The article discusses:

• The demand for rapid and scalable automation, and higher ROI
• The challenge of discovering right automations
• The Empirical Mode of Discovering Automations
• The Comprehensive Mode of Discovering Automations
• Combining Empirical & Comprehensive methods for efficient automation discovery and execution

The advent of “Digital”, “Transformation” and “Automation” has brought to the forefront two key buzzwords - “Rapid automation” and “Higher Return on Investment (ROI)”.

(A) Higher ROI: Irrespective of the technology lever deployed, i.e. Robotic Process Automation (RPA), Artificial Intelligence (AI), Blockchain, or Process redesign, higher ROI depends on a multitude of factors. These can include having wrong processes/candidates for automation, automating not-so robot-friendly applications, accelerating bot deployment/development cycles, lesser focus on integrating People, Process, Technology, automating largely human decision making driven processes, inability to connect process discoveries to deploying bots, absence of a strong governance structure, and not executing right tests for the deployed bots so that they function seamlessly.

(B) Rapid Automation: Automation is a cyclical activity and all the elements need to be integrated coherently to ensure rapid automation and more importantly, scale-based automation.

Data suggests that up to 70% of RPA project resources are spent on pre-automation1 , i.e. on the discovery phase as around 30-50% of automation projects fail2 because of inefficient automation discovery and over 40% automation projects fail3 due to large implementation timelines and implementation costs. This again rolls back to two fundamental questions that are a part of the “Discovery” step:

• Is there a healthy pipeline of automations in place from where another process can be quickly considered if the existing automations does not yield results?
• Have the right candidates of automation been considered after a robust evaluation process?

These two questions become important as automation is a cyclical process, i.e. (Figure 1).

“Discovering” the right automations?

There is a two-pronged approach that shall work in parallel to discover the right automations, build a robust pipeline of automation, and accelerate pace of automations as per the cycle defined in Figure 1. Identification of right candidates is generally classified into five levels, i.e.

• Level 1 can be an automation tower such as Human Resources (HR).
• Level 2 can be sub towers such as Talent Acquisition (TA) & Talent Retention (TR).
• Level 3 can be sub towers under TA such as TA planning, Candidate Sourcing, Candidate screening & assessment coupled with Employee Onboarding.
• Level 4 can be sub tasks under one of the areas such as Candidate Sourcing, i.e. determining right recruitment methods, performing right recruiting events and managing recruitment vendors
• Level 5 can be capturing Candidate Sourcing as an activity by distinct sub tasks constituting activity

1. Adopting a data driven approach also known as the Empirical Mode Of Discovering Automations:

a) The approach: Empirical approach to discovering automation is a numerical and matrix driven approach that examines data points such as complexity of a Level 4 area generally taking up to 2-3 days per activity. Typical activities carried out include analysis of AS-IS state, data and documents along with SME conversations and creation of a business case. For e.g. a Level 4 activity of Candidate Sourcing captures data points such as, the extent of straight through (STP), number of associates on the task, volume of the activity, average handling time of the activity, extent of automation, business criticality, customer demand etc. and by plotting all of these on matrices or visuals, decision making is enabled. Below is a sample mapping of some L4 level automations for an understanding (See the example in Figure 2).

b) Typical deliverable: RoI business case and heat map of automation

c) Pros:

• Enables quickest view of arriving at an indicative RoI basis 8-10 data points and narrows down the candidates of automation in a defined automation tower, e,g., Finance & Accounting, SCM, procurement or HR.
• Qualifies good & no so good candidates of automation very easily
• Is cost effective compared to the c omprehensive approach
• Generates a higher throughput, i.e. identifies larger number of automations in a shorter span of time and that too at a Level 4.

d) Cons:

• Provides only a high level of direction and excludes an in-depth analysis
• Excludes views of steps, activities involved and any other qualitative inputs

2. Usage of process mining/discovery tools for robust discovery known as the Comprehensive mode of Discovering Automations.

a) The approach: A multi-pronged approach is used for Comprehensive Mode Of Discovery involving process mining tools for cost reduction, elimination of any activity based inefficiencies, sets a precedence for standardization of activities, reduces manual intervention and enables prioritization of automation initiatives. Analysis of an Level 2 activity such Talent Acquisition (TA) as or a Level 3 activity such as Candidate Sourcing typically takes up to 2 weeks using this method. An indicative approach is depicted in (Figure 3).

b) Typical deliverable: Process diagnosis, business case and reimagined Process (To-Be) report

c) Pros:

• Enables detailed assessment of a utomations at a Level 2 or a Level 3
• Is more accurate compared to the Empirical Discovery Method
• Defines scope for more automation levers that include Blockchain, Process Simplification, Artificial Intelligence beyond RPA

             d) Cons:

• Expensive to be executed/deployed because of relatively time-consuming nature
• Requires usage of tools/licenses to be installed and procured
• Involves client readiness to share data logs, event logs and finer process details
• Involves deployment of technically qualified resources which command premium over other roles

Typical scenarios envisaged to use a combination of Empirical & Comprehensive discoveries

Empirical and Comprehensive discoveries can typically work in parallel for the following scenarios, i.e. Enterprises that have…

• ….let’s say 50 or 100 automations in place and want to scale up rapidly across multiple functional areas to realize benefits of automation.
• …not started their automation journey but believe they can do automations, because they have huge number of functions co existing within their ecosystem.
• ….been following a piecemeal approach, i.e. different functions within in enterprise are initiating their own automation journeys in parallel.
• ….been following a data driven approach to find the right automations by considering only fewer parameters such as Average Handling Time, Volume of transactions, No. of employees, % of straight through etc. but still are not able to identify too many opportunities for automation.
• …not been able to go beyond the obvious, i.e. identifying opportunities using RPA, AI, Blockchain, Process simplification, improving customer experience (CX) etc.

The second prong of the two phased approach: Getting it all together for rapid and scalable automation

The Empirical & Comprehensive mode of discovery shall work in parallel and mandates a strong governance structure to monitor execution. The governance structure shall monitor the automation discovery, automation execution and change management in parallel and a lapse on any of these areas impact the overall scalability and pace of automation, ultimately impacting the ROI.

(Figure 4), i.e. the integrated view of execution, indicates that automation discovery teams shall churn the right cases of automation leading to a creation of a healthy funnel of automations. The automation development teams will use the healthy funnel of automations to develop, test and even discard them at development stages in case incorrect automations have been selected. The development teams can then deploy automation bots and pass it on to the support teams so as to ensure stability in automations. The de-selected automations at the stage of development are passed on to the change management teams for further validation and the change management teams communicate with the automation discovery teams for another round of validation, funnel creation and development, thereby making this a cyclical process that originates with discovery.

The conclusion

A multi-disciplinary approach combining both Empirical & Comprehensive discovery can result in massive quantitative benefits for automation stakeholders such as 5-8 times faster discovery of automations, 30-40% employee reduction, 20-25% improvement in turnaround time and 3-5 times faster speed of deploying automations. All of these benefits lead to qualitative benefits such as continuous improvement in operations, higher employee morale and greater focus from tactical to strategic tasks.

In summary, the right amount of focus on due diligence and building a business case ROI on automation with at least a 3-year horizon shall be crucial, and for this “Automation Discovery” needs to have its heart in the right place to guide subsequent teams in the implementation value chain. Moreover, rapid automation and scalability in automation can only be achieved via seamlessness across phases such as automation discovery, automation deployment, automation testing, change management and governance.

References

¹http://www.b2b.com/how-to-cover-the-costs-of-rpa

²https://www.raconteur.net/technology/rpa-failures

³https://blog.aimultiple.com/rpa-pitfalls/

Mehul Damani

Digital Operations & Platforms, Enterprise Operations Transformation, Wipro

Mehul is a seasoned digital and automation thought leader with 12 years of rich exposure dealing with CXO level clients across business domains such as digital transformation, automation, customer experience (CX) journey enhancement, business strategy, B2B/B2C consulting, marketing & sales enablement operations across industry verticals such as Retail, Consumer Goods, Travel, Manufacturing, Hi-Tech & Telecom.