Virtual agents: A thin line between love and hate
September 29, 2020
If you had lingering doubts about using artificial intelligence (AI) in customer service, COVID-19 likely obliterated your uncertainty. As support needs surged, companies successfully tapped into AI to play two important roles: enabling fully digital virtual agents and supporting human agents with real-time tools to make them even more effective.
<<< Start >>>
“There’s a lot to love about AI’s potential to enable faster response times and on-demand scalability while reducing cost to serve.”
<<< End >>>
For customer service organizations, there’s a lot to love about AI’s potential to enable faster response times and on-demand scalability while reducing cost to serve. Yet many organizations are still finding their way. Accenture’s Keep Me Index research – which scores how a company’s customers think, feel and talk about the brand – also offers some valuable insights into what’s working and what needs to be tuned.
One chance to make a first impression
For starters, our research into the moments that impact customer experience suggest that chat-based support doesn’t always delight. For now, that channel is less likely to handle customer needs end-to-end – resulting in high abandonment rates and a consequent negative impact on customer satisfaction.
Chat, including live and AI-based virtual agents, had the lowest first-contact resolution rate of the channels we studied. One-third of customers who started in chat needed or wanted to go elsewhere to complete their inquiry or transaction. In fact, only six of the 40 communication service providers that were part of our research had above-average satisfaction rates for forms of interaction that relied on chat alone.
A major issue: Whether live or virtual, many chat agents just aren’t making a great first impression. Among those who said they wanted to use chat for their next interaction, 62% hadn’t tried it before but were willing to give it shot. But among those who exclusively used online chat, less than half (48%) would want to do so next time.
Our experience suggests that the best solution is one in which virtual agents are designed to handle the bulk of transactional questions and requests while empowering customers to switch seamlessly to a live agent as needed.
A virtual agent alone is not enough to maintain a strong customer relationship in all instances. In our study, 60% of customers who only chatted with a virtual agent were satisfied with their interaction. Satisfaction was 10 percentage points higher when a human was engaged in those interactions.
An integrated approach to chat allows for the largest impact in driving desired experience factors and customer satisfaction.
<<< Start >>>
Source: KMI Moments that Matter Consumer Survey - Global Respondents Surveyed June 2019
(total n=18,003; n=2,737 chat as first channel; n=936 very satisfied with chat)
<<< End >>>
In other words, a human + virtual approach to chat drives the largest impact in driving those experience factors.
<<< Start >>>
<<< End >>>
Strike the right balance
As you work to balance human + virtual resources, keep these factors in mind:
- Consider how virtual interaction will look and feel at different points in the customer journey. Design the customer journey end-to-end, across all channels, using service design and intelligent dialogue flows to drive higher digital channel adoption while improving consistency, quality, productivity and relevance. Not all customer situations are best served by virtual channels; focus on those where AI truly adds value to the customer experience.
- Use AI capabilities to improve customer insights. The measuring rod for omnichannel interactions is the effectiveness of the customer relationship – which hinges on knowing a person. Here’s where advanced AI capabilities come in, as companies can base a virtual interaction on deep, extensive knowledge of a customer’s history with the company. If a switch to a live agent is required, the agent should also have access to that information (along with knowledge of what just transpired with the virtual agent).
- Context is king – Make virtual agents proactive. You don’t want customers to switch channels only because they feel defeated. In fact, effective AI should be able to detect and respond when frustration is setting in. So, start with a virtual agent for straightforward requests; use automation to offer relevant suggestions and resolutions; and be proactive about the virtual-to-human handoff.
- Never stop improving; knowledge is power. Know where the virtual channel is working and where it’s not. Achieving success requires measuring all parts of the experience and understanding what customers are doing before, during or after each interaction. For example, how or where you activate the virtual agent can influence engagement. So can the look and feel, tone and personalization of the experience. Finally, use data and customer feedback, to identify and prioritize improvements to the full customer experience. Align your organizational and live agent incentives to drive superior customer experience outcomes. And consider how automation and AI can be leveraged to drive efficiencies, growth and cost reduction in other channels.
<<< Start >>>
<<< End >>>
The bottom line? Virtual agents aren’t just a cost-cutting solution – they’re a vital component of transformative customer service at scale. As technology continues to advance, and as machine learning makes virtual agents ever smarter, you have even more potential to create AI-based agents to handle a wider range of customer intents. Think about virtual agents as an extension of your workforce. Create a dynamic human + virtual hybrid solution now – and commit to continual improvement – to deliver service your customers will love.
