RPA(Robotic Process Automation): Apply IBM’s Watson conversation service on Calling Center

RPA: Apply IBM’s Watson conversation service on Calling Center

Contents

1. Purpose of paper
2. High level approach and scope of research
3. Research methodology
4. Finding
5.1. What is RPA
5.2. How RPA create value:
5.3. RPA’s Past and Future
5.4. Current Technology Landscape
5.5. How to implement RPA
6. Apply RPA on Calling Center
6.1 How apply Watson’s conversation service will create value in Calling Center
6.2 Software Solution (Watson Conversation Service)
6.3 Architectural Approach and Design
6.4 Challenge of RPA/Watson’s conversation service
7. Conclusion
8. References
___________________________________________________________________________

1. Purpose of paper
calling center should examine the opportunities and challenges of RPA and IBM’s Watson conversation service to cut cost and improve performance.

2. High level approach and scope of research
Examine the current RPA’s potentials, challenges, technology landscape, and recommend some actionable solutions for the company to apply RPA to create value.

3. Research methodology
Utilizing three research’s steps for the problem.
A) Literature review: analyze and interpret the work of other researchers.
B) Conceptualization of solution: conceptualizing the solutions to the research question.
C) Experimentation: Design and develop the prototype solution for the problem.

4. Finding
Companies should consider cost-benefit-analysis (CBA) of RPA to increate competitive advantage. With progress of Deep Learning and GPU, the RPA has potential to create more value for companies.

5. Analysis

5.1. What is RPA

RPA is the use of robot, usually software with machine learning (ML) and Artificial Intelligence, to replace the human on the tasks that are high-volume and repeatable.

5.2. How RPA create value:

 

Current Processing

RPA Enabled Processing

Cost

High Total Cost of ownership due to high resource requirement

Potential reduction up to 80% of total cost

Processing time

Typical days and hours

Minutes and seconds

Productivity

Staff focused on low-value, iterative task

Free up staff to focus on handling value-added, creative, and complex tasks.

Quality/ Accuracy

High error rate due to natural of human

Low error rate due to the accuracy of computer

Audit Trail

Lack of audit trail due to natural of human

Fully maintained logs essential for compliance

Reliability

Not always available 24/7

24/7, all year around

Staff retention

Low staff retention due to low-value repetitive task

Shift human-resource to creative and value-added task, lead to higher retention

Scalability & flexibility

Low scalability due to high-cost of human-resource and limited office-space

Instant ramp up/down to deal with spikes and troughs in demand

5.3. RPA’s Past and Future

The RPA expected to grow exponentially for the next 5-10 years, and replace up to 140 million full-time employees by 2025.
RPA

5.4. Current Technology Landscape

The current technology landscape can divided into 4 levels of sophisticated.
Level 1: the vendors offer very simple solution for the RPA. No-AI and can only handle 1 single data source structure.
Level 2: vendors offer a more complex solution. No-AI and can handle multi-source unstructured data.
Level 3: vendors offer a more comprehensive solution with AI.
Level 4: a complex-comprehensive solution with AI and can handle multi source unstructured Data. Both Microsoft-Cortana and IBM-Watson are the big vendors with their own Cloud service and Ai-engine.

RPALandScape

5.5. How to implement RPA

RPA-Implement

6. Apply RPA on Calling Center

6.1 How apply Watson’s conversation service will create value in Calling Center
In 2010, IBM’s Watson had beaten the human contestants in Jeopardy and set a milestone that AI could be as smart as human.
With the advantage of Natural Language Processing (NLP) in the last decade, calling center could utilize Watson’s conversation service and RPA to cut cost and improve performance.

Watson

6.2 Software Solution (Watson Conversation Service)
Watson-solutioin

6.3 Architectural Approach and Design

Architectural-design

6.4 Challenge of RPA/Watson’s conversation service

- Current NLP (Natural Language Processing) and Watson’s conversation service are still not perfect; there is a moderate rate of inaccuracy. IT department should continuously monitor and update the system to improve performance and effectiveness.
- Current AI (Artificial Intelligence) is still immature; employees needed to involve in processing to avoid error.
- Current RPA has limited human interaction which may reduce the customer’s satisfaction.
- Company should examine if the tasks have a good ROI (Return on Invest) when apply RPA.

7. Conclusion

Companies with repetitive tasks should consider apply RPA to cut cost:

  • Small companies should consider using free RPA solution, UiPath ‘s free community edition or WorkFusion RPA express.
  • Big companies should invest in more expensive and complicate RPA solution.
  • Large company with aggressive growth strategy should consider investing in RPA with advantage AI to maximize automatic processing; for example: IBM’sWatson and Microsoft’s Cortana.

8. References

1.  Robotic Process Automation, Wikipedia.com. Retrieved from: https://en.wikipedia.org/wiki/Robotic_process_automation
2. 12 proven ways successfully implement robotic process, 2016, Pranay Kushwaha, Retrieved from:
https://www.linkedin.com/pulse/12-proven-ways-successfully-implement-robotic-process-pranay-kushwaha/
3. How has RPA played a role in increasing data accuracy and predictability in your healthcare operations?, 2016, Barbra McGann, Retrieved from:
https://www.horsesforsources.com/how-has-rpa-played-a-role-in-increasing-data-accuracy-and-predictability-in-your-healthcare-operations
4. How can we complete X process in less time, for less money, while maintaining or exceeding the current quality?, Jessica Lewis, Retrieved from:
https://www.pwc.com/ca/en/industries/financial-services/insurance-speak-blog/rise-of-the-robots.html
5. Cognitive Robotic Process Automation for a Digital workforce, Ernst & Young, Retrieved from: http://www.ey.com/in/en/services/advisory/performance-improvement/ey-robotic-process-automation
6. Robotic Process automation (RPA) and Avaloq, 2017, Philipp Frauenfelder, Retrieved from:
https://www.linkedin.com/pulse/robotic-process-automation-rpa-avaloq-philipp-frauenfelder/
7. Robotic Process Automation (RPA) Evolution, 2017, Everest Group, Retrieved from:
http://ww https://almato.de/en/news/blog/trend-details/press-review-what-you-should-know-about-implementing-robotic-process-automation/w.everestgrp.com/tag/robotic-process-automation/
8. What you should know about Implementing Robotic Process Automation, 2016, Sara Gebhardt, Retrieved from:  https://almato.de/en/news/blog/trend-details/press-review-what-you-should-know-about-implementing-robotic-process-automation/
9. Watson API explorer, 2017, IBM, Retrieved from:
https://www.ibm.com/watson/developercloud/conversation/api/v1/#get_workspace