Three-month pilot compared Bots-equipped agent care against control group performance. Covid-19 Disruption
A managed care provider to more than six million members, our client is a Fortune 500 insurance company whose partnership
with ResultsCX is more than 14 years established. Over the course of the partnership, ResultsCX has routinely scaled the operation during Open Enrollment.
The steep volume increases sparked annually by open enrollment require rapid staffing ramps. Regardless of the agent tenure,
the ability to provide simple and accurate responses is critical−especially during Open Enrollment where new and vulnerable
populations are looking for plans and options for access to care. Moreover, with increased use of self-service channels, apps,
and portals, agent support covers increasingly complex questions that required navigation of multiple tools and resources.
This complexity causes longer training cycles and increases the time it takes to reach proficiency.
Difficulties experienced by agents in their first 90 days on the job are often marked by delayed responses to members’ questions,
long hold times, lack of confidence, and incomplete information–factors that ultimately degrade quality scores, customer satisfaction, and average handle time (AHT). With 2019 staffing ramps happening just in time for Open Enrollment volume increases, reducing speed to proficiency became our priority.
ResultsCX invests heavily in training, process improvement, and technology aimed at identifying and resolving customer issues.
To uncover the best use of our resources for accelerating speed to proficiency for new agents, we launched a pilot program comparing use of SupportPredict Bots to traditional agent support without Bots.
To prepare for the pilot, ResultsCX gathered voice of the customer intelligence using speech analytics, quality audits, and CSAT feedback. Identifying root causes of dissatisfaction from these sources is a proven means for bringing about sustainable change.
We then stood up SupportPredict Bots to streamline agent workflows and guide our agents through their calls with real time, relevant, and specific information. Now, instead of having to remember complex procedural steps and which tool to use in what instance, the right information was clearly presented to agents at the right time. As agents spend less time hunting for answers, they are freed up to focus on building rapport with members and creating an effortless, enjoyable customer experience.
Deployment of this initiative was led by a crossfunctional project team of IT engineers, Account and Site Operations, Training, Quality and Business Intelligence (BI). As training and reporting were customized to include the new Bot process, a brand champion was assigned to drive adoption and support. Bot utilization was tracked to ensure adoption, and the brand champion executed a communication campaign to ensure early and accurate use. Weekly performance reviews were held to identify and resolve any issues.
To minimize impact of external variables on the outcomes of the study, the Pilot trainees and Control group profiles matched in terms of demographics and work experience. Each group was trained and supported by the same resources and under the same
work conditions (schedules, environment, location).
Quality, CSAT and AHT were target areas for improvement.
- Quality audits measured accuracy and completeness of information provided to the caller.
- CSAT was based on a post-call survey.
- AHT was used to measure efficiency; prior studies indicate that customer experience is negatively impacted by long calls, and simple responses are valued.
Beginning Week 4, the Pilot group’s first week being measured, it outperformed the Control group across all three metrics.
Most importantly, the risk of new hires was curbed as evidenced by significantly higher performance in all KPIs in Month 1 of
Measuring for Accurate and Complete Information
In month 1, the Pilot group using Bots outscored the
Control group by 15%, achieving the desired speed to
proficiency. The Pilot group continued to outperform
the Control group in months 2 and 3.
Measuring for Customer Experience
The Pilot group outperformed the Control group by 12% in the first month and continued to earn higher CSAT scores in months 2 and 3.
Average Handle Time (AHT)
Measuring for Efficiency
The Pilot group again showed immediate and sustained improvement over the Control group when AHT was measured. In month 1, the Pilot group’s AHT was more than a minute lower than that of the Control group. The Pilot group continued to achieve a similar lead to the Control group in months 2 and 3 and beyond.
The pilot achieved its objectives across all three metrics. Use of SupportPredict Bots by the Pilot agent team reduced the number
of steps and time required to resolve members’ needs. The Pilot group’s lower handle times were also affected by shorter
after-call work time and reduced hold times. Additionally, the information provided to members was more accurate, timely, and
complete overall than that provided by Control group agents without the benefit of Bots.