Will AI Replace Help Desks? What the Research Says in 2025

Help desks are under pressure to deliver faster, smarter support. With rising customer expectations and more businesses using AI to gain an edge, service teams in 2025 are looking for ways to keep up.

Sixty-three per cent of service professionals say AI is already helping them respond to customers more quickly. As the technology improves, support teams are rethinking how they work, which tasks AI can handle, and where humans add the most value.

In this article, we’ll explore what’s changing. Using insights from more than 2,000 IT and development leaders in our latest State of IT: AI and App Development report, we’ll look at how AI is being used in help desks, where human input still matters, and how service teams can succeed in 2025 and beyond.

A help desk is a support team that people contact when they need help with a product, service, or system. It’s often the first point of contact when something goes wrong or when a user has a question. Its typical core functions include:

• Fixing technical issues
• Answering questions
• Logging and managing requests
• Escalating complex problems
• Keeping users informed

There are also different types of help desks. For example, IT support, customer service, internal employee help desks, and external support for customers or clients provide very different services.

AI agents are digital assistants that use data to assess a situation, decide what to do, and take action without needing human input. They use artificial intelligence to learn, adapt, and help people get things done without requiring manual work.

In customer service, they often used to answer questions, find solutions to simple issues, and support human agents behind the scenes. In fact, 83% of decision makers plan to increase investments in automation over the next year.

One example of an AI agent that can support your helpdesk is Agentforce. It’s an AI agent built to help your service teams respond faster and more accurately.

The future of customer service

While AI is making help desk jobs easier, it’s not a replacement for human support. It can handle simple tasks, but people are still needed for complex questions, sensitive issues, and moments that require empathy or good judgment.

Here’s a quick breakdown of the things AI can do in 2025 and what still requires a human touch.

As we’ve said, AI can automate repetitive jobs, recommend next steps, and guide users through common issues without needing a human. We found that desk workers using AI tools spend 26% less time on admin work, freeing them up for more strategic and meaningful work.

On top of this, behind the scenes, it helps agents by drafting replies and keeping knowledge articles up to date. As AI learns from customer interactions, it gets better at offering more personalised support.

There are still many areas where AI can’t replace humans, especially in scenarios that are complex or sensitive.

Some customers also expect to speak with a person, especially when things go wrong, so keeping that option matters for trust. We found that 82% of developers say the real advantage comes from humans and AI working together.

AI is already automating many of the simpler tasks that help desks manage. Jobs like password resets, system checks, and basic troubleshooting can often be done faster and more accurately with AI.

This shift may reduce the number of entry-level roles based on repetitive tasks, impacting those important foot-in-the-door opportunities in the workforce. At the same time, demand is growing for specialised skills in areas like IT support, data analysis, and AI operations .

For people with experience, roles are evolving. Many professionals are now focusing on complex problem-solving, managing AI systems, and improving the overall support experience.

According to our research, 72% of executives expect AI to have a positive or neutral impact on hiring. This suggests the future of support is not about job loss, but job change.

Now that we’ve covered what AI can do in theory, let’s look at how it’s being used in practice. Here are five real-world examples from Australia showing how teams are using AI to improve the customer experience.

We use our own Agentforce on our help site. Since its introduction, it now handles more than 85% of support queries without human intervention. It guides users to the right content and automates common issues, helping hundreds of thousands of people every week. This has given our support teams more time to focus on complex cases.

Team Medical is a family-owned and operated company that delivers medical supplies to 10,000 customers throughout Australia, with customers ranging from small GP clinics to large healthcare groups. They’ve built their success on being able to rapidly respond to customers. As the company has grown, so has the complexity of maintaining high service standards.

With approximately 25% of Team Medical’s 90,000 support cases handled annually being requests to place orders, using an agent to take customer orders would save the customer service team significant time while elevating the customer experience.

Formula 1 uses Agentforce to automate support and personalise service for millions of fans. The AI handles common issues like login and streaming errors, cutting response times by 80% and boosting first-call resolution to 95%.

TripADeal uses Agentforce to help customers navigate the website and get their questions answered quickly, enabling the travel agency's Travel Consultants to work more efficiently.

TripADeal’s first agent acts as a virtual travel consultant, conversing with customers in natural language to answer questions and understand their travel preferences so it can recommend the right deals. If customers have more complex questions or would like help to make a booking, the agent can seamlessly hand over to one of TripADeal’s travel experts. With Agentforce deeply integrated with Service Cloud, travel experts can see the full history of incoming conversations and easily pick up where the agent left off.

Luxury appliance brand Fisher & Paykel started using Agentforce to deliver faster and more connected support. By turning a 10,000+ article knowledge base into instant step-by-step help, they now expect online self-service to rise from 33% to 65%, all while reducing training time and manual effort.

Before you bring AI into your help desk, you need to get the basics right. AI is only as useful as the data behind it, the systems that support it, and the people trained to work with it. If those pieces aren’t in place, you could spend more time facing problems than solving them.

This is a real issue for a lot of businesses. Only 53% of developers say their data is high quality, and 69% say their infrastructure needs an upgrade before they can even roll out AI agents. So – it’s worth getting the basics sorted before you dive in.

Here’s a quick checklist to help you get started:

• Clean your data: Fix duplicates and remove outdated records.
• Connect your systems: Link your CRM and knowledge base so your AI can access the full picture.
• Built-in privacy and security: Use DevSecOps principles to keep customer data safe from the start.
• Design for human handover: Set rules for when a case should go to a real person.
• Train your team: Teach support staff how the AI works, when to step in, and how to improve it with feedback.
• Test before scaling: Run a pilot and fix any identified issues.
• Get customer feedback: Ask your customers to have their say about their experience using your AI. Use it to improve the experience going forward.

Getting this right early on makes implementation smoother and the tool more useful in the long run.