For years, businesses have been using AI agents for customer service to make support faster and more efficient. Now, customers are bringing their own AI into the conversation.
They’re using AI to research company policies, prepare questions, challenge answers and even handle parts of customer service interactions on their behalf, changing the dynamic between businesses and the people they support.
In a new survey of 1,000 U.S. consumers, Capacity found that 35% have already used their own AI agent or agentic AI in general to handle a customer service request at least once. Among Gen Z and Millennials, that number jumps to 49%.
Our new report, the Agentic Consumer Report, explores how personal AI tools are changing customer behavior and what businesses need to do to prepare.
Customers are bringing their own AI into the conversation
More than half of consumers surveyed (55%) have already used an AI tool to assist with a customer service request.
For some, that means asking ChatGPT to explain a return policy before contacting a retailer. For others, it might mean using AI to research a billing dispute, prepare questions or determine whether a company representative is giving them the right information.
And consumers aren’t necessarily accepting the first answer they receive.
Our research found:
- 44% have encountered information from AI that contradicted a company representative or policy.
- 34% say AI makes them ask more specific questions.
- 27% feel more confident asking for what they want.
- 23% say AI leads them to ask representatives to explain or cite company policies.
The challenge for businesses is that they can control the information provided by their own AI agents and human representatives, but they can’t control what a customer’s AI tool tells them.
That means customer service teams increasingly need to be prepared to explain not just what the answer is but why it’s correct.
AI is giving consumers more confidence to push back
AI is changing consumers’ willingness to challenge customer service decisions.
When asked whether using AI would make them more likely to push back after a company denied their request, 58% said they would be somewhat or much more likely to do so.
And when information from AI conflicts with a company representative, consumers are just as likely to trust AI outright as they are to trust the representative outright, at 14% each.
This doesn’t necessarily mean AI is providing the correct information. AI tools can misunderstand policies, rely on outdated sources or generate incorrect answers.
But consumers may enter conversations believing they already know what the company should do.
For contact centers, this growing consumer confidence creates a new challenge. Representatives need to be ready for more informed customers, more specific questions and potentially more complex conversations.
Consumers want AI to help, but not always take over
While personal AI agents are gaining traction, our research suggests consumers aren’t ready to hand over every customer service interaction.
Only 14% say they would trust an AI agent to manage an entire interaction from start to finish. Another 21% would trust an agent to reach a human representative before taking over the conversation themselves, and they expect the agent to handle the AI-to-human handoff smoothly. And 26% wouldn’t let an AI agent handle any part of a customer service interaction on their behalf.
Consumer AI isn’t necessarily replacing the customer in every interaction. It’s helping them navigate the process, whether that’s researching an issue, getting through a queue or preparing to speak with a representative.
Businesses need to be prepared for both possibilities: AI agents acting on behalf of customers and customers arriving with AI-generated information.
What does this mean for contact center leaders?
As consumers become more comfortable using their own AI tools, businesses will need to adapt how they support customers with contact center AI agents.
Start by understanding what customers are challenging.
When consumers question policies or present conflicting information, those interactions can reveal where communication is unclear, knowledge needs updating or agents need additional guidance.
Conversation Intelligence can help businesses analyze interactions to identify recurring questions, sources of confusion and opportunities to improve support. Rather than treating each challenged answer as an isolated incident, teams can use those insights to improve future interactions.
Make sure AI agents for customer service and human agents have access to the same knowledge.
A customer shouldn’t receive different answers depending on whether they interact with a chatbot, voice agent or human representative.
Connecting company knowledge across channels helps businesses maintain consistent answers and makes updates easier to manage. When a policy changes, the information supporting both AI and human agents should change with it.
Give human representatives the information they need in real time.
When a customer arrives with AI-generated information that contradicts a company policy, representatives need to be able to respond quickly and accurately.
Real-Time Agent Assist can surface relevant knowledge and guidance during conversations, helping representatives explain policies, address discrepancies and resolve issues without searching through multiple systems. By reviewing and improving interactions more comprehensively with agent assist tools, contact centers can achieve real ROI.
Just as importantly, businesses need to learn from these interactions. Identifying where customers challenge answers or bring conflicting information into conversations can help teams continually improve their knowledge and better prepare both human and AI agents for future interactions.
Preparing for the next generation of customer service
The rise of personal AI doesn’t mean every customer service interaction will become AI-to-AI. Our research shows consumers still want control, particularly when issues become complex or require a human decision.
But the expectations are changing.
Customers have more ways to research, question and challenge the information businesses provide. That raises the importance of accurate knowledge, well-prepared agents and the ability to identify where customer experiences need improvement.
At Capacity, we believe the next step isn’t simply deploying more AI agents for customer service. It’s connecting knowledge, AI agents, human teams and conversational insights so businesses can learn from every interaction, improve their responses and deliver answers customers can trust.
Want to see how Capacity helps contact centers prepare for changing customer expectations? Book a demo.
with Capacity agentic AI.
FAQs
According to our research, 55% of consumers have used an AI tool to assist with a customer service request. Usage of personal AI agents is highest among younger customers: 69% of Gen Z and Millennials have used their own AI agent to handle a request, compared with 35% of all respondents.
Consumer AI gives customers more confidence to challenge support decisions. 58% say using AI would make them somewhat or much more likely to push back after a company denied their request. Customers also use it to sharpen their approach: 35% say AI makes them ask more specific questions, 27% feel more confident asking for what they want and 23% ask representatives to explain or cite company policies.
Not entirely. Only 14% of consumers would trust an AI agent to manage an entire interaction from start to finish, and 26% wouldn’t let an AI agent handle any part of an interaction on their behalf. Another 21% would trust an agent to reach a human representative before taking over the conversation themselves. Personal AI is mostly helping customers navigate the process, so businesses need to prepare for both AI agents acting for customers and customers arriving with AI-generated information.
Start with three steps. First, use Conversation Intelligence tools to find which policies and answers customers challenge most often. Second, connect company knowledge with AI Orchestration so your AI agents for customer service and human agents give the same answer, every time, on every channel. Third, guide representatives with Real-Time Agent Assist so they can explain policies and address discrepancies quickly, without searching and building up handle time.
