The State of Customer Support Automation in 2026

Two years ago, AI in customer service was mostly a demo. Today, service teams expect AI to handle around half of all customer service cases within the next year, according to Salesforce research in Australia and New Zealand. The shift happened faster than almost anyone predicted, and it is reshaping how support leaders think about staffing, tooling, and the customer experience itself.

Customer support automation has moved past the point where a company can call it innovation just because there is a bot on the website. In 2026, the question is no longer whether AI belongs in customer service. The question is whether the business has built the support model around it properly.

AI is moving from assistance to resolution

The strongest sign of change is the amount of work service teams now expect AI to absorb. Salesforce's Australia and New Zealand research puts the current figure at 31% of customer service cases, with expectations that AI will handle around half by 2027. In Australia, the projection is even higher, moving from 31% today to 60% by 2027.

That is not a small adjustment to the support stack. It changes how leaders think about demand, staffing, response times, and what human agents should spend their time doing.

Gartner's forecast adds to the same picture. It predicted that agentic AI could resolve 80% of common customer service requests, without needing any human intervention, by around 2029, with operational costs potentially reduced by 30%. For support teams dealing with growing volumes and tighter budgets, those numbers explain why AI is getting so much attention.

The shift is also changing what people mean when they talk about automation. Earlier support tools mostly helped with the work around the conversation. They could classify a ticket, suggest a help article, summarise what happened, or draft a reply for an agent to review. Agentic AI goes further than that. It can follow steps, complete parts of a process, and resolve some issues before a human agent needs to intervene.

That is where the opportunity becomes more serious, but so does the risk. Suggesting a response is one thing. Letting AI change an order, update an account, or guide a customer through a more sensitive issue is something else. The more responsibility AI takes on, the more important it becomes to define where it should stop.

The work is already shifting for agents. Salesforce reported that APAC service representatives using AI spend 20% less time on routine cases, giving them around four hours a week back for more complex work. The same research found that representatives using agentic AI spend 25% of their week on highly complex issues.

That feels like the more honest version of automation. AI may take away some repetitive work, but the work left behind is often harder. Complaints, exceptions, unclear policies, messy customer histories, and situations that need judgement do not disappear. They become a bigger part of the human role.

Adoption is growing, but maturity remains low

Intercom's 2026 Customer Service Transformation Report shows how quickly AI has entered the support function. Intercom reported that AI spending is already well underway, with 82% of senior leaders investing in customer service AI over the past year and 87% expecting to invest again in 2026.

The part that stands out more is the maturity gap. Only 10% of respondents said their AI deployment had reached a mature stage. So, while many teams now have AI in place, far fewer have reached the point where it is properly embedded into the way support works day to day.

That distinction is important. Turning on a chatbot is not the same as building a reliable automated support experience. The harder work sits behind the scenes: keeping the knowledge base accurate, connecting AI to the right systems, deciding what it can and cannot do, and making sure the handoff to a person is smooth when the issue needs one.

Intercom's findings suggest that teams with more developed AI deployments are more likely to see better performance after implementation. That makes sense. The value is not only in the software. It comes from the work around it.

This is also changing what support teams actually do. Intercoms report also shows that AI is creating new work for support teams, with 40% saying agents are spending more time training and refining these systems. That is an important detail because it pushes against the simple idea that automation just removes work. Someone still needs to review answers, update workflows, train and improve AI responses and workflows, and notice when the same problem keeps appearing.

For smaller teams, this might not become a dedicated AI role. But the responsibility still needs to sit somewhere. If nobody owns the quality of the automated experience, it will start to drift.

Governance and infrastructure are becoming critical

Sinch's 2026 research shows the more difficult side of this shift. It found that 62% of enterprises already have AI agents live in customer communications. At the same time, 74% have rolled back or shut down a deployed AI customer communications agent after a governance failure. Sinch also reported that 98% of enterprises are increasing investment in AI communications in 2026.

That combination says a lot. Companies are still investing, but they are also learning that a live customer-facing AI system is very different from a trial. Once AI is speaking to customers, the issues become more serious. Accuracy matters. So does security, compliance, context, escalation, and whether the system can be trusted to stay within the right limits.

Sinch's research also shows where much of the effort is going. Enterprises are putting heavy focus on trust, security, and compliance alongside AI development. Sinch also found that safety work is taking up a large share of AI engineering time, with 84% of teams spending at least half their time on safety infrastructure and 55% needing custom systems to keep context consistent across channels.

For support leaders, this is a useful reality check. AI will not fix a weak service operation by itself. If the help centre is out of date, CRM data is inconsistent, or escalation rules are unclear, automation will usually make those issues more visible rather than less.

The state of customer support automation in 2026 is moving in two directions at once. AI is taking on more frontline work, investment is increasing, and some teams are starting to see real value. At the same time, many organisations are still working out how to manage AI safely and consistently once it is live.

The companies that get this right will treat AI as part of the support model, not as a standalone fix. They will use it where it makes the customer experience faster and easier, while keeping human agents close to the moments that need judgement, accountability, and trust.

Sources

Salesforce. "AI Expected to Resolve Half of Service Cases in ANZ by 2027, Data Shows." Salesforce Newsroom, 18 November 2025. https://www.salesforce.com/au/news/stories/state-of-service-report-announcement-2025/

Gartner. "Gartner Predicts Agentic AI Will Autonomously Resolve 80% of Common Customer Service Issues Without Human Intervention by 2029." Gartner Newsroom, 5 March 2025. https://www.gartner.com/en/newsroom/press-releases/2025-03-05-gartner-predicts-agentic-ai-will-autonomously-resolve-80-percent-of-common-customer-service-issues-without-human-intervention-by-20290

Intercom. "2026 Customer Service Transformation Report." Intercom. https://www.intercom.com/customer-transformation-report

Sinch. "Sinch research reveals 74% of enterprises have rolled back live AI customer communications agents." Sinch Newsroom, 13 May 2026. https://www.group.sinch.com/media/press-releases-and-news/2026/sinch-research-reveals-74-of-enterprises-have-rolled-back-live-ai-customer-communications-agents/