How Contact Centers are Using AIHow Contact Centers are Using AI
People are excited about the potential for agent assist or customer-facing scenarios, but at the moment, the highest value is coming from more mature use cases.
October 6, 2023
It’s always seemed logical that AI’s first strong use case in the contact center would be in an agent assist capacity. Generative AI in particular might not be ready to show its face to the customer just yet, but surely it can help agents perform tasks like call summarization, and provide real-time information and recommendations to the agent while the human is interacting live with a customer. But it turns out that while agent assist may be the most popular use case, other functions are more effective, at least for now.
That was a key takeaway from a presentation David Myron, principal analyst at Omdia, delivered during a No Jitter/Enterprise Connect webinar last week. Myron offered up some findings from Omdia’s recently released State of Digital CX:2023 survey, focusing on where AI is being implemented now, and where enterprises believe it’s working best.
Agent assistance topped the list of AI features already deployed: 64% of respondents currently have AI in place for agent assist, and another 30% will deploy it within the next 12 – 18 months. Other functions with over 50% deployment included call classification, survey and review analytics, Web-based chatbot, and intelligent call routing.
But Myron pointed out that when Omdia asked which AI features added value in the contact center, agent assistance ranked eighth, with 48% saying it has “significant value.” And Web-based chatbots were dead last on this measure, with 41% of respondents ascribing “significant value” to them. Myron noted the good news here—when you add in those who assign these features “moderate” value, you get a total of 89% saying that both AI has at least some value in agent assist and Web-based chatbot use cases.
Still, the most valuable AI features, according to the Omdia survey, are intelligent call routing and social listening and ticketing. For each of these features, 56% of respondents ascribed “significant” value, and another 41% assigned “moderate” value. Also near the top were digital CX analytics for QM, CSAT, and training; call classification; caller intent; intelligent knowledge base; and survey and review analytics, all of which received a ranking of “significant” value by more than 50% of respondents.
Another interesting set of datapoints: Omdia asked the top reasons why respondents felt AI delivers significant value on the one hand, or limited value on the other, and here were the results:
Significant value:
Employees know AI’s importance; it’s adopted broadly (66%)
A strategy is in place that prioritizes AI outcomes (63%)
Have comprehensive AI data management and governance programs (62%)
Limited value:
Lack of applicable use cases (43%)
Not capturing or quantifying results (43%)
AI doesn’t integrate with existing business systems (35%)
These results shouldn’t be too surprising, and they resonate with the state of the industry right now. People are excited about the potential for agent assist or customer-facing scenarios, but at the moment, the highest value is coming from more mature use cases, or those that don’t require AI-human interaction.
The message for enterprise decision-makers seems pretty straightforward: Plan for AI’s disruptive, transformative use cases, but don’t miss out on the incremental improvements available in the short term.
