Enterprises Have Deployed CX AI. Now to Get it Working.

Most conversations about artificial intelligence in customer experience have centred on the same questions for the past two years. How widely is AI being adopted? Which functions is it transforming? Which vendors are winning? New research from TELUS Digital indicates those questions have largely been answered. Enterprise organisations have already deployed AI across their customer service operations. The challenge that now commands their attention is a harder one: ensuring those investments actually deliver measurable results.

The TELUS Digital Enterprise CX AI: 2026 Global Survey, based on interviews with 815 enterprise CX leaders conducted by Ryan Strategic Advisory in Q1 2026, makes clear that AI deployment is now a baseline expectation rather than a differentiator. Human agents assisted by AI have become the dominant delivery model across six of the seven CX functions the study measured, including customer onboarding, technical support, retention, and revenue generation.

The lone exception is concierge services, where high-touch interactions still favour human-only teams. Across every other function, the human-plus-AI model has taken hold, and the data shows that enterprises have reached a working consensus: AI performs best as an augmentation layer, extending the capabilities of human agents rather than replacing them outright.

Separate research from Metrigy reinforces this picture. The firm's global Customer Experience Optimization: 2025-26 study found that more than four in five organisations already operate with humans and AI working alongside each other to serve customers. The adoption question, in other words, is effectively settled.

The more significant finding in the TELUS Digital research is not how widely AI has been deployed, but how much of the supporting infrastructure necessary to optimise those deployments is still missing.

The survey charts a consistent gap between what enterprises are currently using and what they plan to invest in over the next 12 to 24 months. AI copilots for real-time agent assistance are currently deployed by 36% of respondents, yet 56% plan to prioritise investment in them. Intelligent knowledge management sits at 34% current deployment against 51% planned investment. Automated quality assurance and coaching stands at 32% today, with 46% intending to build it out.

The pattern is striking. Many enterprises have deployed customer-facing AI, such as chatbots, virtual assistants, and self-service tools, without yet building the agent-side infrastructure needed to optimise performance and measure outcomes. The customer-facing layer has gone live but the operational foundation has not kept pace.

Enterprises are not primarily chasing fully autonomous AI experiences. They are focused on making human agents more effective in an environment where AI is already present.

The capabilities attracting the most planned investment, namely copilots, intelligent knowledge management, automated QA and coaching, and predictive analytics, all serve that purpose. Copilots surface real-time guidance during live interactions. Knowledge management ensures agents have accurate, contextually relevant information at the moment they need it. Automated QA compresses the feedback loop between performance and improvement.

Metrigy's AI for Business Success research found that the majority of companies had already recorded measurable returns on their AI investments, with gains reported across sales performance, operational costs, employee efficiency, and customer satisfaction, but also identified areas of ongoing concern. The companies seeing the strongest returns are those investing in the capabilities that directly support how AI and humans work together, not simply those that have deployed the most AI tools. Understanding what that return actually looks like in practice is increasingly where CX strategy conversations are beginning.

As AI becomes embedded across the customer journey, the ability to understand whether it is performing as intended has become essential. The challenge is not a small one. A KPMG poll found that just one in seven business leaders has put formal metrics in place to measure AI returns, pointing to a wider enterprise struggle to connect AI investments with tangible business outcomes. In customer experience, where AI operates in real time across high-stakes interactions, capturing performance and building an effective measurement framework is particularly importance as it carries such direct consequences.

This is where a wave of vendor investment is now concentrated. Cisco's announcements at Cisco Live 2026 are indicative of the direction. The company unveiled AI Agent 360, a monitoring and observability capability for AI agents that provides real-time visibility into agent behaviour, performance, and escalation patterns across the full lifecycle. Alongside it, Cisco launched a rebuilt workforce engagement management suite designed specifically for environments where human and AI agents coexist. These point to a recognition that managing a blended workforce requires fundamentally different infrastructure from anything built for human-only teams.

Cognizant has approached the same problem from a governance angle, integrating its Neuro AI Trust platform with ServiceNow's AI Control Tower to provide persistent oversight of AI behaviour across the full AI lifecycle. The goal is to move organisations from fragmented, periodic compliance checks to continuous, real-time visibility into every AI asset in operation. For CX leaders managing customer-facing AI at scale, that kind of persistent assurance is quickly becoming a practical requirement rather than an optional layer.

Genesys has taken a similar approach, embedding governance controls and transparency mechanisms directly into its cloud platform for agentic AI. A Genesys survey found that more than a third of CX leaders do not have formal AI governance policies, a proportion that represents a significant operational risk as AI systems take on more complex and consequential responsibilities.

Jamie Timm, Global SVP of Service Delivery and Operations at TELUS Digital, summarised the challenge facing CX leaders: “What we see in our delivery centers every day is that the enterprises getting the most from CX AI are the ones who started with outcomes, not tools. Having the right foundations in place to access intelligence and measure for results is what separates deployment from performance. This research reflects that pattern.”

The next phase of enterprise CX AI is unlikely to be defined by who has deployed the most AI. The organisations that emerge ahead will be those that treat AI as an operational capability requiring continuous management. Deployment was just the starting line. The race is now for performance.