Customer service channel preferences in 2026 split by the kind of problem: phone is the first choice again for complex or urgent issues, while most customers still try self-service first. ContactBabel measured phone preference for complex inquiries rising from 28% in 2018 to 47% in 2024. Gartner finds 73% of customers use self-service at some point in a service journey, but only 14% of issues are fully resolved there. The channel customers start in and the channel that solves the problem are usually not the same.
This article compares recent independent studies on channel preference: what each one asked, who it asked, and what it found for phone, email, chat, self-service and AI. The short answer up front: no channel wins on its own, and the most consistent finding across the studies is that customers want to move between channels without starting over. The fix that pays off in every channel is a knowledge base that stays correct, because self-service, chat, AI assistants and human agents all answer from it. For SaaS teams, HappySupport runs the help center, AI chat, ticket inbox and in-app widget from one set of articles, so every channel inherits the same current answer.
Every number below links to the study it comes from. Where a figure comes from a vendor that sells one of the channels it measures, the vendor is named.
How we selected the studies
We used independent research, plus research from contact center vendors labelled as such. The independent sources are Gartner, McKinsey, Capgemini Research Institute, Bitkom, YouGov, ContactBabel and SQM Group. The vendor sources are Five9, Genesys and Cyara, which sell voice, chat or AI channels, so we name them wherever their figures appear. We did not use statistics from help desk and help center vendors, because they sell the channels they measure and because many of their figures circulate without a traceable report behind them.
Studies ask different questions. "Which channel do you prefer?" is not the same question as "which channel would you use for an urgent problem?" or "how likely are you to call after self-service failed?". We report each figure with the question behind it and do not average across studies.
Disclosure: HappySupport sells help center software, which feeds the self-service channel. The self-service data below includes the findings that are unflattering for self-service.
This article was rewritten in October 2026. An earlier version cited figures we could not trace back to the named reports; they have been removed.
Key findings
Across the studies, customers choose the channel by the problem, try self-service first, and judge the experience by whether they have to start over when they switch.
- Phone is back for hard problems. Phone preference for complex inquiries rose from 28% in 2018 to 47% in 2024, and phone was the first choice for urgent interactions for 49% of US consumers in 2024 (ContactBabel via ICMI).
- Self-service is tried first and rarely finishes the job. 73% of customers use self-service at some point, but only 14% of issues are fully resolved there, and only 36% even for issues customers call "very simple" (Gartner, 2024).
- Self-service fails on content. In 43% of failed self-service attempts, customers could not find content relevant to their issue (Gartner, 2024).
- Email and phone lead stated preference. In Britain, email (33%) and phone (31%) are the most preferred channels; only 1% prefer chatbots, though 18% use them (YouGov, 2025).
- Customers want a person within reach. 62% of German online shoppers would most like a quickly reachable human when something goes wrong, and 36% want a chatbot (Bitkom, 2025).
- Continuity is the biggest gap. 97% of consumers prioritize moving between channels without repeating themselves, but only 16% of CX leaders say their organization delivers fully integrated omnichannel service (Genesys, 2025, vendor research).
- Seamless handoff doubles satisfaction. Customers who used more than one channel for the same issue rated a seamless experience at 67% CSAT and a non-seamless one at 28% (SQM Group).
- AI chat has a trust problem. 64% of customers would prefer companies did not use AI in customer service, and the top concern is that it will get harder to reach a person (Gartner, 2024).
Channel preference: what each study measured
Each study asked a different question, so the table lists who was asked and what was asked next to the headline number.
| Study | Year | Who was asked | Question | Headline finding |
|---|---|---|---|---|
| ContactBabel via ICMI | 2024 | US consumers | Preferred channel by inquiry type | Phone first for urgent issues (49%); phone for complex inquiries up from 28% (2018) to 47% |
| Five9 Customer Experience Report (vendor) | 2025 | 1,006 consumers, US, UK, Canada | Preferred support channel | 56% prefer phone for general issues, 74% for complex or urgent ones |
| YouGov | 2025 | British adults | Preferred and used channels | Email 33%, phone 31% preferred; chatbots 1% preferred, 18% used |
| Bitkom | 2025 | 978 German online shoppers | Wanted contact options for order problems | Reachable human 62%, email 52%, callback 40%, chatbot 36%, FAQ 28% |
| Capgemini Research Institute | 2025 | 9,500 consumers | Human vs automated service | More than 70% prefer a human, although most say chatbots have improved |
| Gartner | 2024 | 5,728 customers | Self-service use and resolution | 73% use self-service at some point; 14% of issues fully resolved there |
| McKinsey | 2024 | 3,500 consumers | Channel used after self-service failed | Phone likely for 71% (Gen Z) to 94% (boomers) |
Three patterns hold across these studies. Phone gains share as problems get harder or more urgent. Email remains a top stated preference wherever it is asked about. And self-service is the most used entry point but resolves a small share of issues on its own.
Phone: where it still wins
Phone is the channel customers pick when the stakes go up. ContactBabel's US consumer data shows phone preference for complex inquiries rising from 28% in 2018 to 47% in 2024, and phone as the first choice for urgent interactions at 49% in 2024 (ICMI). ContactBabel's newer guide reports that the complex-inquiry share eased slightly after the 2024 peak but phone stayed on top, and that 42% would choose phone to report an incorrect delivery (Contact Center Pipeline, 2026).
Five9's 2025 survey of 1,006 consumers points the same way: 56% prefer phone for general issues, rising to 74% for complex or urgent matters (Five9, a contact center software vendor). In a separate 2024 Five9 study of 4,000 US and UK adults, 75% preferred talking to a real human in person or by phone for support (Five9, 2024).
The lesson for a support team is not "staff more phones". It is that customers escalate to the most direct channel when the earlier ones failed them, which makes the earlier ones the cheaper place to fix.
Self-service: tried first, resolved rarely
Self-service is where most customer journeys begin. Five9 finds 86% of consumers explore self-service options before reaching out to support (Five9, 2025, vendor research), and Gartner finds 73% of customers use self-service at some point in their journey (Gartner, 2024). Expectation is long established: in Microsoft's 2018 survey of 5,000 consumers in Brazil, Germany, Japan, the UK and the US, 88% expected a brand to offer an online self-service portal (Microsoft State of Global Customer Service, 2018; survey run in 2018, published 2019).
Resolution is the problem. Gartner's survey of 5,728 customers found that only 14% of customer service issues are fully resolved in self-service, and only 36% of issues that customers describe as "very simple". The most common reason for failure, in 43% of cases, was that customers could not find content relevant to their issue. And 45% of customers who started in self-service said the company did not understand what they were trying to do.
“AI should absorb complexity for the customer, not create new complexity around the customer.
”
That is the self-service gap in one line. The channel is not failing because customers avoid it; they try it first. It fails because the content behind it is missing, outdated or impossible to find. For the metric side of this gap, see the self-service rate metric explained. For the reading-behavior side, see why users do not read the help center.
Email: the preference nobody talks about
Email is rarely the headline channel and consistently one of the most preferred. In YouGov's 2025 British data, email is the most preferred channel at 33%, just ahead of phone at 31%, and 74% of Britons tend to contact businesses by email. Preference peaks among Gen X at 36% (YouGov). In Germany, 52% of online shoppers want an email service for problems, and 77% of those who used one were satisfied with it (Bitkom, 2025).
For B2B SaaS, email (or a ticket that replies by email) matters even more than in consumer markets: issues are technical, take more than one exchange, and the customer wants a written record. Deprioritizing it to push customers into chat usually moves the conversation, not the problem.
Chat and AI: high use, low trust
Chat is where the gap between use and preference is widest. YouGov finds 18% of Britons use chatbots but only 1% prefer them. Gen Z is the exception for live chat: 15% of Gen Z prefer it, against 2% of the Silent Generation (YouGov, 2025).
Satisfaction data explains why. Bitkom finds 86% of German online shoppers satisfied after contact with a human, against 50% of chatbot users satisfied with the chatbot. In a Forrester Consulting study commissioned by Cyara, consumers rated their chatbot experiences 6.4 out of 10 on average, and 30% said they would take their purchase to another brand after a bad chatbot experience (Cyara, 2023). Genesys found only 21% of consumers highly satisfied with a chatbot; the top frustrations were not being able to reach a live agent and having to repeat the conversation to an agent (Genesys, 2023, vendor research).
Customers are not against automation as such. Five9 finds 72% are open to AI-powered interactions provided they can escalate to a human, and 59% prefer an instant AI chatbot over waiting for a live agent (Five9, 2025). What they reject is being trapped: Gartner's top customer concern about AI in service is that it will become harder to reach a person, and 64% would prefer companies did not use AI at all (Gartner, 2024).
“The most successful customer-facing AI focuses on automating CRaP: Confident, Routine, Predictable.
”
Jeff's framing matches the data. Routine questions with a known answer convert well into chat and AI. Questions the documentation does not cover need a human, or a knowledge base that has caught up with the latest release. For how the knowledge layer determines chat quality, see help center vs knowledge base.
The new entry point: third-party AI and search
A growing share of service journeys never starts in a company's own channels. Gartner's 2025 survey of 5,801 customers found that 51% of customer service journeys now begin on third-party platforms such as Google, YouTube and ChatGPT, and only 22% start, stay and resolve entirely in the company's own channels (Gartner, 2025). In 2026, Gartner found customers about three times more likely to use third-party GenAI tools than company-provided chatbots to resolve service issues (Gartner, 2026).
For a support team this changes what "self-service channel" means. Your public help center is now also the source that search engines and AI assistants answer from. If it is wrong, the customer gets the wrong answer before they ever reach you.
Generational split
Generation shifts the default channel, but it does not remove the phone. McKinsey asked consumers who had already tried self-service how likely they were to contact customer care through each channel (McKinsey, 2024, survey of 3,500 consumers, May 2023):
| Channel after self-service | Gen Z | Millennials | Gen X | Boomers |
|---|---|---|---|---|
| Phone | 71% | 81% | 86% | 94% |
| Live chat | 74% | 82% | 81% | 76% |
| 75% | 84% | 87% | 86% | |
| Social media | 57% | 59% | 34% | 19% |
Gen Z is less phone-first than older customers, not phone-averse: 71% would still call. McKinsey's text puts it plainly: respondents of all ages said live phone conversations were among their most preferred ways to contact companies for help.
Other studies show the same shape. In Britain, phone is the top channel for half of the Silent Generation but for 16% of Gen Z (YouGov, 2025). Five9 found 66% of Gen Z prefer a real human in person or by phone, against 76% of Gen X and 86% of boomers and the Silent Generation (Five9, 2024). In Germany, Genesys found a live phone agent among the preferred channels for 51% of Gen Z and millennials and 75% of Gen X and boomers, while 46% of the younger group value self-service options such as chatbots and FAQs, against 20% of the older group (Genesys, 2024, vendor research).
Omnichannel expectation vs reality
The most consistent finding across the studies is that customers want to switch channels mid-issue without repeating themselves.
- 97% of consumers prioritize seamless omnichannel experiences and expect to move between channels without repeating themselves; 16% of CX leaders say they provide fully integrated omnichannel experiences (Genesys, 2025, vendor research).
- Consumers in the same Genesys study ranked first-contact resolution as the most valued part of a service interaction; CX leaders ranked it ninth.
- Customers who used more than one channel for the same inquiry gave a seamless experience 67% CSAT and a non-seamless one 28% (SQM Group).
- SQM also finds one-contact resolution earns the highest CSAT at 89%, while 42% of contacts take two or more attempts to resolve.
- In Microsoft's 2018 survey, more than 75% expected agents to see their previous interactions, and customers ranked an agent's lack of knowledge about their issue and having to repeat themselves as the two most frustrating parts of poor service (Microsoft, 2018).
The gap is rarely solved by adding another channel. It is solved by making the knowledge and context beneath every channel consistent and current. For the ticket-volume side of that mechanism, see reducing support tickets with a help center.
What this means for B2B SaaS support
Most of these studies survey consumers. B2B SaaS support has different dynamics, and the data points to a clear order of priorities:
- Fix self-service content first. Gartner's 43% "could not find relevant content" is the largest single cause of failure in the data, and it is a content problem, not a channel problem. A SaaS help center is the substrate every other channel reads from.
- Keep a written channel. Email or a ticket that replies by email covers the multi-step, technical issues typical of B2B, and it is a top stated preference in every study that asks about it.
- Use AI chat for routine answers, with a visible way out. 72% accept AI when they can escalate to a human (Five9, 2025); the top fear is being unable to reach one (Gartner, 2024).
- Pass context on handoff. The 67% vs 28% CSAT gap (SQM) is about continuity: the agent should see what the customer already tried.
- Assume customers ask an AI assistant first. With 51% of journeys starting on third-party platforms (Gartner, 2025), public documentation has to be correct and crawlable.
The shared failure mode across every channel is the documentation underneath. When the help center is stale, the chatbot answers wrongly, the live chat agent improvises, and the customer switches channels looking for someone who knows the current state.
How HappySupport fits the channel question
Every channel in this article reads from the same knowledge layer. The chatbot is grounded in it, the agent answering an email searches it, and third-party AI assistants quote it. When the knowledge layer drifts behind the product, every channel inherits the drift.
HappySupport is the help center built around that constraint. Git Handoff reads code changes on GitHub, GitLab or Bitbucket, finds the help articles they affect and drafts updates that a human reviews before publishing. The HappyRecorder captures UI walkthroughs as DOM and CSS metadata so screenshots can be refreshed when the interface changes. For the architecture, see how a self-updating help center works and the hidden cost of documentation decay.
As self-service software, HappySupport covers the help center, the AI chat, the ticket inbox and the in-app widget in one product with one flat monthly price and no per-agent pricing. The AI chat answers only from your own articles, with citations, and escalates to a human by opening a ticket with the visitor's email and an AI-written summary; the team replies by email from the ticket inbox. If you want to keep Intercom, Zendesk, HubSpot, Help Scout, Front or Freshdesk, you can: 1-click import from Zendesk and Intercom, migration from the others, automatic redirects for every old URL. Keeping the existing helpdesk is an option, not a requirement.
Book a demo