Customer self-service: why it resolves so few requests

Customers use self-service tools, from help centers to chatbots, more than ever. Most still leave without a solution and end up calling an agent. Here is what the data shows about customer self-service, and what to measure to fix it.

Key takeaways

• Gartner finds that 73% of customers use self-service, yet only 14% of issues are fully resolved there. The most common cause of failure: customers can’t find content that matches their issue.
• Help center traffic doesn't tell you whether self-service works. Track the share of requests resolved without an agent.

What is customer self-service?

Customer self-service covers every tool that lets customers find an answer or complete a task on their own, without contacting an agent. Its job isn't just to prevent a call. It's to resolve the request.

Service leaders are betting heavily on it. In a 2025 Gartner survey of 265 customer service leaders, self-service and live chat are expected to overtake phone and email as the top customer service technologies by 2027.

Types of customer self-service

Self-service comes in three broad forms, and each one answers a different kind of need.

Help centers, FAQs and knowledge bases

Open-access content answers the most common questions, such as shipping times or return policies. It's usually the first place customers look.

Chatbots and community forums

Conversational tools guide customers toward the right answer. A chatbot narrows down the question, while a community forum lets customers learn from each other.

Customer accounts and authenticated portals

Once logged in, customers see their own file and can complete tasks online, like tracking a request or updating a contract. An enterprise web portal connected to the CRM and back-office tools can show each customer their contracts and the status of their requests. We come back to this below.

How many issues does customer self-service actually resolve?

According to a Gartner survey published in August 2024, only 14% of customer service issues are fully resolved in self-service. The same survey shows that 73% of customers use self-service at some point in their journey.

The two figures don't share a base, since the first counts issues and the second counts customers. Gartner surveyed 5,728 customers in December 2023.

Complexity doesn't explain it all. Even for issues customers describe as “very simple,” only 36% are fully resolved in self-service.

More recent data points the same way. In a second survey of 5,801 customers run in early 2025, Gartner found that only 22% of customers resolve their issue end to end within a company's own channels. More than half of service journeys (51%) now begin elsewhere, on a search engine or on ChatGPT.

Why customers can't find the answer

Most of the time, customers simply can't find content that matches their issue. It's the most common reason cited in the Gartner survey, behind 43% of self-service failures.

Missing content isn't the only problem. Nearly half of customers who started in self-service (45%) said the company didn't understand what they were trying to do.

This often happens when the question is about the customer's own situation, such as where their claim stands. A general article won't help: they need an answer about their own case, often a task they can finish online.

Agents play a part too. A 2025 Gartner survey found that 60% of customer service agents don't promote self-service options. Yet customers whose agent does mention them are about twice as likely to try self-service next time.

When customers can't find what they need, they end up calling an agent. Or they look elsewhere and land on answers the company no longer controls. Either way, the contact wasn't avoided. It was only delayed.

Which metric shows whether your self-service works?

The right metric is the share of requests resolved without an agent. It's also what Gartner measures in its survey, rather than traffic.

Many teams mainly track help center traffic or the deflection rate, the share of visitors who didn't contact an agent. Those numbers show that self-service is being used. They say nothing about the outcome, since a customer who gives up without calling counts as a success.

Two sources help you estimate the resolution rate.

1. What your tools show. A completed task in the customer account, such as a request submitted or a contract updated, is a good sign. A customer who calls or writes again about the same topic within a few days, on the other hand, wasn't served.
2. Customer feedback, collected with a short question at the end of the journey (“Is your request resolved?”).

Put this rate at the top of the dashboard and the conversation changes. The question is no longer whether customers find the help center, but whether it actually helps them.

How to improve self-service resolution

Gartner pairs its survey with four recommendations. Taken in order, they make a simple method.

1. Have the people who handle requests write the content. Agents know which questions keep coming back and how they get solved. Capturing that knowledge while they resolve a case feeds self-service directly.
2. Anticipate needs with customer data. Account details and past interactions make it possible to offer the right answer before the customer even looks for it.
3. Offer a single front door. Gartner gives the example of a prominent GenAI-powered assistant that points customers to the right resource.
4. Assess content continuously. Customers and agents should be able to flag a page that doesn't help, so it gets fixed quickly.

One more rule is common sense. When self-service isn't enough, customers should be able to reach an agent without explaining everything again. Otherwise a self-service failure just adds to the wait on the phone.

What an authenticated customer portal changes

A public help center shows everyone the same thing. An authenticated customer portal knows who is logged in, and that's what lets it put these recommendations into practice.

It starts with customer data. User roles and groups determine what each customer sees, from their contracts to the tasks that apply to them. The data comes from the CRM and back-office tools, which Jahia connects through stackConnect, 1000+ connectors, so customers can submit a request and then track it without switching channels. Content also stays current, because business teams update pages themselves and fix the ones customers flag.

The gap with a public help center shows most in industries where customers ask about a case in progress. In insurance, a policyholder who has just filed a claim wants to know which documents are still missing before they get paid. Only a customer account connected to the claims system can tell them.

Housing providers are in the same situation. Tenant requests come in by phone or email, and nobody really knows where they stand. For tenants to track a request online, the portal has to be connected to the housing provider's CRM, the tool where staff handle these requests. Jahia presented this case with Efficy, a partner CRM vendor, in a webinar on the tenant portal connected to a CRM (in French). The page on the Jahia and Efficy integration explains how the two tools fit together.