcare.

Prasun is a seasoned python coder, and has published python libraries in Pypi.org.