As 2025 kicks into gear, it’s clear that last year was pivotal for customer experience (CX). Generative artificial intelligence (GAI) continued to monopolise attention, making bold promises for the way businesses interact with customers. While some AI early adopters saw gains, others struggled to demonstrate a clear return on investment.
AI regulators globally made significant strides following the release of ChatGPT in November 2022, laying the groundwork for more robust AI governance, notably through the EU AI Act.
These trends have set the stage for even greater change in 2025, especially as the UK government declares AI a strategic business imperative and a new US government. Organisations will need to learn from the lessons of 2024 while embracing new opportunities. This article makes three predictions for AI in CX this year.
Prediction #1: Generative AI gets a reality check – but succeeds in CX
2025 looks like a year of setbacks for generative AI and its hopes of world domination. The technology is descending from the second of the five phases in the Gartner Hype Cycle, the “Peak of Inflated Expectations”, characterised by more buzz than proof, and into the third: the “Trough of Disillusionment”.
Many organisations are finding that realising efficiency gains from AI is a more complex business than initially expected and early AI adopters see their initial excitement (and funding) dry up, with businesses reporting performance issues and lower-than-expected ROI. Gartner predicts that 30% of generative AI projects will be abandoned after proof of concept by the end of this year, due to poor data quality, inadequate risk controls, escalating costs, and unclear business value.
The notable exception will be in CX, where generative AI is set to prove its worth in practical ways and begin scaling the hype cycle’s “Slope of Enlightenment”. AI’s central role in creating streamlined, hyper-personalised interactions are getting started. In 2025, we’ll see AI shift from high-risk experiment or IT hobby project to an essential tool that helps businesses deliver more personalised, more efficient customer interactions at scale.
Prediction #2: AI regulations tighten
With AI dominating the public conversation for over two years, it’s no surprise that regulators are stepping up to the challenge. As global governments react, and sometimes overreact, to public and industry concerns, AI’s rapid development will be matched by an increase in rules and compliance requirements over the coming twelve months.
While a global AI standard remains unlikely, probably ever, 2025 will see regional regulators adopting frameworks that closely align with the rules of AI’s pacesetters, the EU, where impressive progress has already been made to define standards and address key issues around transparency, data security, and ethics.
With AI’s widespread adoption in the CX space, leaders should keep a close eye on proceedings to stay ahead of upcoming regulatory changes, whilst embedding a dynamic compliance regime into their operational processes. The challenge will be the pace of upcoming regulations- and the winners will be those organisations who can keep their AI-powered CX on the right side of a constantly shifting legal line without disrupting service.
Prediction #3: Agentic AI steps into the spotlight
The next step in AI development will focus on autonomous decision-making, shifting from “human- in-the-lead” to “human-in-the-loop” systems. 2025 will see so-called ‘Agentic AI’ begin to carve out a niche in CX, focusing initially on low-risk, goal-driven applications where AI autonomy can deliver measurable value.
Until now, Large Language Models (LLMs) were commonly used to generate content; Agentic AI enables LLMs to take on more agency and carry out tasks based on prompts from external events. Under adequate human supervision and with appropriate regulation, such capabilities will enable the emergence of a goal-driven machine workforce that autonomously makes plans and takes actions at scale, using a degree of sophistication that was not previously possible.
As Agentic AI takes on more decision-making roles, emphasis on transparency and governance will increase. Businesses adopting Agentic AI will need to balance the benefits against the risks of diminished human oversight, such as the legal repercussions of any hallucinations or other wrong information, and the need to remain compliant with future data regulations, such as the UK’s Data (Use and Access) Bill.
Embracing the future of CX and artificial intelligence
As we race into 2025, the CX industry has a unique opportunity to redefine itself from its contact centre origins. And the contact centre, now fully omnichannel and AI-powered, will, of course, remain a key component in the CX machine. Generative AI may be facing its reality check year elsewhere, but when applied to CX, it’s set to be transformative in creating personalised engagement that creates a competitive advantage at all levels.