The support & AI glossary.

Agent Memory

Agent memory is an AI agent's ability to retain and recall information across a conversation — or across sessions — such as earlier messages, customer details, and past interactions, so it stays coherent instead of treating every turn as brand new..

Agentic AI

Agentic AI is AI that doesn't just answer questions but takes actions — it plans multi-step tasks, calls tools and APIs, and works toward a goal end to end with little or no human intervention..

AI Agent

An AI agent is a software system that uses a language model to understand a request, decide what to do, and take actions — calling tools, retrieving data, and completing tasks — rather than just returning a single scripted answer..

AI Agent vs Chatbot

The difference between an AI agent and a chatbot is that a chatbot answers questions with scripted or generated replies, while an AI agent can reason, use tools, and take actions to resolve a request end to end rather than just responding..

AI Copilot

An AI copilot is an assistant that works alongside a human — suggesting replies, summarizing context, and surfacing information — while the person stays in control and makes the final decision, rather than acting autonomously..

AI Guardrails

AI guardrails are the rules, checks, and constraints placed around an AI system to keep its behavior safe, on-topic, and within policy — controlling what it can say, do, and access..

AI Hallucination

An AI hallucination is when a language model generates a response that is fluent and confident but factually wrong or fabricated — inventing details, policies, or sources that don't exist..

AI Orchestration

AI orchestration is the coordination of multiple AI models, agents, tools, and data sources into a single reliable workflow — deciding what runs, in what order, and how results are passed along to complete a task..

AI Summarization

AI summarization is the use of a language model to condense long text — a ticket thread, chat transcript, or knowledge article — into a shorter, accurate summary that captures the key points and context..

AI Support Agent

An AI support agent is a software system that uses AI — typically a large language model plus tools and access to your knowledge sources — to understand customer requests and respond to or resolve them, either autonomously or by assisting a human agent..

AI Ticket Automation

AI ticket automation is the use of AI to handle support tickets automatically — classifying, tagging, routing, drafting replies, or fully resolving them — reducing the manual work agents do on each ticket..

Autonomous Agent

An autonomous agent is an AI system that pursues a goal on its own — deciding what steps to take, calling tools, and acting on the results — with little or no step-by-step human direction..

Chatbot

A chatbot is a software program that simulates conversation with users through text, ranging from simple rule-based scripts that follow decision trees to AI-powered bots that understand natural language..

Confidence Score

A confidence score is a value an AI system assigns to a prediction or answer that estimates how likely it is to be correct — used to decide whether the AI should act automatically, ask for clarification, or hand off to a human..

Context Window

The context window is the maximum amount of text — measured in tokens — that a language model can consider at once, including both the input you give it and the output it generates..

Conversational AI

Conversational AI is technology that lets people interact with software through natural back-and-forth dialogue — understanding intent, maintaining context across turns, and responding in natural language via chat or voice..

Deflection vs Resolution

Deflection vs resolution is the distinction between preventing a ticket from reaching a human (deflection) and actually solving the customer's problem (resolution) — two outcomes that are easy to confuse but measure very different things..

Embeddings

Embeddings are numeric vector representations of text (or images, code, etc.) that capture meaning, so that pieces of content with similar meaning sit close together in a high-dimensional space..

Entity Extraction

Entity extraction is the process of pulling specific pieces of structured information — like names, order numbers, dates, products, or amounts — out of unstructured text so a system can act on them..

Few-Shot Learning

Few-shot learning is when you guide an AI model to perform a task by including a small number of examples in the prompt, so it can follow the demonstrated pattern without any retraining..

Fine-Tuning

Fine-tuning is the process of further training a pre-trained language model on a smaller, task-specific dataset so it adapts to a particular domain, style, or behavior — updating the model's weights rather than just its instructions..

Foundation Model

A foundation model is a large AI model trained on broad, general-purpose data at scale that can be adapted to many downstream tasks — through prompting, retrieval, or fine-tuning — rather than being built for one narrow job..

Function Calling

Function calling is a capability that lets a language model request that a predefined function or API be run — returning the function name and structured arguments — so the model can fetch live data or take actions instead of only generating text..

Generative AI

Generative AI is a class of AI that creates new content — text, images, code, or audio — by learning patterns from training data, rather than only classifying or predicting from existing inputs..

Grounding

Grounding is the practice of tying an AI model's answers to verified source material — your documentation, live data, or knowledge base — so responses reflect real facts rather than the model's own guesses..

Intent Recognition

Intent recognition is the process of identifying what a customer is trying to accomplish from their message — such as "track an order," "request a refund," or "cancel a subscription" — so the system can route or resolve it correctly..

Large Language Model (LLM)

A large language model (LLM) is a neural network trained on vast amounts of text to predict and generate language, enabling it to understand questions, summarize, classify, and write human-like responses..

Machine Learning

Machine learning is a branch of AI in which systems learn patterns from data to make predictions or decisions, rather than following rules a programmer wrote by hand..

Model Context Protocol (MCP)

The Model Context Protocol (MCP) is an open standard that defines a common way for AI applications to connect to external tools, data sources, and systems — so a model can access them through a single, consistent interface instead of a custom integration for each..

Multi-Agent System

A multi-agent system is a setup where several specialized AI agents work together — each handling a specific task or domain — coordinated so they can solve problems that a single agent would handle less reliably..

Natural Language Processing (NLP)

Natural language processing (NLP) is the field of AI focused on enabling computers to understand, interpret, and generate human language — the foundation for tasks like intent recognition, sentiment analysis, and text classification..

Prompt Engineering

Prompt engineering is the practice of designing and refining the instructions given to a language model to get accurate, relevant, and reliable outputs for a specific task..

Prompt Injection

Prompt injection is a security attack where malicious instructions are hidden in user input or external content to manipulate an AI system into ignoring its original instructions or performing unintended actions..

RAG vs Fine-Tuning

RAG vs fine-tuning is the choice between grounding an AI model in external knowledge at query time (retrieval-augmented generation) versus adjusting the model's own weights on your data (fine-tuning) — two different ways to make a general model useful for your specific needs..

Retrieval-Augmented Generation (RAG)

Retrieval-augmented generation (RAG) is a technique where an AI model retrieves relevant information from your own knowledge sources at query time and uses it to ground its answer, instead of relying only on what it memorized during training..

Semantic Search

Semantic search is a search technique that matches content by meaning rather than by exact keywords, using embeddings to find results that are conceptually related to the query even when they share no words with it..

Sentiment Analysis

Sentiment analysis is the automated classification of the emotional tone of a message — typically positive, negative, or neutral — so teams can gauge how a customer feels and respond appropriately..

System Prompt

A system prompt is the underlying set of instructions that defines an AI model's role, behavior, tone, and boundaries for a conversation — set by the developer, not the end user, and applied to every interaction..

Temperature (LLM)

Temperature is a setting that controls how random or deterministic a language model's output is: low temperature produces focused, predictable responses, while high temperature produces more varied and creative ones..

Text Classification

Text classification is the task of automatically assigning a category or label to a piece of text — such as tagging a support ticket by topic, language, or priority..

Tokens (LLM)

Tokens are the small chunks of text — words, parts of words, or characters — that a language model reads and generates; they're the fundamental unit models use to measure input, output, context limits, and usually billing..

Tool Use

Tool use is the ability of an AI model to invoke external functions, APIs, or systems — like looking up an order or issuing a refund — instead of only generating text, so it can act on real data rather than just describe it..

Transformer Model

A transformer model is a neural network architecture that processes an entire sequence of text at once using an attention mechanism to weigh how much each word relates to every other word, and it's the foundation of nearly every modern large language model..

Vector Database

A vector database is a database built to store and search embeddings — high-dimensional vectors — by similarity, so you can quickly find the content whose meaning is closest to a query..

Virtual Agent

A virtual agent is an AI-powered system that interacts with customers on its own to answer questions and resolve requests, functioning as an automated stand-in for a human support agent..

Zero-Shot Learning

Zero-shot learning is when an AI model performs a task it was never explicitly trained or given examples for, relying instead on the general knowledge it learned during pre-training and a clear description of the task..

Abandonment Rate

Abandonment rate is the percentage of customers who leave a support queue — hanging up a call or closing a chat — before ever reaching an agent..

Agent Utilization

Agent utilization is the percentage of an agent's paid time spent on productive, support-related work — handling contacts and related tasks — versus total time on the clock..

Average Handle Time (AHT)

Average Handle Time (AHT) is the average total time an agent spends handling a single customer interaction, including talk or chat time, hold time, and after-contact work..

Average Resolution Time

Average Resolution Time is the average length of time it takes to fully resolve a support ticket, measured from when it's created to when it's marked solved..

Average Speed of Answer (ASA)

Average speed of answer (ASA) is the average time a customer waits in the queue before their contact is answered by an agent, measured from entering the queue to connection..

Backlog Rate

Backlog rate is a measure of how the volume of unresolved support tickets is changing — typically the ratio of open or incoming tickets to those being resolved over the same period..

Contact Rate

Contact rate is the number of support contacts a business receives relative to its customer base or transaction volume — a measure of how often customers need to reach out for help..

Containment Rate

Containment rate is the percentage of interactions that a bot or automated system handles to completion without escalating to a human agent..

Cost per Contact

Cost per contact is the fully loaded operating cost of a support operation divided by the number of customer contacts it handles, showing what each interaction costs to service..

Cost per Resolution

Cost per resolution is the total support cost divided by the number of fully resolved issues, capturing what it costs to actually solve a customer's problem rather than just handle a single interaction..

CSAT vs NPS

CSAT vs NPS is the comparison between two customer experience metrics: CSAT (Customer Satisfaction Score) measures satisfaction with a specific interaction, while NPS (Net Promoter Score) measures overall loyalty and willingness to recommend the brand..

Customer Churn Rate

Customer churn rate is the percentage of customers who stop doing business with a company over a given period — canceling, not renewing, or lapsing — relative to the total at the start of that period..

Customer Effort Score (CES)

Customer Effort Score (CES) is a metric that measures how much effort a customer had to expend to get their issue resolved or their request completed, usually from a short post-interaction survey..

Customer Health Score

A customer health score is a composite metric that combines signals like product usage, support activity, satisfaction, and engagement into a single rating that predicts whether a customer is likely to renew, grow, or churn..

Customer Lifetime Value (CLV)

Customer lifetime value (CLV) is the total net revenue a business can expect from a single customer across the entire length of their relationship..

Customer Retention Rate

Customer retention rate is the percentage of existing customers a company keeps over a given period, excluding any new customers acquired during that time..

Customer Satisfaction Score (CSAT)

Customer Satisfaction Score (CSAT) is a metric that measures how satisfied customers are with a specific interaction or experience, usually from a short post-contact survey..

Escalation Rate

Escalation rate is the percentage of support tickets that are escalated beyond the first agent or tier — to a specialist, manager, or higher support level — before being resolved..

First Contact Resolution (FCR)

First Contact Resolution (FCR) is the percentage of customer issues fully resolved in a single interaction, without any follow-up, callback, or reopened ticket..

First Response Time (FRT)

First Response Time (FRT) is the average time between when a customer submits a support request and when they receive the first human or meaningful reply..

Forecast Accuracy

Forecast accuracy measures how close a support team's predicted contact volume (or workload) was to the volume that actually arrived, expressed as a percentage that shows how reliable the forecast was..

Handle Time

Handle time is the total time an agent spends handling a single customer contact, including the active conversation plus any after-contact work like notes, tagging, or follow-up..

Net Promoter Score (NPS)

Net Promoter Score (NPS) is a customer loyalty metric that measures how likely customers are to recommend a company, product, or service to others, on a 0–10 scale..

Occupancy Rate

Occupancy rate is the percentage of an agent's logged-in, available time that is actually spent handling contacts — including talk, chat, and after-contact work — versus waiting for the next contact..

Positive Feedback Rate

Positive feedback rate is the percentage of customer feedback responses that are positive — for example, the share of CSAT surveys rated "good" or "satisfied" out of all rated responses..

Quality Assurance (QA) Score

A Quality Assurance (QA) score measures how well a support interaction met a team's quality standards, based on a scorecard that grades things like accuracy, tone, policy adherence, and resolution..

Reopen Rate

Reopen rate is the percentage of resolved or closed support tickets that are subsequently reopened, usually because the customer's issue wasn't actually fixed..

Resolution Rate

Resolution rate is the percentage of support tickets or conversations that are successfully resolved out of the total received in a given period..

Schedule Adherence

Schedule adherence is a workforce management metric that measures how closely agents follow their assigned schedules — being available for work (on calls, chats, or tickets) during the times they're scheduled to be..

Self-Service Rate

Self-service rate is the share of customer questions resolved through self-service resources — help center, FAQs, portals, or automation — without a customer needing to contact a support agent..

Service Level

Service level is the percentage of contacts answered within a defined target time, most often expressed as an "X/Y" goal — for example, answering 80% of calls or chats within 20 seconds..

Service Level Agreement (SLA)

A Service Level Agreement (SLA) is a defined commitment to respond to or resolve support requests within a set time, often varying by priority, channel, or customer tier..

Shrinkage

Shrinkage is a workforce management metric that captures the percentage of paid agent time not spent handling customer contacts — including breaks, meetings, training, absenteeism, and other non-productive activities..

SLA Breach

An SLA breach occurs when a support team fails to meet a target defined in a Service Level Agreement — for example, not sending a first response or resolving a ticket within the promised time window..

Ticket Aging

Ticket aging is a measure of how long open support tickets have been unresolved, usually grouped into buckets (for example 0–1 days, 1–3 days, 3–7 days, 7+ days) to show where tickets are getting stuck..

Ticket Backlog

Ticket backlog is the number of open, unresolved support tickets that have accumulated and are still waiting to be handled at a given point in time..

Ticket Deflection

Ticket deflection is the share of customer questions resolved through self-service or automation before they ever become a human-handled support ticket..

Ticket Volume

Ticket volume is the total number of support requests received over a given period, across all channels a team handles..

Voice of the Customer (VoC)

Voice of the Customer (VoC) is the practice of systematically capturing and analyzing customer feedback — across surveys, support tickets, reviews, and conversations — to understand what customers need, expect, and experience..

Wait Time

Wait time is how long a customer waits before their contact is answered by an agent — the time spent in a queue, on hold, or waiting for a reply after reaching out..

Agent Collision Detection

Agent collision detection is a help-desk feature that warns support agents when someone else is already viewing or replying to the same ticket, preventing duplicate or conflicting responses..

Agent Workspace

An agent workspace is the unified interface where a support agent handles their work — tickets, live chats, customer context, and channels — in one screen, so they don't have to switch between separate tools to respond..

Answer Bot

An answer bot is an automated assistant that suggests or delivers relevant help-center articles and answers to customer questions, aiming to resolve common queries through self-service before a human agent gets involved..

Audit Log

An audit log is a chronological, tamper-resistant record of who did what and when in a system — capturing actions like logins, permission changes, ticket edits, and setting changes for security and compliance..

Business Hours

Business hours are the working schedule a help desk defines — days, times, and time zone — that determines when SLA timers count and when the team is expected to respond..

Business Rules

Business rules are configurable logic in a help desk that automatically govern how tickets behave — controlling routing, prioritization, SLAs, field values, and notifications — so support runs consistently without manual intervention..

Canned Response

A canned response is a pre-written, reusable reply that agents insert into a ticket or chat to answer common questions consistently and quickly, without retyping the same message each time..

Community Forum

A community forum is an online space where customers ask questions, share solutions, and help each other, with company staff moderating and contributing — turning peer knowledge into a searchable, self-service support resource..

CSAT Survey

A CSAT survey is a short, usually single-question survey sent after a support interaction to measure how satisfied the customer was, most often asking them to rate the experience on a simple scale..

Custom Fields

Custom fields are configurable data fields you add to help-desk tickets, users, or organizations to capture information beyond the built-in defaults — such as order number, product version, subscription tier, or affected region..

Escalation

Escalation is the process of moving a support ticket to a higher tier, a specialist, or a manager when the current agent can't resolve it or when it risks breaching an SLA..

Help Center

A help center is the customer-facing website where a company publishes its knowledge base, FAQs, and support options, giving customers a single place to find answers and, when needed, contact support..

Help Desk

A help desk is a tool (and the team behind it) that receives, organizes, and resolves customer or user support requests — turning emails, chats, and messages into trackable tickets that agents work through to resolution..

Help Desk vs Service Desk

A help desk is a tactical tool focused on resolving individual support requests and incidents, while a service desk is a broader ITSM function that manages the full lifecycle of services — including requests, incidents, problems, and changes — aligned to frameworks like ITIL..

Helpdesk API

A helpdesk API is the programmatic interface a support platform exposes so external systems can read and write data — creating tickets, updating status, pulling reports, or syncing users — without going through the agent UI..

Incident Management

Incident management is the process of detecting, logging, and resolving unplanned disruptions to a service as quickly as possible, with the goal of restoring normal operation and minimizing impact on users..

IT Service Management (ITSM)

IT service management (ITSM) is the set of practices, processes, and tools for designing, delivering, and supporting IT services — covering how an IT team manages requests, incidents, changes, and assets throughout their lifecycle..

Knowledge Base

A knowledge base is a structured, searchable library of articles — how-tos, FAQs, troubleshooting guides, and policies — that lets customers or agents find answers without contacting support directly..

Live Chat

Live chat is a support channel that lets customers message a company in real time — usually through a widget on a website or app — and get immediate responses from an agent or AI, without waiting for email or phone queues..

Macro (Helpdesk)

A macro is a saved, reusable set of actions in a help desk — most often a canned reply plus updates like tags, status, or assignee — that an agent can apply to a ticket in one click..

Messaging Widget

A messaging widget is the embeddable chat interface — usually a floating button in the corner of a website or app — that lets customers start a conversation with support, a bot, or an AI agent without leaving the page..

Multichannel Support

Multichannel support is offering customer service across multiple channels — such as email, phone, chat, and social media — where each channel is available but managed separately rather than unified into a single connected experience..

Omnichannel Support

Omnichannel support is a customer-service approach where all channels — email, chat, phone, social, messaging — are integrated into a single connected experience, so context and history follow the customer seamlessly across every channel..

Priority Matrix

A priority matrix is a grid that sets a ticket's priority from two inputs — impact (how much the issue affects the customer or business) and urgency (how quickly it needs resolving) — so priority is assigned consistently rather than by gut feel..

Proactive Support

Proactive support is reaching out to customers before they contact you — anticipating issues, questions, or opportunities and addressing them ahead of time, rather than only reacting to inbound tickets..

Round-Robin Assignment

Round-robin assignment is a ticket-distribution method that hands incoming tickets to available agents in rotation, one after another, so the workload spreads evenly across the team..

Self-Service Portal

A self-service portal is a customer-facing hub where people can resolve issues on their own — searching a knowledge base, checking ticket status, submitting requests, or managing their account — without contacting a support agent..

Service Desk

A service desk is the single point of contact between an IT organization and its users, managing not just support requests but also incidents, service requests, changes, and other IT service management (ITSM) processes..

Shared Inbox

A shared inbox is a single email address (like support@ or help@) that a whole team can access and manage together, with visibility into who's handling which message so replies aren't missed or duplicated..

Side Conversation

A side conversation is a separate thread branched off a main support ticket — with a different set of participants, such as an internal team, a vendor, or another department — used to gather information without cluttering the customer-facing conversation..

Single Sign-On (SSO)

Single sign-on (SSO) is an authentication method that lets users log in once with one set of credentials and access multiple applications — including a help desk — without signing in separately to each..

Skills-Based Routing

Skills-based routing is a ticket-assignment method that matches each incoming ticket to an agent with the right skills — such as language, product expertise, or seniority — so the best-equipped person handles it..

SLA Policy

An SLA policy is the configured set of rules inside a help desk that assigns service-level targets — like first-response and resolution times — to tickets based on conditions such as priority, channel, or customer tier..

Support Ticket

A support ticket is a single trackable record of a customer's request or issue, holding the full conversation and metadata — status, priority, assignee, tags — from when it's opened until it's resolved..

Tags (Helpdesk)

Tags are short keyword labels attached to support tickets to categorize them by topic, product, issue type, or any custom dimension — making tickets easy to filter, report on, and trigger automations from..

Ticket Lifecycle

The ticket lifecycle is the full sequence of stages a support ticket moves through from creation to closure — typically created, assigned, in progress, resolved, and closed..

Ticket Merge

Ticket merge is the action of combining two or more support tickets about the same issue into a single ticket, consolidating their conversations so the request is handled in one place..

Ticket Priority

Ticket priority is a field that ranks how urgently a support ticket needs attention — commonly low, normal, high, and urgent — so agents and automations can work the most important issues first..

Ticket Routing

Ticket routing is the process of directing incoming support tickets to the right agent, team, or queue based on rules like topic, channel, language, priority, or customer tier..

Ticket Status

Ticket status is the field on a support ticket that shows where it is in the resolution process — common values are new, open, pending, on-hold, solved, and closed..

Ticketing System

A ticketing system is software that captures customer or employee requests as trackable records called tickets, then routes, prioritizes, and manages them through to resolution so nothing gets lost..

Tiered Support (L1/L2/L3)

Tiered support is a model that organizes a support team into escalating levels — commonly L1, L2, and L3 — where frontline agents handle simple, high-volume issues and progressively more specialized staff take on harder problems..

Trigger (Helpdesk Automation)

A trigger is a helpdesk automation rule that runs the moment a ticket is created or updated: when its conditions are met, it automatically performs actions like assigning, tagging, notifying, or replying..

Webhook

A webhook is an automated HTTP callback that sends real-time data from one system to another when a specific event happens — for example, a help desk posting ticket details to an external URL the moment a ticket is created or updated..

Agent Evaluation

Agent Evaluation is Macha's built-in way to score how well an agent actually did its job.

Agent Orchestration

Agent orchestration is coordinating multiple Macha agents — and the tools each can call — so a complex request is broken into the right steps.

Agent Templates

Agent templates are pre-built Macha agent configurations — name, instructions, recommended connectors, and pre-selected tools — that spin up a working agent for a common workflow in one click..

AI Action Credits

Credits are Macha's usage currency: one deduction per complete AI response an agent produces, priced by the model that agent runs.

AI Agent Builder

The AI Agent Builder is Macha's Build-with-AI flow: describe the agent you want in plain language and it configures the name, instructions, tools, triggers, and even sub-agents based on the integrations you've connected..

AI Resolution

AI resolution is when a Macha agent actually resolves a request — reading data, taking an action through its tools, and closing the loop — rather than just answering a question.

Confirmation Gates

A confirmation gate is the pause Macha puts in front of any write action in interactive chat: before the agent sends a reply, updates a ticket, or issues a refund, it shows you exactly what it wants to do and waits for you to Confirm or Cancel..

Connectors

Connectors are Macha's built-in integrations.

Custom Tools

In Macha, a custom tool lets an AI agent call any HTTP API endpoint you configure — so an agent can read live data or take an action in a system Macha doesn't have a built-in connector for, without waiting for a marketplace app..

Data Sources

Data sources (also called knowledge sources) are the knowledge an agent reads from — uploaded documents, indexed websites, and live content from connected apps like Google Docs, Notion, and Confluence.

Embedded Chatbot

An embedded chatbot is a Macha agent surfaced as a chat widget you embed — so the same agent you built and tested inside Macha can answer directly wherever you place it, with those conversations flowing back into history and evaluation..

Human Handoff

Human handoff is Macha's principle of keeping a person in the loop on anything customer-facing — most concretely through confirmation gates that pause an agent before any write action in interactive chat, and escalation patterns that route hard cases to your team..

Plain-English Agent Configuration

In Macha, you configure an agent by writing its rules in plain English in its instructions field — how to triage, what tone to use, what to escalate, what's off-limits — rather than coding logic.

Scheduled Agents

A scheduled agent is a Macha agent attached to a scheduled (cron-style) trigger, so it runs itself on a recurring interval — daily digests, periodic audits, routine checks — with no event or human needed to start it..

Sidekick

Sidekick is the AI copilot built into the Macha dashboard that works on Macha itself — building agents, wiring up custom tools, auditing and testing how your agents perform, answering docs questions, and surfacing workspace status.

Studies (AI Analysis)

A Study runs an AI analysis across a list of records and returns a spreadsheet-shaped grid of results — one row per record, one column per field you define.

Sub-Agents

In Macha, a sub-agent is an agent that a parent agent can delegate to.

Triggers (Macha)

In Macha, a trigger is the event that wakes an agent up without anyone typing — a new ticket, a Slack mention, an inbound webhook, or a schedule.