CX Moments That Defined 2025

CX Moments That Defined 2025

Big players are placing bigger bets on adding agentic capabilities to accelerate the speed to insights, context-aware conversations, and help the enterprise and its employees drive more value. Here are the big moments in 2025 that cemented the milestones in the future CX roadmap…

Parloa’s CEO and Co-Founder, Malte Kosub, was standing in a room full of CX leaders at WAVE 2025 in Berlin when he said, “Meaningful relationships are easy to build with people we know. Replicating that feeling between companies and customers… that’s the challenge artificial intelligence must solve.” 

In his opening keynote, Malte put forth a radical idea: every customer deserves their own AI agent – one that understands context, remembers every interaction, and can anticipate what the customer needs next. In short, if you have a million customers, you’ll have a million personal AI agents. 

The next era in CX is being built on the backbone of agentic AI. 

Leading CEM and CXM platforms are pouring massive investments into agentic AI capabilities—systems designed not just to automate tasks, but to act with autonomy, intelligence, and context‑awareness. 

These platforms are evolving to act as proactive partners that can anticipate customer needs, orchestrate journeys across channels, and empower human teams towards better CX delivery.

The future of seamless experience now depends on seamless collaboration between people and AI.

Logo Parloa

In October, Parloa demonstrated some of its new noteworthy product capabilities, including:

  • Adaptive voice guidance: The agent responds to what’s on screen, aligning contextually without overwhelming the user.
  • Barge-in functionality: Customers can interrupt or redirect at any point, and the agent pivots seamlessly.
  • Live API calls: Information isn’t static; the agent pulls real-time flight data, hotel availability, or claim status.
  • Background task handling: Agents don’t just chat — they act, sending emails, booking reservations, and closing loops.

Logo Salesforce

Earlier in March, Salesforce had championed its unified, LLM-agnostic platform. Agent-based AI is projected to automate tasks worth over $6 trillion by 2030, so while enterprises are focused on ROI, it has business leaders stuck wondering if a new LLM development will render previous partnerships obsolete

Salesforce’s suite of solutions offers customers flexibility and independence from any single model. AI solutions are becoming increasingly commoditised. So, this seems like a significant move. 

At the TDX ’25, Salesforce’s spring developer conference,  it unveiled its vision for how autonomous AI agents could usurp apps as the new user interface for companies interacting with their data. It also launched the world’s first agent marketplace, AgentExchange.

Logo Zendesk

In March at its annual Relate conference, Zendesk announced the launch of its Resolution Platform, a complete AI-first solution – purpose-built for service – to empower Customer Service, Employee Service and Contact Centre teams. 

The Resolution Platform, the company says, powers AI Agents that handle a significant portion of nearly 5 billion issues resolved annually, reaching more channels and customers than any other solution. 

These agents outperform legacy systems and traditional bots by effortlessly managing complex, multi-step problems using advanced LLMs like GPT-5 and the Model Context Protocol (MCP) for instant data access, streamlining workflows and accelerating problem resolution. 

This foundation has driven strong market adoption, with nearly 20,000 customers using Zendesk AI and a projected AI Annual Recurring Revenue of $200 million this year, marking significant growth.

Logo Qualtrics

Qualtrics announced the company was accelerating AI investments with a series of senior appointments to enable businesses to drive faster returns on their investments in its purpose-built AI capabilities. Mark Hammond joined as SVP, Core AI, and Jeff Gelfuso was promoted to SVP, Chief Product Experience Officer.

The moves come as adoption of Qualtrics’ specialised AI capabilities for experience management accelerates to deliver measurable performance and value. More than a third of its customers—including adidas, Stanford Health Care, Autodesk, Dollar Shave Club, and Stripe—have upgraded to Qualtrics AI innovations, including Qualtrics Conversational Feedback and Experience Agent, the company said. 

About 90% of the company’s top 50 customers use at least one AI capability, and the company’s purpose-built synthetic model, Qualtrics Edge Audiences, claims to produce results nearly identical to human responses, outperforming generic LLMs.

Logo Adobe

In September, Adobe announced the general availability of AI agents, powered by the Adobe Experience Platform (AEP) Agent Orchestrator. The company is also creating an AI platform for businesses to customise agents from Adobe and across third-party ecosystems — ensuring agents can understand context, plan multi-step actions, refine responses and more. 

AEP provides the foundation for agents to take contextually relevant actions and deliver ROI. Strategic partnerships with major companies, including Acxiom, Amazon Web Services, Genesys, IBM, Microsoft, RainFocus, SAP, ServiceNow and Workday, aim to ensure seamless execution of use cases. 

Previously, at the Adobe Summit in March, leadership announced its intentions to cater to a “new breed of consumers,” a new generation of consumers who see value in having conversational experiences with their favourite brands. 

According to insights from Adobe Analytics, a 1,200% increase in traffic to US retail sites from generative AI sources — and a 1,700% increase to travel sites proves that consumers looked for purchase guidance, inspiration and available deals. Agentic AI ushers in a new era for how businesses and consumers interact, with AI agents that will be able to handle more complex tasks and make highly tailored recommendations. 

Logo Intercom

In October, Intercom, at Pioneer (its AI customer service summit),  announced some significant updates to Fin, its AI customer service agent, in its third iteration. Fin 3 is custom-built for customer service, and operates as a flywheel with four distinct stages: Train, Test, Deploy, and Analyse.  

As part of the Train stage of the flywheel, Fin 3 includes Procedures, which enable companies to set Fin up for complex queries. With procedures, Fin is trained on everything human reps do.

There are four parts to creating procedures:

  • Natural Language instructions: These describe when Fin should use a procedure and outline the specific steps to perform. Adams said you could also copy and paste your Standard Operating Procedures (SOPs) in.
  • Deterministic control: Tell Fin to follow specific rules, like create a data connection, apply branching logic, or write code directly to have complete control of what Fin does.
  • Agentic behaviour: Fin reasons, thinks things through, and works out the next best step.
  • AI Assistants: You can have the AI assistant draft new procedures for you. You can tell it what you want, and it will pull context from other places to draft the procedure.

Logo SAP Emarsys

This past year, SAP Emarsys focused on turning chaos into control. Across four major product releases, the company delivered a wave of innovations grounded in real-time data, AI-driven automation, deeper channel integration, and increased platform flexibility.

The year began with foundational upgrades as it rolled out the Batch Sales Data API to make it easier to load large sales datasets with self-service configuration and data-quality monitoring. Engagement Events enables brands to connect external events from any data source to power segmentation, personalisation, and real-time marketing activations. 

These enhancements offer a real-time, data-driven engagement engine: data from anywhere → unified model → segmentation or automation triggers across channels. 

Logo medallia

Medallia’s new vision and AI capabilities were unveiled at Experience ‘25 at the Wynn Las Vegas in March. 

“This is a pivotal point for the industry and a time when enterprise organisations must move beyond siloed, survey-centric programs. These new AI capabilities enable our customers to understand and act quickly on all unstructured data from digital behaviour and voice and chat conversations, not just structured survey feedback,” said Mark Bishof, CEO of Medallia. 

In the past year, Medallia released more than 100 new features, including seven AI-powered product capabilities that accelerate the speed to insights and action using Medallia’s Text Analytics, unified platform architecture, and new generative AI. 

Other interesting AI-powered capabilities include:

Digital Experience

  • Prescriptive Digital Experience Insights proactively provide users with insights and recommendations to resolve digital experience issues.
  • Digital Session Summarisation enables organisations to quickly understand key behavioural events for digital sessions without watching a replay.

Contact Center

  • Coaching Intelligence empowers managers with summaries of previous sessions and Gen AI-driven, personalised coaching topics recommended for each agent.
  • Intelligent Summaries enable agents to act quickly. Call and chat summaries are automatically populated into feedback records and dashboard views for quick access.

Omnichannel

  • Smart Response creates personalised replies based on the content of a feedback record, freeing up employees to spend more time with customers.
  • Themes with Generative AI reduce time and effort to find more precise emerging trends with detailed, user-friendly labels and more frequent updates.
  • Root Cause Assist automatically generates a summarised root-cause analysis with the ability to drill down for more details, quickly answering pressing business questions without the need for time-consuming creation and modification of analytics and dashboards.

Logo Acquia

Acquia expanded its AI-powered SaaS CMS, Acquia Source, with the launch of three new integrated AI agents aimed at solving content bottlenecks faced by modern marketing teams. 

Announced in December, Acquia Source is built on Drupal and designed as a fully managed, enterprise-grade CMS that allows digital teams to create and manage content without relying heavily on IT support. 

The platform addresses key pain points such as fragmented toolsets, rising maintenance costs, and the growing risk of unstructured AI-generated content slipping through governance gaps.

The new AI agents embedded within Acquia Source include:

  • Source Site Builder Agent: Converts creative briefs into live, multi-page campaign sites, reducing build time and simplifying the path from concept to execution.
  • AI Writing Assistant Agent (Beta): Supports content generation optimised for both traditional SEO and emerging AI answer engines, applying Answer Engine Optimisation (AEO) best practices.
  • AI Web Governance Agent: Identifies and resolves accessibility and policy compliance issues to reduce brand and legal risks during content creation.

In an earlier blog, Jennifer Griffin Smith, Chief Market Officer at Acquia, stated that the time has come to move to adaptive storytelling – where the listener shapes the tale. In the age of AI, every user interaction changes the next one; this is a need-to-have. 

Like a seasoned storyteller adjusting their performance to the room, your site must adapt based on what it sees: device, language, previous actions, time of day, and more, she explains.

Think of modern digital storytelling as a modular approach. Each content block, CTA, product description, or testimonial is a piece of a much larger narrative. AI assembles and sequences these micro-stories into personalised journeys that change every time the story is told.

And not every audience is human. Jennifer suggests that content must satisfy multiple interpreters: the customer, the search engine, the large language model summarising your site, and the recommendation system feeding results downstream. 

If the story can’t be parsed and prioritised by machines, it may never reach a human reader at all. This is why structure matters. Metadata, tagging, schema markup: these aren’t just technical chores. They’re how you ensure your story is legible to both people and algorithms.

Conclusion 

The Agentic AI contributions to the CX team flip the old customer-service model on its head. Instead of routing people through departments, the experience follows the person wherever they go: across calls, chat, apps, and devices. 

Could this be the Rosetta Stone to decipher all customers while treating them like a segment of one? 

ALSO READ: The State of Customer Experience: What 2025 Taught Us