Increasing scheduling efficiency and transparency

I led the design and research for a resource booking tool used by administrators and customer service agents. The React app was integrated into our CRM and built with Microsoft’s Fluent Design system. It helped agents search TELUS Health’s roster of clinical professionals, understand the differences between suitable options, and complete a booking with more confidence.

The interface needed to move quickly without becoming a black box. We exposed the additional information that mattered at the point of decision so an agent could explain the recommendation, read the details back to the customer, and confirm the selection before booking.

The customer service team needed a better way to search for and book clinical resources. Agents were still relying on a legacy system that was inflexible, unstable, and missing some of the filters that mattered most. Location and distance were particularly important gaps.

The MVP only needed to support mental health services, but we did not want to design a mental health-only tool. The longer-term vision was a shared way to search and book any service in the TELUS Health ecosystem, including financial and nutrition services.

We also knew the Trauma counselling team would need the foundation later. Their workflow placed much more weight on geography and travel distance, so the search model had to be flexible enough to absorb more sophisticated location requirements without being rebuilt.

The business need came from our operations team, but the useful detail came from the people doing the work. I worked with stakeholders to define what success looked like, then connected directly with customer service agents through interviews and shadowing.

Shadowing was especially valuable. Booking was only one part of a live customer conversation, and agents were often listening, searching, comparing, and explaining at the same time. A flow that looked simple in isolation could become demanding when someone also had to keep the customer informed and avoid long stretches of silence.

We documented the current workflow, identified process gaps, and looked for the workarounds agents had normalized over time. We also reviewed existing scheduling products and Fluent UI patterns to understand which behaviours would already feel familiar inside the CRM.

Search broadly, then narrow with confidence

We organized the core experience around a filter bar and a results grid. The grid made it easy to scan across a roster, compare availability and other important attributes, and narrow the results without losing the context of the search.

Not every detail belonged in the grid. Additional information opened in a side panel so agents could investigate a promising professional without leaving the result set. This kept the primary comparison readable while still making the reasoning behind a selection transparent.

Pause before committing

After an agent selected a resource, a preview and confirmation page brought the important details together one last time. The page was designed for a practical behaviour we heard repeatedly: read the appointment information back to the customer, confirm that it was correct, then book.

We treated success and error handling as part of that same flow. Clear confirmation removed doubt after a successful booking, while actionable errors helped the agent recover without starting the search again or leaving the customer unsure about what had happened.

An early version separated common filters into their own section. The intent was to help agents get started, but the interface was making an assumption about how every search should begin.

In testing, agents did not need that distinction. It introduced another model to understand and used valuable vertical space, leaving less room for the results they were actually trying to scan.

We returned to the main success criteria: quick, error-free booking. We flattened the filter hierarchy, scaled back secondary UI, and focused the grid and confirmation experience on the information an agent could easily understand and read back to a customer.

Some of the efficiency gain came from the improved stability of the new platform, not the interface alone. I still consider the lower average time and smaller time-sink outliers meaningful because the complete experience helped agents finish more consistently.

The strongest longer-term signal was reuse. The Trauma team could build on the search and booking foundations with design involved in a lighter capacity, validating the decision to treat mental health as the first service rather than the only service.

A passionate group aligned around a clear mission can raise the ambition of an operational tool. By stating our UX goals early, we created allies within operations who helped us move beyond a like-for-like replacement and advocate for the details agents needed.

The project was also a reminder that the small improvements matter. Fixing the things users have quietly worked around for years can generate a lot of goodwill. It signals that the team sees the work as it is actually done and is willing to care about the details.