Natural Language Processing is changing the game for GX (Government Experience)
If you’ve made a purchase or renewed an insurance policy online in recent years, you’ve experienced how much customer service has improved. Pioneered by retailers, financial service providers, and the service industry – American households are served by technologies that have improved consumer engagement in myriad ways. One significant driver for these improvements has been better qualitative and quantitative analysis of interactions with customers. The resulting data and insights often challenge previously held service assumptions and offer businesses clear roadmaps for improving the customer experience (CX).
Maximus is now applying these best practices, including speech analytics and natural language processing (NLP) technologies, to enhance the delivery of government services, including enrollment to government-sponsored health programs. The insights we gain enable us to make government services more accessible for the public – even as we improve the accuracy of eligibility determinations and the effectiveness of achieving desired program outcomes.
Traditionally, key performance indicators (KPI) have been essential in measuring how well a contact center performs. For example, measures focusing on Average Handle Time have offered the principal means for evaluating a contact center’s contribution to the overall efficiency of the operation – and for business process services (BPS), measuring net profitability. Moreover, the growing reliance on workforce management has ensured that consistency and reliability in performance drive floor culture and management decisions.
Given the rise in speech analytics and natural language processing (NLP) technology, the government has a panoply of new tools to ensure consistency and performance in program efficiency. Moreover, the right technology can also ensure that customer experience is aligned to government outcomes and holistic definitions of program effectiveness. Maximus analytics-driven solutions leverage insights from each engagement to support the overall success of our programs and pave the way for the social and equity analysis of the future. Put more plainly; performance management can go far beyond compliance and performance measures – getting right to the heart of how government programs impact beneficiaries and ensure outcomes help people more fairly, efficiently, and equitably.
Natural Language Processing (NLP) overcomes the limitations of Average Handle Time (AHT) for evaluating program effectiveness
Health-plan education and annual enrollment periods have historically been one of the busiest seasons for government contact center operations. Millions of beneficiaries seek information on complex options every year, challenging agencies to meet the surge in demand for information and account services. Traditionally, Average Handle Time (AHT) has been one of the key KPIs that drive the training, capacity planning, and as a result, the approach to staffing to support the enrollment function.
This strategic program outcome has traditionally led to an intuitive understanding from both customers and operators; the lower the AHT, the poorer the experience and the lower the likelihood of meaningful health plan enrollment. A lack of program data is often cited as an issue validating the specific correlation. However, projects instinctively push for more efficient AHT to improve agent productivity, and therefore, AHT serves as a critical agent performance indicator. As a result, interaction analysis can operate as the missing link between customer intent and the alignment to their individual and program outcomes.
Maximus has developed a process of operational and interaction analysis to decompose KPIs using innovative techniques like topic mining and NLP to measure, as an example, the long understood but the little-measured impact of AHT on quality and increased rates of enrollment. With the ability to practically measure the topics, sentiment, and language used in serving and educating beneficiaries on their plan and provider choices, operations can now zero in on a full spectrum of relevant measures of quality. Managers can assess and address specific drivers of handle time, quantify meaningful time spent with each engagement, and prioritize the training and approaches used by successful agents, leading to higher customer satisfaction and more active enrollment.
Using the tools of KPI decomposition, new targets for AHT management can be evaluated to strike a balance for the employee and outcome. For example, Maximus’ experience has shown that higher AHT is driven by positive engagements (information sharing) and negative ones (poor system performance). Moreover, they’re often correlated. When used effectively, projects empowered with these insights can address the issues most meaningful to their agents, experiences, and outcomes to measure performance holistically and align KPIs to broader program and social outcomes in the future.
Where we go from here
Ultimately, this conversation reflects a broader set of questions about how aligning technology and people and processes can improve government services and deliver more for less in the public sector. For example, aligning performance monitoring – taking KPIs from rudimentary measures of compliance and cost and focusing on specific and relevant quality and efficiency benchmarks offers tremendous opportunities to improve the impact of programs at every level of government.
About the authors:
These capabilities are driven by the Maximus Performance Analytics Organization (PAX) to develop advanced analytics capabilities and total experience and outcome improvement for our government partners. Megha Gupta is a Product Owner, Julia San Roman is a PAX Data Analyst, Eric Stewart, Vice President, and Oana Cheta, Senior Vice President.