Industries
Why Chat Support Is Not Just Cheaper Phone Support, and How to Staff It Right
Why Chat Support Is Not Just Cheaper Phone Support, and How to Staff It Right
Key Takeaways
By Andy Schachtel, CEO of Sourcefit | Global Talent and Elevated Outsourcing
- Chat support has become the preferred channel for customers under 45, with adoption growing 30 to 40% year-over-year across most industries, yet most companies staff it as an extension of their phone team rather than as a distinct discipline with its own skill requirements, training needs, and quality standards.
- A skilled chat agent handles three to four simultaneous conversations with 30-second or faster response times, which requires a cognitive skill set fundamentally different from phone support: rapid written communication, parallel context management, and the ability to convey empathy and brand personality through text alone.
- Chat produces higher customer satisfaction than phone when staffed correctly, because customers choose chat for its speed and convenience; when those expectations are met, satisfaction scores of 88 to 93% are achievable, but when response times exceed 90 seconds, abandonment rates spike and satisfaction craters.
- The economics of chat are compelling: a chat agent handling three concurrent conversations effectively triples the throughput of a phone agent at the same hourly cost, but only if the agent is recruited and trained specifically for chat rather than reassigned from phone duty.
An AI-powered identity verification company learned the distinction between chat and phone the expensive way. In 2023, they expanded their customer support channels to include live chat alongside their existing phone and email operation. The logic seemed sound: the existing agents were experienced professionals who knew the product, understood the highly technical customer base spanning financial services and fintech sectors, and had strong satisfaction scores on phone. The logic was wrong.
Within the first month, the data told the story. Average chat response time was 72 seconds, nearly two and a half times the 30-second benchmark. Chat CSAT was 64%, compared to the team’s 91% phone CSAT. Abandonment rate was 35%. The agents were not failing because they lacked product knowledge or customer empathy. They were failing because they were applying phone skills to a chat medium. They wrote long, conversational messages that took 45 seconds to compose. They handled one chat at a time because managing multiple conversations simultaneously was a skill they had never developed. They tried to build the same verbal rapport on chat that worked on phone, which felt warm on a call but felt slow and inefficient in a text window.
The company rebuilt its chat operation from scratch with its outsourcing partner, recruiting specifically for chat aptitude: typing speed above 60 words per minute, comfort with concurrent task management, and the ability to communicate clearly and warmly in short written messages. The new team hit 92% CSAT within 90 days. Same product. Same customers. Different skill set.
The Cognitive Difference Between Phone and Chat
The reason phone agents struggle with chat is not effort or attitude. It is cognitive architecture. Phone support is a serial, auditory, verbal task. The agent focuses on one customer at a time, processes information through listening, and responds through speaking. The interaction flows linearly: customer speaks, agent listens, agent responds, customer speaks again. The agent’s working memory manages one conversational thread.
Chat support is a parallel, visual, written task. The agent manages three to four conversations simultaneously, processes information by reading, and responds by typing. The interactions are interleaved: while waiting for Customer A to respond, the agent is composing a message for Customer B, reviewing the history of Customer C’s issue, and noting that Customer D has been waiting 20 seconds and needs a check-in message. The agent’s working memory manages four independent conversational threads, each at a different stage of resolution, while maintaining the accuracy and tone appropriate to each.
This is a fundamentally different cognitive demand. Research on task switching shows that most people experience a 20 to 40% efficiency loss when alternating between tasks. Chat agents who have been specifically recruited for parallel processing ability and trained to manage concurrent conversations do not experience this loss because the skill has been selected for and developed. Phone agents who are asked to switch to chat without this selection and training experience the full efficiency penalty, which manifests as slow response times, confused context between conversations, and the frustrating sense that they are performing worse despite working harder.
Phone vs. Chat: Agent Skill Comparison
| Skill Dimension | Phone Agent | Chat Agent |
|---|---|---|
| Communication Mode | Verbal; real-time spoken conversation | Written; rapid text composition |
| Typing Speed Requirement | Minimal; notes and documentation only | 60+ WPM with accuracy under pressure |
| Concurrent Capacity | 1 conversation | 3-4 simultaneous conversations |
| Empathy Expression | Through vocal tone, pacing, warmth | Through word choice, punctuation, response speed |
| Response Time Expectation | Immediate (real-time verbal) | < 30 seconds between messages |
| Information Processing | Auditory; linear conversation flow | Visual; parallel thread management |
| Resolution Style | Guided conversation toward resolution | Concise, direct answers with links/resources |
| Quality Measurement | Call recording audits; tone and resolution | Transcript audits; speed, accuracy, concurrency |
Recruiting for Chat: What to Screen For
The recruiting process for chat agents should be as distinct from phone agent recruiting as the job itself is distinct. The core competencies to screen for are typing speed and accuracy, parallel task management, written communication quality, and temperament under time pressure.
Typing speed is the most objective filter. An agent who types 40 words per minute cannot maintain 30-second response times across three concurrent conversations. The minimum threshold for chat agents should be 55 to 60 WPM with 95% accuracy. This is assessed through a timed typing test during the screening process, and it eliminates roughly 30 to 40% of applicants who would otherwise appear qualified based on their resume.
Parallel task management is assessed through simulation. The candidate is presented with two to three simultaneous chat scenarios and asked to manage them concurrently, switching between conversations while maintaining accuracy and appropriate response times. The assessment measures not just whether the candidate can handle the volume, but whether they maintain quality, tone, and context accuracy across all conversations. Candidates who freeze when the second conversation begins, or who start confusing details between conversations, are not suited for chat regardless of their other qualifications.
Written communication quality is evaluated through the chat simulation as well as a separate written assessment. Chat messages must be clear, grammatically correct, appropriately toned, and concise. A chat message that reads like a spoken sentence transcribed verbatim usually feels awkward in a chat window. The best chat agents write in a style that is professional but conversational, warm but efficient. They know when to use a complete sentence and when a fragment with the right punctuation conveys the same meaning in half the words.
The Economics: Why Chat Changes the CX Cost Equation
The economic argument for chat is built on concurrency. A phone agent handles one interaction at a time. At an average handle time of 8 minutes per call, a phone agent completes approximately 7 interactions per hour. A chat agent handling three concurrent conversations with an average handle time of 12 minutes per conversation completes approximately 15 interactions per hour. The chat agent produces more than double the throughput at the same hourly labor cost.
For an outsourced operation, this concurrency advantage translates directly to cost per interaction. If a Philippine-based CX agent costs $2,000 per month on a per-agent model, the phone cost per interaction is approximately $1.70 assuming 7 interactions per hour over 168 productive hours per month. The chat cost per interaction at the same agent cost is approximately $0.80. The per-interaction cost of chat is less than half the cost of phone, which means every customer interaction shifted from phone to chat reduces the overall CX cost without reducing quality, provided the chat operation is properly staffed and managed.
This is why chat adoption is growing so rapidly in CX operations worldwide. It is not just that customers prefer it. It is that the unit economics are dramatically better for the company as well. The alignment of customer preference and company economics in the same direction is rare, and it makes chat investment one of the highest-ROI decisions a CX leader can make.
When Chat Is Not the Right Channel
Chat is not universally superior to phone. There are interaction types where phone remains the better channel, and forcing those interactions into chat degrades the experience for both the customer and the agent.
Emotionally charged interactions, where a customer is upset, confused, or dealing with a stressful situation, are generally better handled by phone. The human voice conveys empathy more effectively than text. The real-time nature of verbal conversation allows the agent to hear frustration building and adjust tone and approach in real time. Chat can handle mild dissatisfaction, but genuine distress benefits from the warmth and immediacy of a phone conversation.
Complex technical troubleshooting that requires a guided, step-by-step walkthrough is also better suited to phone. When the agent needs to direct the customer through a sequence of actions, confirming each step before proceeding to the next, the back-and-forth of chat introduces latency that extends the interaction and increases the risk of miscommunication. A phone agent can guide a customer through a 10-step troubleshooting sequence in 8 minutes. The same sequence on chat may take 20 minutes and generate more errors.
The optimal approach is channel routing: directing each interaction to the channel that serves it best. Routine inquiries, order status checks, simple questions, and straightforward transactions route to chat. Complaints, technical escalations, and complex account discussions route to phone. The routing can be managed through the customer’s initial channel choice, supplemented by agent-initiated transfers when a chat interaction reveals that the issue would be better served by phone.
Frequently Asked Questions
How many concurrent chats should we expect agents to handle?
Three simultaneous conversations is the standard for trained chat agents handling moderate-complexity interactions. Agents handling simple, highly repetitive interactions like order status or shipping inquiries can manage four to five. Agents handling complex technical or billing issues should be limited to two concurrent conversations. The capacity should be calibrated to the complexity of the interactions, not set at a universal number, and should be adjusted based on quality data. If accuracy or CSAT begins to decline, reduce the concurrency target before the quality gap becomes visible to customers.
What is an acceptable chat response time?
Industry best practice is an initial response within 15 to 30 seconds and subsequent responses within 30 to 60 seconds. Response times beyond 90 seconds trigger significant abandonment, typically 20 to 30% of waiting customers will leave. The standard should be measured as an average across all conversations, with a maximum threshold that no individual response should exceed. An average of 25 seconds with a maximum of 60 seconds is a strong target for most operations.
Should we use chatbots to supplement human chat agents?
Chatbots are effective for initial triage and for resolving simple, highly structured interactions: account balance inquiries, order status, password resets, and store location lookups. They are not effective for anything that requires judgment, empathy, or creative problem-solving. The best implementation uses a chatbot to handle the first interaction, attempt resolution for simple issues, and seamlessly transfer to a human agent when the issue exceeds the chatbot’s capability. The transfer must include the full conversation context so the customer does not repeat themselves.
How do we measure chat quality differently from phone quality?
Chat quality audits evaluate written communication quality including grammar, clarity, tone, and brand voice; response time adherence; concurrent conversation management without context confusion; accuracy of information provided; and resolution completeness. Unlike phone audits that assess vocal tone and listening skills, chat audits focus on the written interaction and the agent’s ability to maintain quality across multiple simultaneous conversations. Both channel audits should contribute to an overall quality score weighted by interaction volume.
Can chat support work for B2B companies with complex products?
Yes, particularly for Tier 1 support functions: initial issue triage, common question resolution, documentation and resource sharing, and ticket creation for issues requiring specialist involvement. B2B chat agents need deeper product training and access to more comprehensive knowledge bases than B2C agents, but the channel mechanics are the same. Many B2B companies find that chat is preferred by their customers for quick questions that do not warrant scheduling a call with a support engineer, and it reduces the load on expensive Tier 2 and Tier 3 specialists.
To learn more about how SourceCX builds high-performance chat support operations, visit sourcecx.com or contact our team for a consultation.
