Implementing Next-Generation RPA with a ‘Minimum Viable Bot’ Strategy
Analyst guest post Jason Bloomberg of Intellyx -- Part 2 of 4 in the Gen2 RPA series
January 4, 2022 – Jason Bloomberg, President, Intellyx
I recently attended the customer conference for one of the first generation or ‘Gen1’ Robotic Process Automation (RPA) vendors. After speaking with a number of this vendor’s customers, one common theme emerged.
Enterprise RPA initiatives are complex, time-consuming, and expensive. They require a diverse team of professionals to create and maintain the bots – typically by bringing SI partners onto the team.
These customers were generally happy with this heavyweight state of affairs, as the cost efficiencies they were able to achieve exceeded the costs, thus providing a positive ROI.
It occurred to me, however, that there must be a better way of implementing RPA – one that doesn’t only focus on the highest value bots that are necessary to cover the significant costs of the effort.
Understanding the RPA Long Tail
Gen1 RPA vendors are as successful as they are because their customers are able to implement a relative handful of bots that deliver this high ROI.
There are many more possible automations, however, that are simply not cost-effective to implement with these heavyweight, enterprise-class tools.
In fact, if you were to rank all possible automations from most valuable to the organization to the least valuable, the high cost of Gen1 RPA would prevent all but a small group of automations at the high end of this value ranking from returning a positive ROI.
The rest of the automations represent the long tail – automations that would only be cost-effective to implement if the per-bot cost of implementation were lower.
Changing the RPA Cost Equation
There are several RPA platforms on the market today that cost less than the offerings from the Gen1 vendors. To be sure, selecting one of these is likely to improve the ROI of some of the bots an organization might implement.
Nevertheless, the two primary limitations of RPA would still apply: RPA adds to technical debt, and bots are fragile, requiring ongoing (typically expensive) maintenance.
Simply buying a cheaper RPA platform won’t address these two challenges – especially if the organization is still approaching RPA implementation and maintenance efforts with a traditional enterprise lifecycle implementation approach. In other words, there is more to a second generation or ‘Gen2’ RPA tool than a lower up-front cost.
Remember, the per-bot implementation cost (factoring the licensing or subscription costs of the platform) may be higher for some RPA vendors than others, but the real long-term costs are in the maintenance.
The solution is to move away from the big-bang bot deployment model to a lightweight, iterative Gen2 approach that focuses on deploying inexpensive, flexible bots that are not only less expensive to deploy, but simpler to maintain than the heavyweight bots of the Gen1 RPA vendors – while simultaneously supporting whatever heavyweight bots the organization wishes to preserve.
Introducing the Minimum Viable Bot
To achieve this shift to lighter weight, easier to maintain bots, take a ‘minimum viable bot’ approach. This technique follows the principles of minimum viable products, or MVPs.
The MVP approach to deploying a finished product (software or otherwise) is to focus on the bare bones functionality that will meet a customer need in order to get the first product out the door. From there, incorporate customer feedback in order to improve the product over time – while it remains on the market.
The minimum viable bot approach focuses on implementing automations that meet the needs of the business in a way that allows for future iterations to improve the bots. Instead of a few, expensive bots, therefore, the minimum viable bot approach facilitates the construction of more bots – thus moving down the long tail of possible automations.
Minimum viable bots are also easier to modify as requirements evolve – a task that is far more difficult with expensive, heavy maintenance bots.
An Example of Minimum Viable Bots
Bots-as-a-Service solutions provider Thoughtful Automation had been building automation solutions with Gen1 RPA vendor UiPath, but the company found the high per-bot implementation costs impractical. “We were losing money to build that bot,” explained Alex Zekoff, CEO of Thoughtful Automation, “Just to prove the model worked.”
Instead, Zekoff wanted to drive his per-bot implementation fees to zero while adopting the minimum viable bot strategy to accelerate delivery times. As a result, clients would see immediate ROI, and Thoughtful Automation could add functionality over time as its customers required.
The company ended up choosing Gen2 RPA vendor Robocorp for its automation platform, because it wanted to control development at the code level via a low-maintenance, scalable cloud solution. Robocorp’s Agile AutomationOps methodology also aligned with Thoughtful Automation’s Agile strategy.
Robocorp fit the bill for supporting the company’s minimum viable bot strategy. “What used to take hours getting up and running, with Robocorp, I don’t have to deploy anything. It’s gone from hours to really nothing,” said Serhii Romanets, engineering lead for Thoughtful Automation. “I just have to focus on automation as code.”
Tackling Technical Debt
Another long-term win for the minimum viable bot technique is how it facilitates the lowering of the technical debt that heavyweight RPA adds to an organization’s legacy landscape.
Expensive bots end up locking in the user interface-based automations of legacy applications, making it difficult for organizations to retire the sunk cost of the bots.
With minimum viable bots, in contrast, the task of refactoring heavyweight bots by transitioning the business logic and workflows they represent is straightforward, especially leveraging Robocorp’s soon to be released automated conversion tools. Such tools also improve the reusability of bot logic, further lowering technical debt.
The Intellyx Take
Starting small and then iterating while taking into account dynamic business requirements is a fundamental Agile principle. Yet while most of today’s enterprises talk the Agile talk, when it comes to RPA, they struggle to walk the walk.
Gen1 RPA vendors have implemented platforms that resist Agile approaches for a reason – it gives their customers the ability to build and support expensive, high maintenance bots that provide a positive ROI.
For organizations who wish to extend RPA to the long tail of potential automations, as well as lowering per-bot costs generally, this heavyweight approach falls short.
The minimum viable bot approach, in contrast, is inherently Agile, as is Robocorp’s AutomationOps approach. Both these techniques can thus help organizations overcome the shortcomings of heavyweight RPA.
Copyright © Intellyx LLC. Robocorp is an Intellyx customer. None of the other organizations mentioned in this article is an Intellyx customer. Intellyx retains final editorial control of this article.