Zendesk Voices: front line teams in the age of AI

The needs of front line teams have definitely changed in the age of AI, but not always in the way people expected. It used to be mostly about giving agents good training, solid knowledge, and the right tools to get through the work. Those things still matter, but now agents are increasingly handling the contacts AI could not resolve. That usually means the work is more complex, more emotional, or requires more judgment.

Because of that, the tools that matter most are the ones that give agents context. Good summaries, accurate suggestions, clear escalation paths, strong knowledge, and routing that gets the right work to the right person all make a big difference.

One thing we have seen is that bad AI can be worse than no AI. If the suggestions are inaccurate or irrelevant, agents quickly learn to ignore them. Giving agents a way to flag those misses, and showing that their feedback actually improves the system, is important.

The human side matters too. When teams hear that AI may deflect a large portion of contacts, it is natural for them to wonder what that means for their role. The best approach is to bring them into the transition. Let them test, share feedback, identify edge cases, and help shape how the tools are used.

AI can remove a lot of repetitive work, which is a real win. But it also raises the bar for the work that remains. Agents need more context, more confidence, and more authority to handle the moments automation cannot.

This is a great response ​@Brandon Tidd ! 

It is so important to note that AI raises the game for frontline teams and really highlights the expertise they have in solving more meaningful interactions.

Having the right information, easily available can really help empower frontline teams to solve these interactions in a frictionless way.

Zendesk following all cutting edge technologies and AI 

Zendesk maintain all process and getting control

One of the biggest changes I've noticed is that agents spend less time searching for information and more time solving customer problems. AI can handle repetitive tasks and suggest helpful responses, but agents still play the most important role when a conversation needs empathy or critical thinking.

For me, keeping the Help Center up to date has become even more important. Better knowledge means better AI suggestions, and that gives agents more confidence while still allowing them to make the final decision.

I am very satisfied 

Our definitions, priorities, and focus have changed significantly over time. However, the outcome we strive for has always remained the same: delivering satisfaction and the best possible experience for our customers.

What was once centered primarily on speed and efficiency has evolved into something much greater: creating opportunities for partnership, understanding, and connection. It goes beyond simply solving problems. It means showing empathy, being present, and genuinely listening to our customers' challenges. At times, it can feel almost like being a psychologist, providing a listening ear and support when it matters most.

There is a quote that resonates strongly with me: "People may forget what you did for them, but they will never forget how you made them feel." This perfectly summarizes what has been capturing our attention, especially in today's era of Artificial Intelligence.

One point I would like to highlight is the transformation we are driving. Through the evolution of our chat experience, we have gained a much deeper level of empathy and understanding regarding the impact of every interaction. It is no longer just a decision tree or a set of predefined instructions. Today, it provides a more personalized, human-centered, and empathetic experience, creating stronger engagement and more meaningful connections with our customers.

For this reason, we are extremely satisfied with the progress we have made and the results we continue to achieve.

Roger Souza.

How our frontline team's needs have changed

One of the biggest changes we've seen in the age of AI is how frontline teams acquire information.

In the past, when a customer reported an error message or shared logs, we had to search through internal documentation, reach out to engineers or developers, and wait for responses before we could provide meaningful guidance. Sometimes the right person was unavailable or focused on higher-priority tasks, which delayed the entire process and left customers waiting.

Today, AI has significantly reduced that dependency. While AI does not always provide perfect answers, it can often deliver 70-80% of the information we need in real time. That is usually enough to point us in the right direction and accelerate problem-solving. As a result, customers receive feedback much faster, and frontline teams spend far less time searching for information.

What feels easier in the age of AI

One area that has become noticeably easier is handling repetitive explanations and information sharing.

Customers often become frustrated by slow responses, inaccurate answers, or interactions that feel inattentive. AI helps address all three. It can provide detailed explanations of complex technical topics consistently and without fatigue. Whether it's the first inquiry of the day or the hundredth, the quality and depth of the explanation remain the same.

Frontline agents frequently encounter customers who ask the same questions multiple times or require step-by-step guidance throughout the troubleshooting process. While that repetition can sometimes be mentally exhausting for people, AI can patiently provide the same level of detail every time. This creates a more consistent experience for customers and reduces the burden on support teams.

How AI helps reduce emotional labor

In my experience, one of the most valuable benefits of AI is the reduction of emotional labor.

When customers are frustrated, frontline teams often absorb that frustration. Long resolution times, repeated explanations, and communication delays can create stress for everyone involved. AI helps remove many of those friction points by providing faster responses and more consistent information.

As a result, agents spend less energy managing repetitive interactions and more energy focusing on meaningful customer engagement. Customers receive quicker answers, while agents can maintain a more positive and patient mindset throughout the day.

Ultimately, AI helps create a healthier service environment. When customers are less frustrated, frontline teams are better positioned to provide friendly, empathetic support, leading to a better experience on both sides.

Frontline teams can focus on solving problems

AI is not just a tool for generating answers. It allows frontline teams to focus on what matters most: solving customer problems.

A significant portion of support work used to involve searching for information, coordinating with other teams, and manually drafting responses. AI reduces much of that effort. It helps gather relevant information more quickly, shortens communication cycles, and decreases the time required to compose responses.

This allows agents to spend more time analyzing root causes and identifying solutions rather than performing administrative or repetitive tasks. Customers receive faster resolutions, and frontline teams can dedicate their expertise to higher-value work. In many ways, it creates a positive cycle that benefits both customers and support organizations.

One more thing: Who will QA the AI?

While AI creates tremendous value, it also introduces a new challenge.

Many organizations expect AI to reduce operational costs, but maintaining AI quality requires ongoing investment. Someone must monitor responses, identify inaccuracies, validate outputs, and continuously improve performance. In some cases, the resources required to manage AI can offset a portion of the efficiency gains it creates.

This is why I believe the next evolution will be QA AI, an AI designed specifically to evaluate other AI systems.

A QA AI would learn from the standards and evaluation criteria defined by frontline team leaders and support managers. It could review AI-generated responses for accuracy, policy compliance, service quality, and customer experience. Instead of relying entirely on human reviewers, organizations could use QA AI to perform continuous monitoring at scale.

Looking further ahead, I believe QA AI should also be capable of learning from feedback and improving over time. In the future, we may see AI supporting customers, AI assisting agents, and AI monitoring AI. That layered approach could help organizations maintain both quality and efficiency as AI adoption continues to grow.

From my perspective, AI is not replacing frontline teams. It is empowering them. It helps people deliver faster, more accurate, and more human-centered support. The real competitive advantage will not come from simply deploying AI, but from effectively combining people, AI, and AI governance into a unified support ecosystem.

Zendesk Voices: front line teams in the age of AI

One AI feature that has proven particularly valuable for our team is the Ticket Summary. By providing an instantaneous distilled version of a ticket, it allows me to get up to speed on a conversation that may have had dozens of public replies, internal notes, and transfers, and quickly ascertain what the problem is, what has been done so for, the status quo, and what if anything needs to be done next. I find that this drastically improves my ability to respond to tickets that have been routed to me, particularly if they have been transferred from another agent .

This in turn improves my ability to provide a first response to a customer after the initial request, and provides greater continuity if I need to pick up where a colleague left off. I think it is especially important in a helpdesk environment that serves students, where a single request may require several interactions with different departments before a satisfactory resolution is reached, and AI-generated summaries provide greater context for less effort compared to reading through a ticket from scratch.

I think this particular feature is valuable for other contexts too, although it would probably be less common that one agent would have to handle a single ticket through multiple stages, and a new agent would need to get context on a single case before passing it on to someone else.