Enterprises are pouring money into contact center AI, with 61% spending over $10 million annually on CX delivery. Yet, a startling two-thirds of these organizations have limited visibility when it comes to quality assurance.
According to new research from TELUS Digital, while companies rush to deploy AI tools, only 32% have the automated QA infrastructure needed to measure their success. This severe enterprise AI performance gap means businesses are scaling automation without proper AI-assisted agent monitoring, risking customer trust and compliance issues.
The Measurement Gap
The TELUS Digital report, Enterprise CX AI: 2026 Global Survey, polled 815 enterprise decision-makers across 12 countries. It indicates that the most common approach to CX, across various teams, is now "human agents assisted by AI."
However, this rapid adoption has created a significant measurement gap. Without proper AI-assisted agent monitoring, organizations cannot evaluate interactions at scale. They are effectively relying on manual, subjective sampling to determine if their AI investments are actually improving the customer journey.
The survey found that 56% of organizations plan to invest in AI copilots for real-time agent assistance. However, only 46% plan to invest in automated QA infrastructure and coaching, with only 32% already deploying such tools. This gap between future AI plans and current QA realities means a significant portion of the market is actively scaling AI without the safety net needed to measure its impact.
As Peter Ryan, President and Principal Analyst at Ryan Strategic Advisory, summarized:
"Adoption of AI-powered solutions in CX has moved fast but enterprises haven't caught up to optimizing it quite yet"
He added that when looking to the future:
"The companies that will get a real return on their AI spend will be the ones that recognize that closing this gap is what turns AI deployment into performance."
The Cost of Unobservable AI
When enterprises deploy AI without robust AI-assisted agent monitoring, they shift the burden of discovering errors directly onto the customer. In traditional deterministic automation, a broken flow is relatively easy to spot. But AI agents and copilots adapt to context and language, meaning they can fail in unpredictable ways, such as hallucinating incorrect information or dropping context during a handoff to a human agent.
Tony Shen, Senior Product Manager at Amazon Connect Customer, recently discussed this exact risk with CX Today, noting that when an AI agent sits in the middle of a transaction, visibility is non-negotiable:
"On the human side, they need to know what decisions that AI agent made and have observability... or else they might be doing the duplicate things or might be confusing the customer"
This lack of visibility is particularly dangerous in highly regulated industries, where AI hallucinations and unpredictable model behaviors are not just technical glitches, but material compliance risks.
Justin DiPietro, Chief Strategy Officer & Co-Founder at banking AI platform Glia, told CX Today about the importance of accurate AI in the finance sector:
"In banking, whenever there’s payments, transactions, whenever people are making life decisions based on the information that their bank is giving, there’s not an option to be wrong"
According to him, the quality of AI output must be held to the highest standard, saying: "You can’t be probabilistically correct. You have to be 100% correct."
