New Research: AI Service Agents Are Scaling and Delivering CSAT

Not long ago, discussions on service teams - and at Salesforce - focused on expectations for AI agents rather than results. That's changing fast. Salesforce's new report, State of Service: AI Agents Edition, surveyed 3,075 customer service professionals worldwide and found that agentic AI in customer service has moved from pilot programs to mainstream deployment - and the most improved metric isn't operational efficiency but customer satisfaction.

"This research shows that agentic AI in customer service has moved from promise to proof. But what's really exciting is that agents are actually improving how customers feel about the service they receive." - Kishan Chetan, EVP & GM, Agentforce Service, Salesforce

Agentic AI in customer service has reached a tipping point

AI has become a standard part of the customer service toolkit, and 85% of service organizations now use at least one form of AI. But the most notable shift is in agentic AI - autonomous systems that take action and support complex workflows. Sixty-six percent of customer service organizations now use agentic AI, up from 39% in 2025 - a 1.7x increase.Seventy-seven percent of service teams with AI agents deploy them in both customer-facing and internal operations. On the customer side, top use cases include proactive outreach, personalized product recommendations, and multichannel case resolution. Behind the scenes, agents handle operational work that supports those interactions, like routing cases to the right agent.

"Limitless Service is the new era where AI doesn't just assist but takes action. When every department that touches a customer - service, sales, marketing, field operations, and employee support - works from the same information, they can act in a coordinated way. The companies that deliver these connected service experiences will move faster and capture more value." - Annie Weinberger, CMO, Agentforce Service, Salesforce

AI service agents improve customer satisfaction

Customer service teams are seeing results quickly: 70% of organizations with AI service agents say they observe measurable value within 60 days of deployment.

When respondents were asked which KPIs have improved the most since deploying AI agents in customer service, customer satisfaction was at the top of the list - ahead of service rep productivity, average handle time, customer retention, and first-response time.

That finding is notable because customer satisfaction isn't strictly an operational metric. It reflects how customers actually experience service - and research consistently links it to retention and lifetime value, suggesting the efficiency gains from AI agents may run deeper than operational efficiency alone.

Ultimately, confidence in AI service agents is high: 89% of service professionals with AI agents say their organization would benefit from expanding their use.

"The goal is to make support so seamless it almost feels invisible. When we get this right, the customer doesn't just walk away with a resolution; they walk away with their time back. And in a world where everyone is stretched thin, giving someone their time back is the highest form of service we can offer." - Bradley Condon, Salesforce Administrator, Waste Solution Services

Customer service teams reimagine their processes for AI

Adopting AI service agents is more than a technological shift. Ninety-seven percent of customer service leaders with AI say it's impacting their approach to workforce planning.

These changes sometimes mean creating entirely new positions to support the technology, like architect roles to oversee deployment and operations and data management roles to maintain knowledge bases for AI.

Survey insights suggest the data management struggle in particular may be underrated.

Seventy-two percent of service operations professionals say data readiness is a major blocker to AI - compared with 59% of customer service leaders. In other words, the people working with the data day to day report higher concern than leaders.

Additionally, organizations that have already deployed AI are more likely to anticipate growth in data management roles than those that haven't yet - suggesting the data challenge may only become fully visible once AI is live.

"I attend as many in-person events as possible. Not only are the sessions informative and engaging, but the Trailhead Community is always there to help each other out. Right now, learning AI tools and understanding data is critical." - Magon Mair, Director, Presales Solution Engineering, Wilco Source

Trust in agents grows With experience

Service professionals who have deployed AI agents are more likely to personally trust autonomous AI than those who haven't - suggesting experience may build trust.

Sixty-five percent of customer service professionals say their customers fully trust AI. However, research suggests customers may be more hesitant than service teams believe. According to the Metrigy Consumer CX Index, consumer trust in AI-powered service remains relatively low: just 44% of consumers trust AI to handle their customer service needs. However, when customers actually experience AI-powered service, skepticism tends to give way, with customers reporting that it exceeds their expectations.

"What we're seeing is that trust in AI agents isn't built through promises - it's built through experience. The more service teams and customers interact with these systems, the more confidence they develop in what AI can handle." - Annie Weinberger, CMO, Agentforce Service, Salesforce

Explore further:

Methodology:

Data in this report are from a double-anonymous survey conducted from March 9 to April 4, 2026. The survey generated 3,075 responses from service professionals across North America, Latin America, Asia-Pacific, and Europe. For more demographic information, please refer to the report.