The Top Customer Service Trends and Technologies for 2025: Agentic AI Is Poised to Remake Self-Service
It’s long been known that excellent customer service is critical for companies’ bottom lines, and the numbers continue to support that. Research from The Harvard Business Review found that customers whose experiences with companies met or exceeded expectations spent 140 percent more than customers who had negative experiences. Additionally, customers with higher satisfaction levels are 74 percent more likely to still be customers a year later, compared to only 43 percent who rate their customer experiences as poor.
Additional research from Pegasystems uncovered the following:
- Seventy-seven percent of consumers say businesses need to invest more in improving how they interact with customers.
- More than half (56 percent) of consumers believe customer experiences have worsened over the past decade, despite advancements like artificial intelligence.
- Forty-eight percent of customers actively warn others about poor service, and more than a third switch to competitors after bad experiences.
It would seem that companies are finally catching on, evolving their use of AI, delivering advanced insight at the supervisory level, providing enhanced multichannel customer support, and enabling easier and more sophisticated customer self-service to enhance their customer service in 2025. Of those, artificial intelligence is the most common, though it is still very much in its infancy.
“AI is still the No. 1 thing that people have talked about,” says David Singer, global vice president of go-to-market strategy at Verint. “In the last couple of years, there were a number of AI pilots and experiments, but most didn’t go into production.”
According to Capgemini, only 13 percent of AI pilots did go into production.
Successful AI pilots have used excellent large language models for training; were trained on customer service data relevant to the company; and were focused on producing relevant outcomes to support the customer or to support the agent, according to Singer.
Firms with successful AI pilots have also established workflows and processes supported by AI rather than looking at AI as a solution in and of itself, Singer adds. “They think about the outcomes that they need, then they enhance and elevate them with AI.”
While generative AI got all the buzz a few years ago, now it’s agentic AI that’s making a huge splash in the contact center world. Combining AI, workflows, and processes, agentic AI, or autonomous AI, runs independently to design, execute, and optimize workflows, with AI agents making decisions, planning, and adapting with little or no human intervention.
Customer service leaders are among the largest early adopters of agentic AI, which Gartner expects to become commonplace by the end of the decade. By 2030, the research firm predicts, 50 percent of all service requests will be initiated by machine customers powered by agentic AI systems. By 2029, agentic AI will autonomously resolve 80 percent of common customer service issues without human intervention, leading to a 30 percent reduction in operational costs, it says.
“For service teams that are used to managing reactive demand from human customers, this shift will be monumental,” the research firm notes. “Overall service volume will likely increase, while access to customer feedback and sentiment will decrease.”
For contact centers, the strength of agentic AI is its ability to understand the context of the interaction and integrate into third-party systems, such as a CRM, to take action, fully resolving more complex customer queries rather than simply providing a response.
“When you put in the right genAI to answer the questions and the right agentic AI to automate the actions in the established workflow, it accelerates actions in the established workflow and elevates the service companies provide at a lower cost,” Singer says.
However, some companies have siloed their AI without combining it with workflows, limiting the benefits, Singer adds.
“Fragmented AI deployments are simply unsustainable,” agrees Patrick Martin, executive vice president of customer experience at Coveo. “As customers engage with businesses across an ever-increasing number of digital platforms, providing a truly seamless experience is paramount. The future of CX rests on cohesive AI strategies that unify data and deliver consistent experiences across all channels.”
Martin adds that true omnichannel support ensures customers can move effortlessly between platforms without losing the context of their interactions. This requires a shift toward flexible, modular technology that enables businesses to adapt quickly and integrate the best AI tools available.
Pegasystems is one of the companies involved in this form of AI. In February, it introduced Pega Agent Experience, a set of API capabilities in Pega’s workflow automation and orchestration solution to guide AI agents. Pega Agent Experience (Pega AgentX for short) enables AI agents to optimally complete tasks while invoking other AI agents to automate additional steps. The company says this approach will enhance customer service, improve employee productivity, and lead to end-to-end automation. The technology is designed to empower any agent, either from Pega or third parties, to identify and execute the best Pega workflow to guide them through requested tasks, including processing orders, updating customer accounts, and filing applications.
But Rahul Garg, vice president of product at Genesys, cautions that agentic AI isn’t a comprehensive solution but rather just a single piece of the much larger customer service puzzle. “In regulated industries, for example, you need AI to do a lot of things, but you also need deterministic workflows.”
Multichannel Experiences Add Embedded Support
While omnichannel strategies continue to be a priority, the next step is to deeply integrate these experiences into customers’ natural workflows, says Terence Chesire, vice president of CRM and industry workflows at ServiceNow. Rather than expecting customers to navigate to a help center or contact a service agent separately, embedded support enhances omnichannel strategies by providing assistance in real time within the applications and services customers are already using.
In retail consumer ordering and loyalty mobile apps, in-app support can offer instant AI-guided help when, for example, a promotional credit is not applied, eliminating the need for customers to switch channels or call customer service, Chesire explains. When using software such as workplace collaboration tools, embedded support might proactively assist employees facing technical issues, such as troubleshooting connectivity problems during a video conference without requiring them to leave the application.
“Rich embedded experiences will extend beyond traditional support to enhance the broader customer life cycle,” Chesire says. “Consider a commercial account onboarding process that appears seamlessly within a business banking portal, enabling a frictionless experience.”
Embedded support takes omnichannel strategies to the next level by reducing friction, ensuring continuity across channels, and delivering proactive, context-aware assistance precisely when and where customers need it, Chesire adds. It ensures a true continuum of care, meeting customer needs through seamlessly threaded interactions that maintain context across channels and agents.
Enhancing Supervisory Insight
AI not only helps agents provide better customer service but also helps supervisors coach those agents in offering the better service, Garg says.
“In a contact center, there might be only one supervisor for every 100 agents,” Garg says. “Companies are looking for ways to use AI to make that job much easier for them. We’re seeing a lot of push from our customers to bring AI into the back office so that supervisors’ tasks can be automated.”
That demand prompted Genesys to develop Cloud Supervisor Copilot and Genesys Cloud Virtual Supervisor, which were released in mid-March. Both products are designed to enable organizations to automate routine tasks and provide managers the support and insights to accelerate speed, improve work quality, and increase overall effectiveness. The capabilities offer organizations real-time assistance with analyzing data, training employees, overseeing processes, and handling critical business operations.
Genesys estimates that it’s possible to see a 40 percent reduction in quality evaluation time, a 25 percent reduction in multilingual evaluations, and a 38 percent decrease in quality management administrative costs.
Supervisor Copilot builds on last year’s launch of Genesys Cloud Agent Copilot to streamline contact center operations. It automatically summarizes interactions, highlighting key insights such as the reason for contact, the resolution, action items, and sentiment for supervisors to review.
Virtual Supervisor enables managers to automatically evaluate interactions. The AI summarizes all calls, including those escalated to the supervisor level, enabling better customer service because the supervisor doesn’t need to review an entire interaction transcript or have the customer repeat all of the information that was provided at the agent level. The technology provides customer sentiment details and can even provide real-time, in-call coaching for agents, informing them when they’ve gone off script, reminding them of mandatory disclosures, or providing other advice.
“Because this is improving coaching in real time, customer service is better,” Garg says. “By delivering information from the company knowledge base to the agent and in escalated interactions to the supervisor, the AI is also ensuring quicker interactions, which also enhances customer service.”
Easier Self-Service
AI’s growing impact is affecting other areas of customer service as well.
“From 2024 to 2025, we’ve seen a massive shift in terms of how customer service teams are leveraging AI and workflow automation,” says Rebecca Miller, senior product manager for CRM at Pega.
The focus in 2024 was on productivity help, such as meeting summarization and digital messaging. While those features are still used in 2025, the AI tools are now taking on more complex tasks, with AI agents completing tasks that previously required human intervention and automating end-to-end customer journeys, she maintains.
Combining conversational AI with workflow automation is critical for self-service, Miller adds. “We’ve been talking about self-service for many years, but we’re finally at the point where self-service is going into an entirely new landscape where customers are able to resolve more than just pretty simple tasks.”
Disputed credit card charges and other somewhat challenging customer service support queries can now be handled through AI agents, Miller explains. “Before, you had to escalate to a more traditional channel.”
By offering advanced self-service capabilities, companies can help control costs, but they can also service more customers via their desired channels, Miller says. She acknowledges, though, that it’s still challenging to balance self-service with the agent-assisted help that’s needed for certain interactions.
A Pega study found that 58 percent of workers are already using AI agents today. Among these early adopters, 41percent highlighted automation of tedious tasks as the primary benefit, followed by reduced time spent searching for job-related information (36 percent), and quick meeting summarization (34 percent).
Adoption of AI tools for customer service and support will continue to accelerate for another year or two, Miller predicts. “The goal is to provide an action network that is highly personalized. AI agents are going to become much more commonplace.”
The Rise of Empathetic AI
The contact center sector is witnessing a surge in new vendors offering faster, more cost-efficient, and improved capabilities. This, along with rising customer expectations around personalized interactions, is driving companies to set their sights on empathetic AI, the next evolution in personalized customer experiences, according to Sam Danby, head of voice at Boost.ai.
“Integrating AI agents with centralized knowledge bases has enabled AI agents to respond with account-specific information, an initial level of personalized interaction for every customer interaction,” Danby says. “However, while this shift toward empathetic AI is becoming an expectation, it is important not to forget the basics. Without a fundamental understanding of each caller and their interactions and without providing clear, accurate answers that enable self-service, any emotional or intelligent response adds little real value. Like humans, AI agents must make users feel heard and understood; otherwise, trust is lost immediately.”
To enable responses that include the appropriate level and type of empathetic response, a different dataset is necessary, Danby explains.
While customer account information is relatively easy to track using existing CRM-centric solutions, delivering empathy via AI agents requires the ability to dynamically adjust responses based on context and sentiment. By detecting either micro or macro variations in tone, pitch, cadence, or speech patterns, AI agents will be able to infer emotional states such as frustration, excitement, or distress, according to Danby. And that information is the key to effective customer service in 2025 and beyond.
Phillip Britt is a freelance writer based in the Chicago area. He can be reached at spenterprises1@comcast.net.