Generative AI in the Contact Center: What’s New in 2025?

As contact centers experiment with generative AI (GenAI), they're moving from promising pilot projects to full-scale deployments.

These deployments are slowly reshaping how agents work, how leaders drive contact center efficiency, and how customers engage.

This roundtable unpicks the latest advancements shaping GenAI in the contact center and what innovations lie ahead for brands looking to stay ahead of the curve.

The industry experts participating this month include:

• Amy Roberge, Head of Global Contact Center Solution Engineering at Zoom
• Martin Taylor, Co-Founder and Deputy CEO of Content Guru
• Jay Patel, SVP and GM, Webex Customer Experience Solutions at Cisco
• Thomas John, VP EMEA at Five9
• Mike Szilagyi, SVP/ GM Product Management at Genesys
• Crystal Miceli, Vice President of Product and Industry Marketing at Talkdesk

Below, they share emerging use cases, challenges, and predictions for GenAI in contact centers.

Emerging Use Cases of Contact Center GenAI

Next-Generation Virtual Agents

Roberge: In 2025, the contact center is entering a new era of intelligence, one in which AI isn’t just generating responses; it’s taking action.

As this shift takes hold, virtual agents will move from static flows to autonomously completing tasks like order updates, appointment changes, or data retrieval.

However, automation alone isn’t enough. When escalation is necessary, human agents step in, and they must be just as empowered.

Thankfully, real breakthroughs are happening at the intersection of agentic AI and agent assist, supporting agents as they engage in complex, emotionally-charged conversations.

The future of CX isn’t AI vs. human; it’s AI setting the pace and humans delivering the finish.

Automated AI Quality Management

Taylor: Whether scoring for CSAT, empathy, or regulatory adherence, Auto QM is redefining the speed and scale at which contact centers can evaluate customer conversations.

Instead of sampling a handful of interactions, AI can audit 100 percent of calls and messages almost instantly, flagging risks and surfacing coaching opportunities in real time.

This shift transforms the auditor's role from a highly-paid finder of call recordings to someone delivering real value, spotting trends, improving processes, and driving performance.

For the agent, it replaces the dreaded, delayed performance review with almost instant, constructive feedback that supports continuous improvement.

Auto QM isn’t just faster; it’s fairer, more consistent, and allows quality teams to focus on what matters: enabling better experiences for customers and agents.

Real-Time Agent Assist

John: One of the most impactful emerging use cases for generative AI in contact centers is real-time agent assist.

Businesses are leveraging GenAI to surface contextual prompts, suggest next-best actions, and summarize interactions on the fly.

Indeed, 94 percent of business leaders are using AI to support agents live during customer interactions, per a recent Five9 survey.

Ultimately, that not only boosts productivity but also empowers agents to focus on complex, emotionally charged issues, where empathy and human judgment matter most.

In this sense, GenAI helps bridge the gap between speed and quality in service delivery.

Advanced Contact Center Reporting

Miceli: GenAI has revolutionized how contact centers create, manage, and deliver customer service.

GenAI supported self-service use cases using natural language, serving up just-in-time knowledge to agents.

Additionally, it improved contact center reporting, enabling supervisors and administrators to pull themes and opportunities for improvement from call recordings and transcripts.

The next wave of AI innovation that is taking hold in 2025 is agentic AI.

Now, the most innovative contact center solutions include not just agentic AI but also multi-agent orchestration in an end-to-end platform.

With multiple agents, brands can deploy a network of AI agents, each with a specific specialty, to collaborate and solve advanced customer queries.

Automating Routine Agent Tasks

Patel:

Generative AI first helped agents by automating note-taking, drafting customer responses, and automating after call work (ACW).

Now, AI agents are taking it one step further by automating routine queries, allowing human agents to focus on more complex, high-value inquiries.

These agents act as a 24/7 front door to a business - available across any channel, whether voice or digital - ensuring it can meet customers wherever they are.

Recent advancements in generative AI have enabled more natural, human-like interactions, delivering a truly conversational AI experience.

The ability to process and analyze vast amounts of data drives more personalized and accurate interactions at scale.

Additionally, integrating "agentic AI", which combines workflows and automation, enables seamless fulfillment of customer requests.

Challenges Companies Have Faced with GenAI in the Contact Center

Disconnected Systems

Roberge: For many companies, the promise of GenAI fell short because of a foundational issue: disconnected systems.

AI can only be as effective as the data it has access to. However, in most contact centers, customer data is trapped across siloed tools for voice, chat, CRM, WEM, and support.

This fragmentation limits the context GenAI can use to generate meaningful insights or take informed actions, creating inconsistent experiences for both customers and agents.

What we’re seeing in 2025 is a shift toward platform consolidation, where GenAI is embedded across a unified CX ecosystem, like Zoom’s.

Foundation Models Operating Within The Strict Guardrails

Szilagyi: One significant challenge companies face with contact center GenAI is ensuring that foundation models operate within the strict guardrails required by enterprise environments, while also delivering the responsiveness and accuracy essential for customer service.

Out-of-the-box models often lack the domain-specific knowledge and control mechanisms needed to meet these high standards.

To overcome this, organizations are investing in domain expertise to fine-tune models with the right data and applying rigorous governance to establish guardrails for accuracy, transparency, and consistency.

AI Hallucinations

Taylor: Hallucinations from GenAI haven’t gone away. They are real and can have catastrophic consequences in critical contact center scenarios. They’re also a fast way to erode trust among agents and customers alike.

Leading organizations are taking steps to improve AI accuracy and transparency to overcome hallucinations. Retrieval Augmented Generation (RAG) techniques are now widely adopted, ensuring generative responses are grounded against frameworks built on trusted internal knowledge bases.

In contact centers, especially, providing links to original sources allows employees to verify context in real time, forcing the AI to double-check its thinking and drastically reducing the risk of hallucinations.

Security And Data Privacy

Patel: One major challenge businesses face with contact center GenAI is ensuring security and data privacy, which remains a recurring concern when adopting AI solutions.