You are here

How to choose the right RPA tool

public://pictures/ericka_c_0.jpg
Ericka Chickowski, Freelance writer

The robotic process automation (RPA) market has heated up as enterprises increasingly look to RPA bots to achieve their long-term automation ambitions.

Meanwhile the RPA tools market has been rapidly evolving as new vendors enter the market, mature vendors add new features, and RPA use cases expand beyond traditional applications in response to the platforms adding cognitive and AI capabilities.

However welcome these new capabilities may be, they muddy the waters for enterprises that are evaluating RPA tools.

TechBeacon talked to top experts and analysts to provide some clarity. Here are the variables to consider as you shop for the right RPA tools for your organization.

[ Learn how robotic process automation (RPA) can pay off—if you first tackle underlying problems. See TechBeacon's guide. Plus: Get the white paper on enterprise requirements. ]

Understand the different levels of RPA automation

RPA excels at helping organizations chain established processes with automation, all without ripping and replacing legacy systems.

The typical use cases have organizations leveraging RPA for what automation experts call "swivel chair" processes, where people engage in manual, repetitive labor that has them using data, information, or alerts from one system and taking an action in one or more other systems. Applications range from finance teams approving and processing employee expense reimbursements to help desk provisioning of new passwords.

But as RPA tooling has evolved, different strata of automation intelligence have emerged.

At the most basic level, RPA grew out of macros and scripts used to automate simple tasks. Then came basic, programmable bots based on human-defined process rules. When done well, these bots offered a bump in efficiency, but the discovery and definition processes could be time-consuming, and resultant rulesets could be fragile and inflexible to changing business needs.

RPA vendors responded by using machine learning to automatically discover and learn processes, and to define the rules that govern the bots. These kinds of cognitive capabilities are expanding the boundaries of RPA in terms of the types of automation the technology can provide.

RPA functions to look for

At their core, RPA bots are simple chunks of software, said Sairam Bollanpragada, head of the Global Delivery Center at Micro Focus.

 "Anybody who is a software expert can create a bot."
Sairam Bollanpragada

Still, the right RPA tooling can smooth out the process of deciding what that code looks like for any given bot, developing it, and deploying, maintaining, and managing all bots once they're deployed. 

RPA functionality varies, but here are some common functions to keep in mind when evaluating tool options.

How does it handle process discovery and definition?

How does the RPA tool discover and define process steps and components? Increasingly, vendors are building in the ability to record and analyze user actions and then use machine learning to help define process rules and reduce the number of manual steps.

Other tools put the onus on business stakeholders and the automation team to do all that manually. Usually, it requires the aid of consultants and integrators.

Integrators may layer their own tooling on top of the RPA vendor's platform to further automate discovery even before the organization has chosen what processes to automate. For example, DXC Technology has a digital process assessment tool that uses Google's TensorFlow to evaluate user behavior to identify the right RPA candidates.

DXC's tool takes a year's worth of historic data and analyzes transactions from beginning to end to figure out how many times a transaction occurred, and how long is each step taking, said Aniket Maindarkar, vice president and general manager of DXC's business process services practice.

"We get a good sense of whether you have enough volume, and if you have taken enough time to identify whether [a transaction] is a top candidate for automation." 
Aniket Maindarkar

How easy it to create bots?

Many RPA tool vendors claim to offer an easy way to create a bot-making "factory" within your organization. The toolsets typically include a design studio that provides automation experts with an easy way to set up RPA workflows based on discovered tasks within a process.

The amount of coding work that must be done varies, but many tool vendors are moving toward low-code or no-code features. The design capabilities are often aided by bot libraries that provide out-of-the-box functionality for common types of processes. But your "factory" will need to create bots tailored to your organization's needs, and that will take trained staffers to pull off. 

So picking the right tool for your organization may depend how much people power each tool requires to develop and deploy bots. Also, keep in mind that you should make decisions about RPA teaming in tandem with your tooling decision.

Many organizations struggle to find the talent trained in the specific platform they've chosen. The biggest challenge can occur when it's time to scale things out from limited pilot projects to enterprise-wide initiatives, DXC's Maindarkar said.

How well does it do bot deployment, maintenance, and management?

The RPA platform you choose should help you build a centralized function that allows you to use automation to easily deploy, maintain, and manage bots to ensure both the functionality and their uptime/technical performance. 

"You need an intelligent command-and-control center for the robots, which is where all your bots will reside, and where you monitor them for functional and technical performance."
—Sairam Bollanpragada

The centralized control center will help you ensure that the bots' technical and functional performance are at acceptable levels, alerting automation experts, business process owners, and technical leads when problems arise to make adjustments when necessary.

Functional performance should be dictated by your business process owners, who should ensure that each bot is doing the tasks it is supposed to do and that the business is getting the maximum amount of automation from the bot for any given process.

Technical performance is about ensuring that the technical infrastructure effectively supports uptime goals and other application performance KPIs.

[ Learn how to transform your IT with AIOps in TechBeacon's guide. Plus: Download the analyst paper on how AI is changing the role of IT. ]

Other things to consider

As you begin the process of evaluating which tools and platforms are right to develop and manage their RPA bots, keep your requirements front and center, said Mark Davison, partner and global lead for ISG Automation.

"We encourage clients to clearly understand what their requirements and expectations are for a given tool. Then we map those features and capabilities of the tool to those requirements and expectations to determine which best fits the client's needs."
Mark Davison

In other words, your RPA use cases should be driving your tool choices, rather than the other way around. This is fundamental to many IT evaluation processes, but bears repeating for a hype-prone category such as RPA, where bells and whistles may take focus away from business outcomes.

The key pitfall to avoid is to assume that every tool is the same, Davison said.

The tools that are the leaders in the market today offer a similar baseline of features, functions and capabilities, he said. But there are differentiators—how processes are discovered, the types of bot libraries the tools bring to the table, connections with other automation tooling. These characteristics, "when compared to a client's unique and critical requirements, can make a difference," Davison said.

Customers need to have a good understanding of all aspects, requirements, costs, and other factors concerned with installing, supporting, maintaining, and utilizing a tool before making a final decision, he said.

Don't forget to factor in services

Many organizations may find that the platform they choose is less important than how it is integrated into their processes and systems. According to experts, for every $1 spent on tools, enterprises spend approximately $3 on RPA-related services. And that trend will continue for some time, says the automation analyst firm Horses for Sources.

Ultimately, the goal of RPA is to speed up clunky business processes that involve multiple steps across legacy systems. DXC's Maindarkar warned organizations not to get hung up on shopping for features and functionality of RPA tooling that they lose sight of the fact that they're working toward a business outcome.

[ Enterprise Service Management brings innovation to the enterprise. Learn more in TechBeacon's new ESM guide. Plus: Get the 2019 Forrester Wave for ESM. ]