In this two-part series, we take an RPA adventure from the first bot to the one-hundredth. We look at the steps of growth and what to consider when embarking on your own Robotic Process Automation journey.
In part one of this two-part miniseries, our fictional example company embarked on a journey to build an RPA program. Starting with a small proof of concept, we ended up with 10 bots and a Center of Excellence (COE) to support our growing capability. We’ll pick up the story in part two and show how we went from a fledging COE to reach full maturity.
As a reminder, each organization is different, and the path to success will vary. We’re outlining one possible progression based on our experience and observations. I do not mean for this to be a roadmap but a guide to inform your own journey and anticipate challenges.
20 Bots: Grow Without Growing
Our COE is doing great, but we still can’t seem to keep up. Every department wants bots, lead times are getting longer, and more effort must be spent on production support. This is a great problem – there are loads of opportunities to create value – but it’s a challenge we must address.
As the COE continues to buy more licenses and deliver automation, we tackle our capacity issues by more fully utilizing employees from departments looking to benefit from automation. Sure, we started doing this when we launched the COE, but this phase is about turning them into full-fledged RPA champions.
It starts with training, not just for RPA awareness but also for business analysis and even citizen development. Standard training can be self-service or even used by champions within each department to engage their own team members. Next, we encourage the federation of duties, so each department not only identifies its own candidates but also assesses, prioritizes and even documents them. Departments now have a toolkit for onboarding to the COE as well.
We’re still doing the technical work, but business empowerment signifies a major shift from the COE out to the business. It also means we consider RPA for virtually every operational enhancement. In other words, RPA is moving from tactical to strategic.
Had we failed to address these growing pains, we could’ve expected a drop in quality with too much to manage, seen a decline in interest due to slow results, or both. Some departments lobbied not to take ownership at all – why can’t IT just do it? – but RPA programs can and should lean on business users, process owners and folks outside of IT. And where some programs may have stalled or plateaued, we’ve managed to keep excitement high and continue growing.
50 Bots: Enterprise Scale
Flash forward: RPA is thriving. Departments across our company are largely self-sufficient, but we’ve once again become victims of our own success. Problems now revolve around the technology platform itself, which has become unwieldy even for our experienced COE team. There are three central issues: reporting, maintenance and reusability. We’re expanding the COE technical team, but that’s not enough by itself. So, how do we streamline our processes and use automation to manage infrastructure better?
- Reporting: We build metrics into our automation designs, which then feed directly into business-facing, real-time dashboards. No more weekly reports to construct manually! Automated reporting creates a feedback loop for users to assess their automations’ business value better, then identify and help resolve pain points faster to create even more business value.
- Maintenance: We inject automation into our day-to-day COE processes, including automated support for basic tasks (like password resets), IT automation for spinning up new infrastructure, regular platform upgrades to get the latest features, and high availability architecture that has built-in disaster recovery. This increases performance for our bots while reducing the maintenance effort to support them.
- Reusability: We create a standard library of automation frameworks and reusable components along with a clear set of technical best practices. Standardized artifacts not only speed up development by providing off-the-shelf automation but also drive consistent, working bots across a larger development team with less oversight. Using pre-built components reduces testing and maintenance issues, too.
Our technical maturity ushers in new COE capabilities, most notably widespread attended automation that quickly increases bot volume. Though we’ve used attended bots for targeted processes before, we’re much more comfortable deploying and managing them broadly across the organization. At this point, everyone is confident in RPA at scale, which also means we can tackle more critical automations and ensure business continuity.
100 Bots: RPA in Your DNA
It’s taken a couple of years, but, suffice it to say, RPA has become a mature and prevalent solution in our example company. The next step is enterprise automation, where we add tools beyond RPA to our core automation competencies. Though we’ve built automation with other tools along the way, until now, it’s been opportunistic and not part of the core strategy. Time to change that! Here are the areas we integrate it into strategy.
- Business Process Management (BPM): RPA and BPM tools have a lot of overlap, with a fundamental difference that BPM is generally much better at human-in-the-loop processes. BPM gives us a whole new level of human integration, process orchestration and end-to-end case management.
- AI and Machine Learning (ML): AI and ML fill subjective gaps in processes that RPA is too rules-based to handle. Even though many RPA platforms have some built-in AI features, it takes dedicated tooling as well as expertise to make our solutions truly valuable.
- Intelligent Document Processing (IDP): IDP is a specific type of AI focused on handling documents (like PDFs or scanned images). We have documents across the organization, so the applications for this tool are pretty massive.
- Process and Task Mining: Over time, crowd-sourced ideas will eventually dry up. These tools help us overcome that, using data to identify and prioritize further automation opportunities that help us keep expanding. They also provide meaningful feedback for continuous improvement of automations already in place.
Adding these tools isn’t easy because they require different skills and collaborating with more teams (for example, we worked with data scientists on a custom ML model). Not only that, but they target the activities of knowledge workers, which gets more pushback than the rote, mundane tasks targeted by RPA. It also means increased support costs, increased licensing costs and making a business case to justify these increases.
Nevertheless, the benefits are clear. Where RPA limited us to only fragments of an end-to-end process, we’re able to capture a lot more of each process by combining complementary automation tools. Furthermore, we can now tackle processes that were totally beyond RPA’s reach – like human-centered processes – making automation not just a goal but an expectation for almost every process improvement.
Mission accomplished. Just kidding (sort of). The work is never done.
Reflecting on the Journey
Through this story, we’ve tried to show the path and pitfalls to RPA maturity. It started with a single bot, using RPA to automate a simple manual task and understand the hype around this emerging tool. It ended with an enterprise automation program encapsulating not just RPA at scale but other intelligent automation tools as well.
The RPA and automation landscape is quickly evolving. New tools will emerge to address some of the pain points we encountered, while new challenges will arise. The one likely constant is that processes will need improvement – and companies will need to consider process automation more and more to help.