Robotic Process Automation at Scale: Innovating with bots

Using RPA, Arrow Digital partnered with a global financial services client to save $8 million annually, reduce risk, and increase quality.

Boosting Productivity with Minimal Process Change

Our client wanted to enable innovation without having to take on a large-scale technological uplift. Working with their leadership and technology team, we set our sights on decoupling a key impediment from its dependencies. Together we developed a solution using RPA to reduce and remove quality and risk management efforts for repetitive error-prone tasks.

Services Provided

  • Application Development
  • Robotic Process Automation

What Is RPA?

Robotic Process Automation (RPA) is the use of non-invasive software that interacts with applications in precisely the same way people do to fill in the gaps between systems currently filled by human workers. RPA liberates time and allows employees to execute higher value activities. It can reduce cost, human errors, and increase compliance and efficiency.

Enables Innovation

Addresses Obstacles

Reduces Vendor Lock-in


Together in the discovery phase, we created a road map tied to business outcomes. We gathered a cross-functional team of IT and business user stakeholders to determine key workflows that we wanted to automate. Then we did an analysis of the top RPA software vendors and determined the vendor(s) that best suited our target outcomes.

Utilizing our RPA maturity curve model as the foundation of our strategic plan, we laid out the key objectives that we wanted to accomplish.


In some use cases, a single bot more efficiently completes the defined tasks of one or more workers

We were able to remove the need for 6,000 hours of manual labor and saved $8M annually after a 3-month project.

For repetitive error-prone tasks, Quality and Risk Management efforts can be removed due to a single point of review.

Lessons Learned

1. Don't be afraid to automate a process that can only be completed 80% by a bot.

For a manual process, we learned that collaboration between a bot and a worker can create efficiency gains - even if only a portion of the process is automated.

2. Don't try to automate processes that you are defining at the same time.

Bots understand explicit instructions. Make sure you have your requirements, defined tasks, data points, and workflow specifications identified before diving into an automation initiative.

3. Identify the well-defined pieces of a process like building blocks.

We were able to leverage sub-bots, sequences of well-defined tasks, as building blocks that could be shared across multiple bots.

4. Decouple capabilities - Design with disposability in mind.

We learned that separating capabilities provides the flexibility to make major changes in the middle of a project with little disruption.

5. Consider security implications of a robot.

A bot runs in the context of a user - likely a user you will create specifically to represent a bot. If you put a bot in charge of doing a task, the bot requires role-based access to the system. If you run multiple bots, they will all need the appropriate access.

6. Don't be afraid of singles and doubles.

Rather than trying to hit a home run by addressing all of the technical capabilities which need extension or replacement, we were able to focus on incremental high-value investments employing RPA.

More Case Studies



CB Old Country Store

CB Old Country Store



Your vision, realized.

Start a Project