Human-centric support is a product decision at Deskpro

“We’re still leaning on the human side of things, because there’s a subtlety there [that’s important.] Our agents need to know how to solve a problem, how to dig deeper and read between the lines.”

That’s what one of our customers recently told us about the way they run their support team, and it’s reflective of a trend we’ve been seeing across our customer base.

The support teams we work with don’t measure success by tickets avoided. They measure it by cases handled well. That sounds obvious, but it’s the opposite of how most of the help desk market is currently selling itself.

There are two ways to use technology to improve support: deflect tickets before they reach a person, through chatbots and help centers, or make the people handling tickets faster and better at the hard ones. Deskpro can do both, and we think about them differently. The chatbot and help center are the front door. The product decision we’ve made is that the front door should hand off cleanly to a person when it needs to, rather than pretend it doesn’t need to.

There will always be complex cases that require human judgment, empathy, and problem solving. And there will always be issues, especially in regulated industries like financial services and healthcare, where the risk of automating support from end-to-end and having something go wrong is simply too high.

Deskpro works with a wide range of customers in B2B customer support, HR, and IT, and I’ve seen firsthand the importance of keeping people at the center of support. It’s why we’ve chosen to double down on the harder half of support: the issues that can’t be deflected or fully automated. One of the clear themes coming out of our customer interviews is that the support leaders we work with don’t see fully autonomous support as the end goal. While they see the value of AI as a productivity tool, they don’t want to deflect every ticket because they know their people bring empathy, expertise, and problem-solving skills that AI can’t reproduce. We’re deliberately prioritizing tools that make human agents more efficient on our roadmap because we know that’s what our customers actually want.

Why human-centric support matters more than ever

91% of service and support leaders report pressure from executive leadership to implement AI, according to a 2026 Gartner survey. On one hand, I understand where the mandate is coming from. Support is one of the most clear-cut use cases for getting a return on your AI investment. The more tickets you deflect, the less money you spend on support interactions, at least in theory. However, this race to deflect as much as possible overlooks the fact that there will always be situations when people want support from a human or that require a human in the loop to manage risk.

Consumer research points the same way: in a 2025-26 Metrigy study, almost 85% of consumers said they would prefer getting support from a human over an AI agent, and 80% said that preference held even if they were assured AI could resolve their issue.

That’s consumer sentiment, and consumer attitudes towards AI are shifting. But the teams we work with aren’t running consumer helplines. They’re resolving complex B2B, HR, and IT issues where the cost of getting it wrong is higher and the preference for a person is even stronger.

One of the big risks is implementing AI for self-service without providing a clear path to a human: only about a quarter of customers say they trust AI in support when there’s no option to escalate to a human, while over half trust it if there is a path to a live agent. Customers want the assurance that when they chat with your business or submit a support ticket, there’s a person on the other end who can empathize and problem solve, especially when their issue is complex or urgent.

Then there are the cases that aren’t just a matter of customer preferences, but of business or legal requirements. I see this frequently with the B2B customer support teams Deskpro works with. These support organizations may have contracts specifying their clients will have access to human agents or consultants, with strict SLAs for first response and time to resolution. Beyond contractual obligations, some decisions simply need a person to own them. An HR team can’t automate the outcome of a sensitive employee case. An IT team granting access to financial systems needs someone accountable for the approval. A B2B support team that misses an SLA for a key account needs someone with the authority to decide whether to offer a service credit, and to stand behind that call when the contract comes up for renewal. In each case, the risk isn’t that AI gets the answer wrong. It’s that nobody owned the decision.

As the agent role evolves, so must agent productivity tools

The need for humans in support isn’t going away, but the role is shifting. As AI tools and help center content aim to deflect the straightforward, human agents spend more time working on complex issues and edge cases. While this work can potentially feel more rewarding than resolving the same routine issues day in and day out, it can also be taxing for agents who don’t have the right tools.

This year we’ve interviewed help desk admins and agents, and the most consistent theme from those conversations is overload. Not ticket volume alone, but the combination of multiple open tickets, different SLA clocks, and the information needed to resolve any of them scattered across several systems or sitting with a specialist. That combination slows resolution and causes errors.

The answer isn’t an AI agent that tries to automate the work away. It’s tooling that takes the gathering and juggling off the human agent so they can spend their attention on the problem.

Deskpro prioritizes the human experience in our product

User research and customer feedback shape the Deskpro roadmap, and what we’ve heard has solidified two focal areas.

The first is configurability. Every support team operates differently, and we’d rather the platform bend to the team than the other way round. Triggers, escalations, follow-up reminders and if/then rules let teams automate the repetitive steps around a ticket while keeping a person on the decision. That’s not unique to Deskpro, but how we apply it is: the automation is there to protect the human’s time rather than remove them from the loop.

The second is how we’re thinking about AI. Our belief is that AI is more effective when it sits beside the support agent, not just in front of the customer: drafting replies the agent reviews, surfacing context from past tickets, recommending what to tackle next. The constraint is deliberate. AI does the gathering and synthesising, and never makes a decision a person should own.

The goal is the same in both cases: free agents from the work that isn’t a good use of their time and judgment, so they can spend it on the work that is.

The future of our help desk is human

One thing that’s become clear from my team’s conversations with our customers is that they don’t want to replace their people with AI. If they’re going to use AI, they want it to be genuinely helpful for their agents.

The support organizations we work with rely on their people to resolve challenging issues and build trust with their customers or employees. B2B support teams need people to resolve high-stakes issues for accounts with annual recurring revenue attached to them. HR teams need people to handle sensitive employee cases. IT teams need people to gather information and find creative ways to solve new problems.

All these teams need tools that prevent cognitive overload, connect agents to information they need in the moment, and keep humans in the loop to maintain trust and reduce risk. People at the center of support operations are not a limitation we’re working around. It’s the thing we’ve chosen to build for and it’s why our product roadmap is focused on the humans behind the help desk.