Set Your Agents Up for +25% More Positive Sentiment with Conversational Intelligence
In the complex, ever-evolving world of healthcare, few things can be more disruptive to the member experience than a poorly communicated change to benefits. To help call centre agents prepare for member inquiries regarding the annual notice of change, a Fortune 100 healthcare insurer implemented a Conversational Intelligence Proof of Concept (POC) leading up to an open enrollment period in which costs would increase while benefits degraded. The initiative aimed to apply natural language processing (NLP) to analyse two key data sources over five months:
- Call Transcripts – Approximately 20,000 records, capturing dialogues between agents and callers, including metadata
- Voice of Customer (VoC) Surveys – Post-call surveys and core metrics (NPS) attached to each of those ~20,000 records
INSIGHTS
While the company believed it had standardised its agent training across all four contact centres, the data analysis yielded discrepancies in call-handling operations, which were leading to wide variances in member experience outcomes. The InMoment team went a click deeper–-using custom queries and Conversational Intelligence analytics—to identify some member-centric best practices that could be implemented across the organisation.
Advocacy Is Mission-Critical
Using text analytics to analyse call transcripts, the InMoment insights team identified that one particular call centre supervisor had been coaching teams to introduce themselves as the member’s ‘Healthcare Advocate’ at the beginning of the conversation—and it was having a big impact. Customer sentiment was 25% more positive on calls where associates introduced themselves as “Healthcare Advocates.” Unfortunately, this was only happening on 15% of calls—representing a quick-win opportunity for training agents.
Agents Must Be Adept at ANOC (Annual Notice of Change):
Calls mentioning the ANOC showed a 4-point higher NPS and were positively correlated with customer understanding and satisfaction. Unsurprisingly, agents that mention ANOC also tend to model other positive behaviors worth reinforcing, such as:
- Asking the member if they need additional help
- Using diffusion techniques
- Showing appreciation
Uncertainty Spurs Negative Sentiment
When members call in, they’re searching for answers from knowledgeable agents. Associates using phrases like “I don’t know” faced a 90% negative sentiment. The phrase was used frequently by agents with 6+ years of experience, indicating a potential area for coaching tenured agents and reinforcing best practices.
Confirmation Is Key to First Contact Resolution
The team also used AI Journey Insights to break down conversations into separate stages—from initial contact to post-call follow-up and everything in between. By breaking the conversation out by speaker, the team was able to apply the AI Journey Insights to the full call transcript as well as the agent’s and member’s individual portions. The analysis showed calls in which the agent confirmed resolution at the end of the call had an 18-point higher NPS.
Confirmation rates, however, were low across business partners at approximately 1 in 5 calls, highlighting a critical area where training could improve first-call resolution and customer satisfaction. While confirming resolution typically adds time (and thus, associated costs) to the call, doing so prevents the need for customers to call in a second time—making it a worthwhile best practice to implement.
BOTTOM LINE
Conversations are full of nuance—using AI-driven analytics to translate those nuances to insights can be a game-changer for your agents, your organisation, and most importantly, your members. This case study demonstrates the potential for using historical data and targeted AI-driven insights to drive measurable improvements in member satisfaction and operational efficiency in the call centre environment.