74% of brands using ‘action-taking’ AI see improved resolution rate
Nearly three-quarters of brands whose AI tools handle issues like processing a product return or updating a customer’s account are accelerating the number of closed cases, according to research published by Forethought AI Agents by Zendesk.
Based in San Francisco, where it provides agentic AI solutions that are designed to be self-improving, the company surveyed more than 600 CX leaders and practitioners across a range of company sizes, industries, and
geographies to produce the third edition of it State of AI in CX report.
The research distinguishes “action-taking” AI as focusing on a single process without human intervention and describes agentic AI as spanning multiple steps and systems. Self-improving AI, on the other hands, is designed to learn from every interaction to offer continuous improvement.
In contrast, “assistive AI” sounded more like generative tools that summarize and surface relevant content. Only 33 per cent of brands using assistive AI saw a positive impact on resolution rates, the report noted. Perhaps not surprisingly, the majority of brands surveyed are still in this stage.
“The pattern is clear: as AI moves from answering questions to owning more of the workflow, programs report stronger performance and
higher confidence,” the report’s authors wrote. “When AI works, buyers want more of it.”
360 Magazine Insight
Forethought’s report argues that using self-improving AI such as the solutions it develops builds the trust and confidence needs to expand the scope of AI-driven CX programs. Yet many brands may want to see how well assistive technologies work first before making a deeper commitment.
Not content to let AI maturity grow at a natural pace, Forethought says more brands should be aiming to operate at what it calls an AI ‘Centre of Excellence” level. It has provided five dimensions to define this, including training sophistication, business discipline and outcomes/measured improvement.
Sadly, despite showing 70 per cent of CX teams have adopted AI, the report found just two per cent have reached the Centre of Excellence nirvana. It’s hard to imagine otherwise given the complexity of many existing tech stacks, the challenges in transitioning workflows and roles and the rising concern about AI costs.
Those who download this gated 59-page report can decide whether they want to make scoring high on Forethought’s index a part of their strategic goals. Alternatively, they can continue applying assistive AI where it poses minimal risk to employee or customer disruption and layer on other capabilities when they’ve proven the value.
That said, there’s a lot more here that is well worth reading, including a detailed breakdown on the intersection of AI and leading CX metrics such as CSAT, cost per resolution and customer retention.
