State of Service: AI Agents Edition

For this special edition of the State of Service report, Salesforce surveyed 3,075 service professionals to learn:

• How to adopt and apply AI agents across service organizations, from customer-facing support to internal operations
• How to deploy AI agents across channels and measure their impact
• How to build trust, transparency, and security into AI deployments

Data in this report is from a double-anonymous survey conducted from March 9 through April 4, 2026. Respondents represent 13 countries across five continents. All respondents are third-party panelists. Click here for further sample details.

AI agents are no longer just experiments. They’re actively embedded in workflows, handling routine tasks, surfacing insights, and freeing up human service reps to focus on higher-value work. Organizations are learning what works, where AI delivers the most impact, and how to build trust while scaling responsibly.

AI adoption is now widespread, and the move toward agentic capabilities is accelerating. Eighty-five percent of service organizations report using at least one form of AI, with agentic AI adoption up from 39% to 66% in just one year.

As organizations deploy AI agents across service channels, they’re beginning to measure where the technology is making the biggest difference. Service professionals report that the top KPI improved by AI agents is customer satisfaction, highlighting the growing role AI plays in improving the customer experience.

Trust remains a top consideration for scaling AI. As AI becomes more embedded in service experiences, organizations are putting guardrails in place to ensure customers remain informed and in control. Seventy-seven percent of companies with customer-facing AI allow customers to request a human representative at any point during an AI interaction, helping maintain trust as AI adoption expands.

Today, AI is a standard part of the service toolkit. Eighty-five percent of service professionals say their organization uses AI.

But the most notable shift is happening within that broader adoption. Organizations are increasingly turning to agentic AI — autonomous systems designed to take action and support more complex service workflows. Sixty-six percent of service organizations report using agentic AI today, up from 39% in our April/May 2025 survey — a big leap in just one year.

While earlier waves of automation focused primarily on rules-based tasks, agentic AI expands those capabilities by helping service teams manage more dynamic interactions and operational processes. With that, AI is becoming less of a standalone tool and more of an active participant across every service interaction.

Organizations adopting AI agents are putting them to work across a wide range of service activities. Seventy-seven percent of companies with AI agents deploy them in both customer-facing and internal service operations.

AI agents can support the entire service lifecycle. The most common use cases include proactive outreach, personalized product recommendations based on customer preferences and history, and resolving cases across multiple channels.

For customers, AI agents can assist via chat or voice — offering recommendations and issue resolution. They can also get ahead of problems through proactive outreach – like catching a billing error and notifying the customer before they run into it.

For service teams, agents can triage incoming cases and route them to the right service reps, surface information (like customer data or knowledge articles) while service reps are handling an issue, and analyze conversations to provide insights and coaching tips.

Together, these capabilities show how AI agents are deeply embedding themselves in how service gets done.

Service teams are already transforming their organizations to support AI. For one, organizations are creating entirely new positions to support the technology, like architect roles to oversee AI deployment and operations, and data management roles to maintain knowledge bases for AI.

And across organizations, service professionals are investing in new skills that help them work effectively as customer service transforms. In fact, the vast majority of service representatives report participating in some form of upskilling, like attending conferences, taking online courses, and earning certifications. Only 3% of service reps report no engagement with upskilling, underscoring just how broadly this shift is taking hold across the workforce.

Service professionals agree that the growth of AI is transforming the skills service reps need to succeed. Top priorities include AI oversight and judgment — knowing when to trust or override AI outputs — along with adaptability and learning agility, and complex problem solving for issues that require human intervention. Together, these reflect a shift in the role of the service representative, from handling routine tasks to managing exceptions, guiding AI-enabled experiences, and building stronger customer relationships that promote long-term satisfaction and loyalty.

As organizations adopt AI for responsibilities like case resolution and knowledge retrieval, service professionals are learning to collaborate with these tools in their daily work. The result? Organizations where humans and AI work together to deliver exceptional service.

AI isn’t just supporting customer interactions — it’s reshaping how service leaders manage their teams. For instance, about half of service leaders say they rely on AI for analyzing trends and monitoring the performance of their teams and of individual service reps.

By surfacing insights on trends, capacity, and performance, AI helps leaders make more informed staffing and coaching decisions. In fact, 92% of service leaders with AI say it improves their ability to coach at scale. AI can also help anticipate customer demand, balance workloads, and optimize resource allocation in real time – which can help orgs control labor costs and make the most of return on investments.

AI agents are no longer confined to a single touchpoint – they’re showing up across the service landscape.

Eighty-three percent of service organizations with AI agents say they’ve deployed them across five or more channels. Top channels are email, online chat, messenger apps, SMS, and phone — common environments where customers seek support. Deploying AI agents across multiple channels allows organizations to support customers wherever interactions begin, while maintaining more consistent experiences across touchpoints.

For service teams, this shift also changes how work flows through the organization. Instead of handling requests that originate from one channel at a time, AI systems can assist with interactions across multiple environments simultaneously.

Also important is what happens when AI hands work to a human. In many service organizations, AI agents surface context, summarize conversations, and pass interactions to service reps when human expertise is needed. This allows service reps to pick up where AI left off without asking customers to repeat information, creating a more seamless experience across channels and between AI and human support.

Organizations adopting AI agents are increasingly able to quantify their impact. Forty percent of the time AI is used in case resolution, that work is done autonomously, showing strong organizational confidence in AI’s capability to act independently. In our previous 7th State of Service report, service leaders and operations professionals using AI agents said they expected service costs and case resolution times to decrease by an average of 20% — suggesting many organizations are already beginning to see the operational shifts they anticipated.

Teams are also seeing results quickly. Seventy percent of organizations with AI agents say they observe measurable value within 60 days of deployment.

As a result, service organizations are developing new ways to track performance. Sixty-six percent say they now measure AI accuracy, and 66% track AI resolution rates to evaluate how effectively AI systems address customer needs.

Service professionals are also noticing workflow improvements. Eighty-eight percent of service reps using AI say it reduces time switching between tools and systems.

Confidence in these technologies is high. Eighty-nine percent of service professionals with AI agents say their company would benefit from expanding their use. Together, these metrics suggest AI agents are helping service teams improve both quality and scale — resolving more requests autonomously while maintaining high levels of accuracy.

AI agents are often introduced to improve operational efficiency. But service professionals say the most meaningful improvements are showing up somewhere else: the customer experience.

When asked which KPIs have improved the most with AI agents, customer satisfaction rises to the top. That’s notable because the metric isn’t strictly about operational speed or internal productivity. Instead, it reflects how customers actually feel about their service experience.

Operational metrics also see impact. Service professionals report improvements in service rep productivity, average handle time, and first-response time, suggesting that AI is helping teams work faster and more efficiently. These gains can translate into meaningful cost savings by reducing the time and resources required to resolve customer issues. But the fact that customer outcomes lead the list suggests those operational gains are translating into better experiences for customers.

By helping service teams respond faster, access information more easily, and resolve issues more consistently, AI agents are improving the quality of interactions customers have with service organizations. And when service gets better, satisfaction and loyalty tend to follow.

Despite growing momentum, implementing AI agents at scale can be complex. Organizations report that the biggest obstacles aren’t the AI technologies themselves, but the systems and data that support them.

Organizations cite challenges like security and compliance concerns, legacy technology environments, and fragmented data across systems. As companies work to improve data quality and integration, they’re laying the foundation needed to scale AI agents with confidence.

More than half of service leaders (59%) say data readiness is a major blocker to AI adoption. Concerns are even higher among the teams: 63% of service reps and 72% of service operations professionals say data readiness is a challenge. This gap may reflect who experiences the issue most directly — reps and operations professionals are the ones working with the data day to day — suggesting some leaders may be underestimating the scale of the problem.

AI systems rely on accurate, accessible, and up-to-date information to deliver reliable results. When data is incomplete or difficult to access across systems, AI outputs can become less reliable.

All AI agents are powered by data — and sometimes that data can be sensitive, such as payment information, health records, and personal identifiers. Most AI agents for customer service have access to at least some sensitive data, though only a minority have full access. In a world of increasingly sophisticated cyberattacks, this brings data privacy concerns increasingly to the fore, with 70% of service professionals saying data privacy concerns delay or limit their AI rollouts.

According to a separate Salesforce study , 88% of data and analytics leaders believe AI advances demand new data governance approaches — but those approaches are still evolving and adoption of them is uneven. For instance, while 69% of organizations conduct real-time data monitoring, under 50% engage in practices like encryption, advanced access controls, and defined incident response plans. In fact, only 43% of data and analytics leaders have even established formal data governance frameworks and policies.

As AI agents become more integrated into service operations, and across the company, the importance of data governance heightens.

Most service professionals say customers are becoming more comfortable with AI-powered service. In fact, 65% say their customers fully trust AI, and just under half say their organization is cautious about using AI in customer interactions.

However, research suggests customers may be more hesitant than service teams believe. According to Metrigy,* consumer trust in AI-powered service remains relatively low compared with how organizations perceive it. Less than 44% of consumers trust AI to handle their customer service needs.

Despite low trust, Metrigy also found AI-powered customer service experiences outpacing consumer expectations. For instance, 73% of healthcare consumers, 69% of financial services consumers, and 64% of retail consumers said AI-powered customer service exceeded their expectations. These results suggest that while skepticism remains, positive experiences with AI may build trust over time.

Across the service organization, trust in AI is high. Most service professionals say they personally trust AI agents to handle tasks autonomously — including sensitive ones — and that confidence holds across roles, from frontline service reps to service leaders.

Not surprisingly, trust correlates with experience. Service professionals at organizations with AI agents are more likely to trust autonomous AI than those at organizations without them. In other words, seeing AI agents in action appears to build confidence in what they can do.

As organizations gain more real-world experience with AI agents, familiarity may strengthen trust. What starts as cautious experimentation can evolve into broader confidence as teams see AI consistently deliver accurate responses, automate routine work, and support faster resolutions.

As AI becomes more deeply enmeshed in the customer experience, organizations are focusing on transparency and choice. Among companies that use customer-facing AI, 77% say they allow customers to request a human representative at any point during an AI interaction with a voice assistant or chatbot.

Over half say they’re taking steps to communicate clearly when AI is involved, such as disclosing to customers when they are interacting with AI or allowing customers to opt out of AI interactions entirely. These practices help ensure that customers maintain control over how their support interactions unfold.

For service organizations, transparency measures can also help address concerns about AI adoption — like reliability, data privacy, and the ability to escalate to a human when needed — by making the role of AI within the service experience more visible. As AI continues to evolve, organizations are exploring how these practices influence both customer confidence and service outcomes.