As 2025 comes to a close, Heads of CX are operating in a very different environment from January. The rush to buy AI tools because of impressive demos has shifted to a trend of tough questions about value, risk, and accountability.
Tim Banting, Head of Research at Techtelligence, says that this year "leadership teams changed the rules" and that CFOs and regulators now require evidence of business value.
Economic pressure and rising privacy enforcement have turned AI in CX from an innovation story into a funding and compliance story. Banting describes 2025 as "the prove it year" for AI in CX.
According to Banting, five CX trends stand out when reviewing the sector's innovation in 2025:
1. From impressive AI demos to outcome obsessed CX
The first trend is a shift from experimentation to accountability. For years, CX teams could gain support by showcasing AI proofs of concept. Leadership scrutiny has pushed CX leaders to define outcomes and success metrics before anything is deployed.
That has raised the bar for AI investments. Operating discipline and governance are now built into project design, not bolted on later. CX leaders who can express AI programs in financial and customer terms are finding it easier to defend their portfolios and keep budgets intact.
2. Compliance and privacy become CX buying criteria
Techtelligence search data shows a sharp pivot in buyer behavior. Banting notes that enterprise users are now digging into compliance and regulations, such as the California Privacy Rights Act (CPRA), and the SEC’s privacy rules for financial institutions (Regulation S-P) .
Topics such as data residency, consent management, and AI governance are also prominent among enterprise search traffic. Indeed, CX security, privacy, and compliance leads all show active topical research behavior, with around 26,000 businesses a week increasing interest over a steady three-month baseline.
This mirrors increased privacy enforcement in the United States, particularly around analytics tracking, session replay, and AI observability.
Banting outlined the four lenses to evaluate CX and AI vendors previously for CX Today.
Vendors have responded by pairing efficiency and revenue narratives with stronger compliance stories. For CX buyers, questions about where data sits, how consent is managed, and how AI decisions can be audited are now central to vendor selection.
3. From chatbots to agentic AI, shaped by customer expectations
At the start of 2025, AI in CX was often framed around generic chatbots and agent assist. Banting now sees a broader move towards agentic AI.
Enterprises want systems that can handle more complex workflows and actions, not just simple queries. At the same time, they are cautious. They want automation benefits only where "control audit trails and clear guard rails" are in place, he points out.