Robotic Process Automation (RPA) is essentially process of digital workforce institutionalization. The digital workforce does not suffer from Monday blues or lack of enthusiasm. Instead, digital workforce is most suited for mundane work most dreaded by human workforce. A good RPA implementation performs such monotonous work with brilliance and perfection. More the mundane and predictable work, higher will be the RPA efficacy. A good side effect of right RPA implementation is heightened thinking scope for human workforce, it can focus more on what it is meant for – thinking and designing. Does that mean is it the right thing to do? It is, as long as there is positive ROI. First step towards calculating ROI is understanding costs and benefits, this post covers some of these common factors for RPA.
What it essentially means is right set of business processes must be considered for RPA and right set of automation should follow for better ROI. Majority of RPA thought processes start as cost saving or productivity improvement initiatives – most of these are relatively easier to measure. Some organizations look past these easy measures to additional measures such as error reduction, repeatability, accuracy, etc. Following are some the key RPA benefits that have worked well for me in justifying RPA benefits:
Direct Cost Saving: One of the most common reason for RPA consideration. Low key repeatable processes are best targets for digital workforce. Paper to digital format conversion, OCR, data entry, etc. are classic examples. Automated under controlled conditions these can produce best results. Associated cost savings are relatively easy to calculate which can be converted to monetary benefits.
Speed: Back-office tasks and multiple applications data entry are classic slow processes. BOTs can ring in significant benefits in hastening such processes. For example, extracting documents from email to make entry into financial system for payment processing is very common automation candidate. Reduced time can then be extrapolated to monetary formats such as delayed payment charges, early payment bonuses or simply in terms of reduced headcount.
Improved business focus: Generally, afterthought, this should ideally be the primary reason for RPA. Automation should take away robotic activities from human workforce to create pro-business environment with focus towards revenue and business expansion. Benefits are not so easy to calculate, though number of new accounts or additional sales can be used as measure. Another alternate is to employ robots for sales, chatbot for structured Q&A is classic example. In this case direct customer satisfaction or number of sales are attached to BOTs.
Reduced Rework: Submitter and Approver approach has worked well so far. This approach works well because people find deficiencies in others work! Human mind is wired like this. These are workflow and people driven, includes rework and duplication, enterprises are full of such implementations. Common inefficiencies such as human error or time lag between state transitions can be addressed through RPA. Some of these benefits can be converted into monetary benefits some may just be qualitative in nature.
Improved Trust: Flip side of reduced rework is improved quality. Combined with consistency the quality gets improved further. Continuous consistency results in significant quality boost thereby increasing trust. It should be noted that trust builds with favourable outcomes and drops with missed targets. Hence one should be mindful of wrong automations as consistent wrong outcomes would impact trust. This is one of the most difficult aspect to measure. While trust cannot be directly measured, consistency or quality can be measures and then translated into tangible means such as reduced customer complaints or improved sales.
Talent Retention: Less the mundane tasks, more the satisfied human workforce. Such an environment results in charged and happy workforce who are willing to take-up new challenges. This results in better business focus and enterprise alignment. Attrition rate, hiring costs, separation costs are example measures.
Licenses Rationalization: Many a times inefficiencies in processes results in increased manpower, this in return results in additional licenses procurement. Implementation of RPA can help in warding of these wastes and help in rationalizing or reducing licensing requirement. This is a direct financial measurement in terms licensing cost saving, one might even add savings in hardware and real estate cost as appropriate.
ROI depends on two primary components – income and investment. While there are many RPA benefits (income), they come with cost (investment). For a profitable and favourable outcome the benefits should outweigh cost. Following are some of the RPA costs:
Licenses: First and foremost, the license cost should be clear. BOTs, BOT managers, reporting engines, BOT editors and optimizers are common RPA licenses. Additional licensing requirements such as third-party products for OCR or cloud services for special processes should be counted. Overall license cost should include underlying OEM licenses for operating system, databases, monitoring tools, antiviruses, etc. Another common lapse in licenses is the impact on surround system licenses, especially the enterprise class systems such as SAP. Automation budget should include all these licenses for investment.
Consulting: Automation requires good understanding of business, without which automation would be meaningless or counterproductive. The skillset requirements for automation vary from business analysis till implementation. Every RPA implementation involves good amount of reverse engineering effort. One must be cognizant of all the consulting effort to arrive at meaningful RPA cost assessment.
Keeping lights on: Focused team is required to sustain and maintain post go-live RPA implementation. Good team comprises of business users, analysts, RPA experts, developers and infrastructure support experts at minimum. Impact on RPA should be considered for every change in underlying system. Seasonal or policy driven changes can wreak havoc in system, oversight may result in wrong outcome. Depending on approach, one may need to realign RPA implementation or suspend it for a certain duration and resume post operations stabilization. Policy driven change such as GST implementation in India is classic example. RPA also must realign to accommodate tax calculation changes. Any insights on such future policy changes should be considered for prioritization as well to align appropriate resources.
Successful RPA also takes continuous improvement needs into account. This requires future proofing in the form of AI capabilities for self-improvement or may require manual tuning for higher accuracy based on experience. In either case, specific skills should to be assessed and aligned for RPA maintenance. Combined with skills, the time critical nature of solution and SLAs should be considered for organizing support team structure.
Infrastructure: Some of the automation tasks might be weekly in nature, some might be daily and some might be monthly. Some of the tasks might take 2 hours and some might take 20 hours. All such load requirements should be considered and assessed to ensure appropriate capacity planning and deployment. Like load factor, the turn-around-time commitments should be assessed to arrive at high availability and disaster recovery plans combined with guidelines as per organization standards and policies. It is also important to note some of the automation frameworks allow deployment of more than one BOT per server while other strictly restrict one BOT per server. Such differences should be identified as they may help in reducing overall OEM costs while the hardware cost may remain same.
OEM costs: Some of the points are already discussed earlier as part of licenses. OEMs such as operating systems, databases and essential software required for RPA deployment are must haves. There would be other set of OEMs such as antivirus software or server monitoring agents that are to be deployed as per enterprise operating policies and guidelines. Licensing agreements with COTS products should be assessed as systems that are being automated may have different licensing structures. It is important to consider impact of automation on existing applications licenses to ensure automation is within acceptable licensing agreements. Other associates costs such as AMC or yearly subscription costs should be taken for having clarity towards OPEX.
Core system changes/integrations: Digital workforce describes RPA at best. This means the automation is acting on behalf of a real user. However, the core systems should demonstrate certain characteristics for automation. These characteristics either exist from day one or they are developed specifically for RPA. This is where one must perform impact analysis on underlying systems to identify changes needed in systems or the integration points to be developed outside the RPA framework. These are generally custom in nature, one should proceed with RPA provided the cost of changes in core systems to produce results without RPA framework are much higher and future needs vary. On a similar note, integration needs are to be assessed as these may result in system changes. One must keep good outlook on underlying systems as frequent changes in systems would require automation changes resulting in cannibalization of all the benefits.
Above are some of the common factors involved in RPA cost-benefit assessment. List may vary based on individual enterprise needs. Once costs and benefits are factored in, one also must put in place both leading and lagging indicators to communicate RPA success. Most individual business use cases take about 8-12 weeks, this short timespan can demonstrate some leading indicators such as number of documents processed. The financial benefits may take longer time; hence, overall benefits should be assessed with 3-5 years horizon for better TCO and ROI insights. Will try and share a template in later posts. Hope these pointers help in planning some of your automation initiatives!
MavenSpeak 