Redesigning enterprise-level AI alerts for Call Center Managers
Lead Designer & Researcher
Figma & Dovetail
Balto listens to 100% of calls and automatically alerts Managers when coaching opportunities arise for their agents within the Real-Time Coaching (RTC) tool. Managers are able to configure alerts for any of their Agents, and if a trigger item is hit during a call, Managers instantly get notified in the Balto Cloud, where they are able to quickly assist the Agent via chat.
I fully transitioned to this project in early 2022 to assist with standardizing & streamlining these Enterprise-level alerts.
Enterprise customers currently utilizing RTC to coach their Agents unfortunately spend most of their time creating, managing and aligning on their teams’ alerts rather than coaching and analyzing metrics. This is a result of the current alert management process being solely based on an individual-user level, which is not optimal for scaling; thus increasing the workload of many Managers-level users.
Enterprise-level organizations can range from having multiple management-level users, teams and playbooks that are interconnected and/or share common themes. This means it is critical that Balto's alert process allows for fast alert standardization across organizations through simple streamlining workflows.
Due to some gaps and questions my team and I had, we decided to continue discovery for Enterprise-level alerts by running a new research project. With time as a constraint, we primarily focused on attaining feedback from a more refined target customer group that this feature would benefit the most. By conducting new discovery sessions that lasted throughout our developmental process, we were able to ensure that we were building something that our customers not only wanted, but would enjoy.
As the primary end-to-end designer for this project, my process was:
Review all previously conducted customer interviews to get caught up-to-speed
done in our research repository, Dovetail
Conduct new research sessions with previous team's prototypes
in pursuit of a desired outcomeL gain a deeper understanding of our larger Enterprise customers–how their orgs are structured & who our key users are
Define our research objective
to understand how organizations are structured, & how different roles & responsibilities play into the creation and management of these alerts
Create affinity maps & personas
helped me visually see our users' pain points and areas that needed attention for the redesign
helped me identify two main user personas:
Admins: Focused on assessing, implementing & measuring team performance
Standards: Focused on coaching their team of agents, & will be main receivers of "assigned alerts"
Map out high-level user flows
Create wire-frame concepts & craft swift prototypes
Affinity mapping 💭
1. Production: Not scalable
RTC is the only area within Balto currently that is not organization-wide. Instead, users are only able to view alerts they have created for themselves. This results in a majority of Manager-level users being tasked with:
Allow for users to select from the top-down who (management-level) they would like to pass (assign) an alert to within the organization.
2. Version 1: Poor naming conventions
When it comes to roles and responsibilities, Enterprise customers are not all structured the same way.
Remove the naming confusion of "Managers" and "Supervisors" when describing the two workflows.
3. Version 1: Confusing views
Participants, regardless of their role, prefer to see all of their alerts in a single condensed view.
Combine all tabs into one view, which will allow for users to quickly see all of their alerts at once.
4. Version 1: Too overwhelming
Filtering capabilities are critical to Enterprise customers that have multiple teams and agents within their organization. Cluttered and lengthy pages lead to cognitive overload.
Add filters to the alert page which will allow for users to quickly view alerts that were assigned to specific users and/or teams on certain days.
TESTING & PROTOTYPING
Now that we had a new set of target opportunities along with possible solutions, I set out to create a new set of mid-fidelity wireframes that we could run A/B tests with. This allowed for us to cancel out any assumptions we might still have before over-committing to another less-than-optimal solution.
Updates 1 & 2
Update Balto's user management to allow for orgs to have Admin & Standard users.
Updates 3 & 4
Restructure alerts into one condensed view by removing the tabs from the Version 1 designs, with the use of filters.
With a continuous mindset, our team aims to deliver customer value by addressing underlining needs, resolving pain points, and satisfying desires.We are currently in the research and iteration phase; trying to gather as much customer feedback as possible before handing off designs to the Engineering team.