This guide is for managers and decision-makers who want the efficiency AI can offer while protecting the empathy customers expect.
What Customers Are Telling Us About AI in Customer Service
Gartner has surveyed customers on this question more than once, and the results point in more than one direction.
In a Gartner survey of 5,728 customers conducted in December 2023:
- 64% said they would prefer that companies didn’t use AI in their customer service.
- 53% said they would consider switching to a competitor if they found out a company was going to use AI for customer service.
- Their top concern was that AI would make it more difficult to reach a person.
A more recent Gartner survey of 3,566 B2B and B2C customers, conducted in February and March 2026, found:
- 50% of customers say their interactions are easier when companies use GenAI.
- 87% of customers say it is essential for companies to provide an option to reach a human agent when using GenAI.
The two surveys asked different questions more than two years apart, so they shouldn’t be read as a before-and-after comparison. They do share one consistent theme: customers want access to a person. Eric Keller, Sr. Director Analyst at Gartner, said, “Service leaders should not use GenAI as a mandatory first step for every issue.”
In a February 2026 release, Emily Potosky, Senior Director, Research in the Gartner Customer Service & Support practice, said, “AI simply isn’t mature enough to fully replace the expertise, empathy, and judgment that human agents provide.” The same release predicts that by 2027, 50% of companies that attributed headcount reduction to AI will rehire staff to perform similar functions, but under different job titles.
The practical takeaway for managers: treat AI as a tool that supports your people, not a replacement for them.
Where AI Tools Can Help Customer Service Teams
AI is most useful when it takes routine work off your agents so they can focus on the customer. Common use cases include:
- After-call work. Automatic call summaries and ticket notes can cut the time agents spend on documentation between interactions.
- Knowledge lookup. An assistant can pull up a policy or troubleshooting step while the agent stays in the conversation. The agent doesn’t have to put the caller on hold to search.
- Simple, low-stakes requests. Password resets, order status and store hours are typical candidates for self-service, as long as a person stays easy to reach.
- Coaching signals. Some tools flag calls where frustration seems to rise or where an agent interrupts the caller. Supervisors can use these flags to decide which calls to review.
In each case, the AI supports the human interaction rather than replacing it.
Where People Should Lead: Keeping the Human Touch
Some service moments call for human judgment and empathy.
- Upset or anxious customers. A customer who was double-charged, or who has already called several times about the same problem, usually needs to feel heard before the problem gets solved. Reading emotional cues and acknowledging feelings are skills agents build through practice.
- Complex or unclear problems. When an issue doesn’t fit a set category, an agent who can ask good questions and use judgment is better equipped than a scripted flow.
- Important relationships. In B2B settings, one poorly handled interaction with a key account can put the relationship at real risk.
- Service recovery. When something has gone wrong, a sincere response from a person gives you a chance to rebuild trust in a way a generic automated message usually can’t.
A useful rule of thumb: if the customer’s emotional state is likely to affect the outcome, have a person handle that moment.
How to Choose Human Touch Customer Service AI Tools
Before you commit to an AI product, ask these questions:
- Does it make it easy to reach a person? Gartner’s research shows customers care about this. Ask the vendor to show you the exact path from the bot to an agent.
- Does it pass along context? When a conversation moves from AI to an agent, check whether the agent can see what the customer already shared, so the customer doesn’t have to start over.
- Does it support your agents or script them? Suggested replies can help newer agents, but replies read word for word can sound robotic. Look for tools that offer guidance your agents can adapt.
- Can you measure the customer experience, not just speed? Handle time and deflection rates are easy to report. Make sure you can also track satisfaction and repeat contacts for AI-handled interactions, so you can see whether the efficiency gains hold up.
Judging human-touch customer service AI tools this way keeps the focus on the customer’s experience instead of the feature list.
The Skills Gap AI Can Create
When AI takes over the simpler, routine requests, more of the conversations that reach your agents may be the hard ones: frustrated, confused, or high-stakes customers. That makes soft skills more important. Agents need to:
- Read emotional cues quickly, on the phone or in writing
- Listen beyond the words to understand what the customer needs
- Stay calm when a customer is already frustrated, including after a bot experience that didn’t work
- Use empathetic language that sounds sincere, not scripted
- Recover well when a customer has been let down
If your budget has gone mostly to new software, check whether your team’s skills have kept up.
Our Emotional Intelligence at Work course, subtitled “Leveling Up Empathy,” includes modules on reading emotional cues, listening beyond words, acknowledging feelings before solving the problem, and staying calm under pressure. For support teams, The Human Side of Tech Support includes modules on de-escalating technical frustration and translating tech into plain language.
To make training last beyond the first rollout, our article on why customer service training programs fail covers common reasons programs underperform and what to do about them.
A Simple Framework for Managers
Here’s a four-step approach to rolling out human touch customer service AI tools:
- Map your interactions. List your most common contact types and sort them by complexity and emotional stakes.
- Start by automating low-stakes, low-emotion work. Order status and simple account questions are common starting points. Keep a clear, fast path to a person.
- Keep high-emotion moments with people. Complaints, cancellations, billing errors and service recovery are good candidates to leave with your agents. Give them the skills and authority to handle these well.
- Measure and adjust. Compare satisfaction scores and repeat contacts for AI-handled interactions against human-handled ones. If AI-handled interactions lead to more callbacks or complaints, change what you automate.
Our guide on how to measure customer service training effectiveness can help you build the reporting side.
Human Touch as a Differentiator
As AI tools become more common, how your team handles the conversations that matter most can be a big part of what sets you apart. Using AI to take routine work off your agents, then investing in their empathy and communication skills, puts each kind of work where it fits best.
Give Your Team the Skills AI Can’t Replace
ServiceSkills training is built on the Telephone Doctor approach to customer service. If your team is adopting AI tools and you want to strengthen the human side of service, explore our content library or book a 15-minute guided tour to see how ServiceSkills fits your training plan.