AI's role in transforming government call centers

Government agencies are under pressure to deliver a quality of customer experience (CX) comparable to the commercial sector, prompting investment in technologies that support more responsive services, increased transparency and improved customer satisfaction.

However, while businesses can measure return on CX investment through sales revenue and customer churn metrics, government agencies lack similar success indicators. This makes ROI analysis less straightforward -- but no less important.

Office of Management and Budget guidance from 2018 called for agencies to invest in call center experience improvements. Section 280 highlights ease and simplicity, efficiency and quality of service in call center engagements, as well as the helpfulness of agents, as key areas of focus.

Most significantly, it draws a direct link between federal call centers and securing the confidence and trust of the American public in government at large. And, with trust at a near-historic 17% low, every opportunity for positive contact counts.

The scope of opportunity

The government provides a range of services through its agencies’ call centers, covering benefits and loans to health care, Social Security, taxes and use of parks or historical sites. Some call centers operate during agency office hours, while others run 24/7, employing hundreds of agent shift-workers and handling upwards of 500,000 interactions per year.

Some operations rarely see change, while others must quickly scale to handle public concern around a national incident, such as health services handling questions on the coronavirus.

Throughout all contact center environments, however, runs one common element: data.

Ease, simplicity and quality

Artificial intelligence is a technology that thrives in data-rich environments. AI can act in several ways during a caller engagement to provide simpler and more natural, or  frictionless, interactions that build confidence and trust.

Its ability to act on natural-language voice prompts supports better call routing to the correct human agent. It can also deliver relevant and timely customer information in advance of -- and during -- the call, to help the agent give the most appropriate advice as well as pick up on action prompts, such as “send form,” “confirm by email” or “book appointment” so that tasks are completed quickly and with minimal human involvement.

For example, a federal employee reporting an accident in a fleet vehicle might need to talk to a human agent about the incident, but in the background an AI bot can commission a tow truck and arrange for a replacement vehicle. The agent has time to engage with the caller to provide better quality support while the bot handles scheduling and logistics.

However, a customer seeking pension advice may want to be able to talk through a range of complex options with an empathic expert -- a feature that an AI-driven bot simply cannot, at least for the moment, provide.

Bots can, though, identify when a customer’s frustrations are increasing. Raised voices or longer pauses in conversation can indicate anger, prompting the AI bot to flag a potential issue to a supervisor or to identify when to bring an agent onto a call during a process-led transaction.

Efficiency and cost

Whilst the federal mandate focuses on investment in the ease, simplicity and quality of service of call centers, AI has most the impact on improving agent efficiency and reducing costs.

Our experience with government and commercial centers suggests that, on average, one in five calls can be handled without human interaction. AI-driven bots can provide a simple, efficient experience for administrative tasks such as updating an address, resetting a password or requesting a loan form. This frees agents from 20% of call handling to be more productive elsewhere, and spend time on delivering a better quality experience where they are needed most.

There is further scope for handling process-driven parts of more complex conversations too. For example, AI can gather early-stage information and verify callers before handing them over to an agent to continue the interaction. Agents receive a briefing from the bot that combines new information received along with historical background, allowing them to join the call fully informed and equipped to deliver a helpful and more efficient service.

So efficient, if fact, that organizations we have worked with estimate AI lowers the average cost per call dramatically from $8.50 to less than $2.50, delivering significant operational savings.

Longer-term opportunities

As voice-based AI continues to mature, it will unlock opportunities to introduce more sophisticated services. With a strong culture of training and support among agency staff as a foundation, AI bots will progress from the routine queries to a more predictive approach that taps intelligent logic, allowing them to ask, for example, “Are you calling about your recent benefits claim?”

By aggregating current and historical information available on the customer, and continually learning from trends and refinements, AI-based systems will gradually learn to identify why customers are calling and enable government call center agents to present a smarter, more intuitive service.

Building confidence and trust in the government’s ability to deliver services has a foundation in the wide-ranging opportunities of AI to unlock new levels of ease, efficiency and quality of service to improve citizens’ everyday contact with the nation’s call centers.

Taken altogether, it's a strong case for investment in AI across the nation’s government call centers.