JORDANA GLOBERMAN

Robotic Process Automation Delivery for Service Canada Call Centres

Problem

I was hired to develop a development and delivery model for Robotic Process Automation in Service Canada call-centres across the nation, to alleviate strain caused by an influx of calls during the pandemic and support Canadians in need.

Role - Manager of UX & Strategy

While managing a UX & Strategy team, I created the development & implementation strategy for the project, established and maintained relationship with partners and stakeholders, designed research activities, led analysis, and managed a team of designers.

Team

Me + 1 senior designer, 1 junior researcher, 5 project partners.

We partnered with the operational branch of the call centres, Deloitte, and two other functional areas of our team (IT build & IT maintenance teams).

Context & Approach

 

In 2020, Covid-19 tore through Canada and the world. Health measures and illness meant many Canadians could not work. The Canadian Emergency Response Benefit (CERB) was implemented to support these Canadians, and to buoy the economy. CERB was new, so citizens had questions. These questions began to overrun Service Canada with calls.

 

Robotic Process Automation (RPA) was identified as a solution to the inundation of requests received by call centre employees. Robotic Process Automations are rules-based actions that can run simple processes for users, such as logging into a program or copying and pasting data. Service Canada employees had many repetitive, manual processes which could have been automated, giving agents time for the nuanced emotional work of supporting callers in distress, or the analytical complexity of offering advice. Our team was given 11 million dollars to build and roll-out processes across Service Canada's call centres. I was hired to design an efficient and human-centered approach to making this happen.

RPA.jpg

Tools

 

  • Digital Transformation Leadership

  • Generative and Evaluative Surveys

  • Co-design

  • Graphic Facilitation

  • Process Mapping & Service Blueprint

  • Affinity Mapping

  • Performance Metrics

  • Analytics Collection & Analysis

  • Remote Ethnographic & Contextual Inquiry

  • Remote User Testing

  • Collaborative Analysis

  • User-led Change Management

  • Behavioural Science

  • Capacity Building

I was charged with creating a development/delivery model for robotic process automation (RPA) in EI call centres across Canada. This included an approach to idea intake, development, and testing with users. This was a complex task on many levels. For one I joined an unstaffed team. As manager, I had to build up a team with skills in user research and design that could deliver these project goals. I recruited externally and brought in the required talent to fulfill our objectives. Another major project challenge was stakeholder management, so I focused on establishing roles and responsibilities with our partner teams at Deloitte and in the call centre operations. Beyond structural ways of working, I also put tremendous effort into building relationships with these partners as this had been neglected in the chaos of the pandemic and there was a fair amount of territorialism before I came onboard. I genuinely like making friends and was able to accomplish this without much trouble.

I believe digital transformation should enhance the established workforce and not the other way around, so I wanted our delivery model to be human-centered . I did this by involving the user at every pivotal juncture in the product delivery roadmap. I began with generative and evaluative surveys with all users to gauge their baseline experience with RPA, assess gaps in our current development and delivery approach, and gather ideas for future automation. I designed, implemented and analyzed these surveys myself while in the midst of recruitment.

 

These surveys revealed a big gap in the agent's understanding of RPA. There were also a lot of fears around this technology - that it would replace jobs for example - and these fears needed to be addressed through better user engagement. Once my team was hired, I led them in building a series of bilingual, collaborative, online workshops with users from across Canada, to address the problems identified in our survey and move forward in bot development. I led the design and coordination of these workshops and provided lead facilitation and graphic facilitation from all of them. Three 2-day workshops were delivered over the span of a month. We led participants in process mapping and the development of service blueprints online, only a few months after the organization had shifted to remote work. This was a huge challenge with a participant group with poor digital literacy but we adapted our activities and made the effort a success. One participant went so far as to email me to say this workshop was the most fun they had experience in their entire career in government.

A photo of seven people sitting around a table in discussion. One person is interpreting. Another man walks by in the background.

Participants told us these were the most inclusive events that they had been a part of, and were moved. It went a long way to build trust in government.

The workshops allowed us to redress misconceptions around RPA and to identify over 30 automation opportunities. We used affinity mapping to identify processes that were time consuming for call centre agents. We validated the potential of automating these processes with call centre performance metrics. This approach became our central process for intake, the cornerstone of our development and delivery model. This approach to intake also involved significant capacity building. We taught users about the benefits of RPA as they taught us about the ways in which we could maximize benefits, by identifying the most tedious and manual components of their jobs. We also created supports to help them better use RPA: a monthly co-design touchpoint with call centre team leads, a digital information hub on RPA for agents, as well as video tutorials and one-pagers to help support their work. We created a multi-channel communication model, allowing users to get support and share intake opportunities through meetings, websites, or email. This allowed users to become involved throughout RPA development cycles.

Development continued to involve users at every point. Once opportunities had been assessed collaboratively with users, the most promising one underwent rigorous user testing. I conducted remote ethnographic research with call centre agents as they carried out tasks with automation potential and documented everything. I gathered analytics on these processes to verify any assumptions that could come from observation alone. Then, our development team used these notes to build robotic process automations using UiPath. When RPA was rolled out, we used the same multi-channel communication model from intake to return to users, communicate the change, and offer further support. This model developed useful bots, while supporting capacity building, change management and human-centered design.

Result

 

Our delivery model involved the user in RPA development from intake straight through to implementation, and allowed for user-led change management to be embedded into the process through constant two-way communication. This ensured continuous improvement of automations based on user-feedback. As a result, RPA adoption increased dramatically across call centres by the spring of 2021. By summer, EI call centres across Canada were using these automations and RPA was providing huge time savings - sometimes nearly 20 minutes of per call - allowing the Government of Canada to serve more Canadians in need.

When the new budget was announced, the project was handed over to the team in charge of call centre operations. This was an achievement because it meant I had set up the delivery model in such a way as to make it sustainable without expert supervision. It could now function outside of a satellite digital transformation team, embedded into the organization's culture. In honesty, I would have loved to have taken the project further, reimagining the service design of EI operations, not just the design of certain tasks, but that, unfortunately, wasn't in scope for the project.