Customer support
glossary
63 terms defined in plain English, from the metrics on your dashboard to the AI vocabulary in your vendor contracts. No jargon for its own sake.
AI and automation
Agent assist Agent assist supports a human agent in real time with suggested replies and retrieved context. Here is where it helps, where it is oversold, and how it is usually priced. Agentic AI Agentic AI plans multi-step work and takes actions rather than only answering. Here is what that means in support, where it helps, and the guardrails it requires. AI agent An AI support agent interprets a customer question, retrieves the right information, and acts to resolve it. Here is how it differs from a chatbot and what it can safely do. Chatbot A chatbot answers customers through a conversational interface, usually from scripted flows. Here is how chatbots differ from AI agents and where they still work well. Conversational AI Conversational AI is the umbrella term for systems that hold natural-language dialogue with customers. Here is what falls under it and how to compare offerings. Grounding Grounding restricts an AI system to answering from verified sources you control. Here is why it matters, how it is enforced, and how to tell whether a vendor really does it. Hallucination A hallucination is a fluent, confident, incorrect AI answer. Here is why it happens, what it costs a support team, and the controls that reduce it. Intent recognition Intent recognition classifies what a customer is actually trying to do. Here is how it drives routing and automation, and why fixed intent taxonomies age badly. Large language model A large language model predicts text and powers modern AI support. Here is what it is good at, where it fails in support, and what it needs around it to be useful. Retrieval augmented generation RAG retrieves relevant documents and gives them to a language model before it answers. Here is why it matters for support accuracy and how it fails in practice. Sentiment analysis Sentiment analysis scores the emotional tone of customer messages. Here is how support teams use it for prioritisation and where the signal is weaker than it looks. Support automation Support automation removes manual steps from the handling of customer issues. Here is what to automate first, what to leave alone, and how to sequence a programme.
Analytics
Conversation analytics Conversation analytics extracts themes and signals from support conversations at scale. Here is what it can reveal and where it misleads. Voice of the customer Voice of the customer is the structured capture of customer feedback across channels. Here is how support fits in and why most VoC programmes stall.
Automation
Automation rate Automation rate is the share of support work handled without human action. Here is how to define it usefully and why the denominator decides everything. Containment rate Containment rate is the share of conversations handled without reaching a human. Here is why it flatters automation and what to measure beside it. Ticket deflection Ticket deflection is any action that stops a support request becoming a human-handled ticket. Here is how it is measured, why the metric misleads, and what to track instead. Zero-touch resolution Zero-touch resolution means a customer issue is solved end to end with no human involvement. Here is how it is measured and which issue types genuinely qualify.
Business metrics
Churn rate Churn rate is the share of customers who leave in a period. Here is how support affects churn and why support-driven churn is systematically underestimated. Customer lifetime value Customer lifetime value estimates total profit from a customer relationship. Here is how support costs and retention both feed into it.
Channels
Multichannel support Multichannel support means offering several contact channels that operate separately. Here is what it costs, when it is enough, and how it differs from omnichannel. Omnichannel support Omnichannel support gives customers one continuous conversation across every channel. Here is how it differs from multichannel and what it actually requires.
Content
Help centre A help centre is the public front end of your support knowledge. Here is how it differs from a knowledge base, and how to make it work for customers and AI alike. Knowledge base A knowledge base is the structured store of answers your customers and AI both read. Here is how to structure it, keep it current, and measure whether it works. Knowledge gap analysis Knowledge gap analysis finds the questions your documentation cannot answer. Here is how to run one and how AI deployment produces the data for free. Self-service rate Self-service rate measures how many customers solve their own problem without contacting support. Here is how to measure it honestly and how to raise it.
Core metrics
Average handle time Average handle time measures the mean time an agent spends on a contact end to end. Here is the formula, what it hides, and why cutting it can backfire. Customer effort score CES measures how hard a customer had to work to get their issue handled. Here is how it is calculated, why it predicts loyalty, and how to lower effort in practice. Customer satisfaction score CSAT measures satisfaction with a single interaction, usually on a one-to-five scale. Here is the formula, the response bias problem, and how to read the number honestly. First contact resolution First contact resolution measures the share of issues solved in a single interaction. Here is how to calculate FCR, why it correlates with satisfaction, and how to raise it. First response time First response time measures how long a customer waits for the first meaningful human or AI reply. Here is how to calculate it, what distorts it, and how to improve it. Net promoter score NPS measures relationship-level loyalty on a zero-to-ten recommendation scale. Here is the formula, what the score can and cannot tell you, and how support affects it. Reopen rate Reopen rate is the share of resolved tickets that come back. Here is why it is the best honesty check on every other support metric. Resolution A resolution is a customer issue that is genuinely closed with a correct answer. Here is how resolutions are defined, counted, and abused in vendor billing. Resolution rate Resolution rate is the share of contacts genuinely resolved, often by AI without a human. Here is the formula, the definitions that inflate it, and how to verify it. Ticket volume Ticket volume is the count of support requests in a period. Here is how to normalise it, what drives it, and why raw volume is a misleading target.
Cost
Cost per contact Cost per contact is the average cost of handling one customer interaction. Here is the formula, what it hides, and why cost per resolution is usually the better metric. Cost per resolution Cost per resolution is the fully loaded cost of solving one customer issue. Here is the formula, what to include, and why it beats cost per ticket as a planning number.
Operations
Canned response A canned response is pre-written reply text an agent inserts into a conversation. Here is where canned responses help, where they hurt, and how they differ from macros. Escalation Escalation moves an issue to greater authority or expertise. Here is how to design escalation paths that resolve faster instead of just moving the ticket. Escalation matrix An escalation matrix maps severity and time to who gets involved. Here is how to build one that works at three in the morning and does not need judgement. Handoff A handoff transfers a conversation from AI to a human agent. Here is what a good handoff carries, what a bad one costs, and why it should never be billable. Macro A macro is a saved set of actions and text an agent applies to a ticket. Here is how macros speed up support, and how macro libraries quietly rot. Service level agreement An SLA is a committed standard for response and resolution times. Here is how support SLAs are structured, why they break, and what to promise. Skills-based routing Skills-based routing matches tickets to agents by capability rather than availability. Here is when it pays off and the failure modes it introduces. Support backlog A support backlog is the accumulated set of unresolved tickets. Here is why backlogs form, why hiring rarely clears them, and how to make the fix stick. Support quality assurance Support QA reviews conversations against a standard to improve quality and coaching. Here is how to run QA that helps rather than polices. Support queue A support queue is the ordered set of tickets waiting for attention. Here is how queue design affects wait times, and why the ordering rule matters more than the length. Ticket ageing Ticket ageing tracks how long unresolved tickets have been open. Here is why age beats volume as a health metric and how to stop tickets going stale. Ticket routing Ticket routing sends each request to the right person or queue automatically. Here is how routing strategies compare and why bad routing is invisible. Ticket tagging Ticket tagging labels requests so they can be reported and automated on. Here is how tag taxonomies decay and how to build one that survives. Triage Triage is the first assessment of an incoming ticket to set priority and destination. Here is how to design a triage step that is fast, consistent, and cheap. Workflow automation Workflow automation applies rules to tickets so routing, priority, and follow-up happen without a person. Here is how to design rules that stay maintainable.
Pricing models
Per-conversation pricing Per-conversation pricing bills for every customer conversation regardless of outcome. Here is how it works, why it can reward failure, and how it compares to per-resolution. Per-resolution pricing Per-resolution pricing bills only when a customer issue is genuinely solved. Here is how the model works, what to check in the definition, and how it compares to seats. Per-seat pricing Per-seat pricing charges a fixed fee per agent licence. Here is how it works, where the hidden costs appear, and how it compares to per-conversation and per-resolution models. Usage-based pricing Usage-based pricing bills for what you actually consume rather than for licences. Here is how it applies to support software, the forecasting problem, and how to bound it.
Tools
Help desk A help desk is the system and team that receives and resolves customer requests. Here is what a help desk includes, how it differs from a service desk, and what to look for. Service desk A service desk is the single point of contact for internal service requests and incidents. Here is how it differs from a help desk and where the practices overlap. Shared inbox A shared inbox lets a team handle one email address collaboratively with assignment and visibility. Here is how it differs from a help desk and when to graduate.
Get Started
See Fidiora resolve a ticket in 60 seconds.
No credit card, no sales call required. Connect your docs and watch it work.
No credit card · Live in under an hour