Glossary
AI Agent & CX Glossary
Master the language of enterprise AI agents. 100+ definitions covering AI SLAs, autonomous resolution, compliance, deployment, and CX automation.
#
A
- ActionsActions let a Maven AI agent read and write data in business systems during a conversation, like order lookups, refunds, and case creation.
- Agent AssistAgent assist provides real-time suggestions, knowledge retrieval, and auto-drafted responses to human agents during live interactions. Agent assist reduces handle time while improving response accuracy and consistency.
- Agent ChannelsAgent Channels are where a Maven AI agent meets customers and employees: chat, SMS, WhatsApp, email, voice, the help desk, Slack, and Teams.
- Agent DesignerAgent Designer is Maven AGI's no-code workspace for configuring, testing, analyzing, and improving enterprise AI agents in one place.
- Agent MavenAgent Maven is Maven AGI's enterprise AI agent. It understands intent, reasons through each request, and takes secure actions across chat, email, and voice.
- Agent ProductivityAgent productivity measures how efficiently human support agents resolve customer issues. AI copilots boost productivity by auto-drafting responses, surfacing relevant knowledge, and handling routine tasks automatically.
- Agentic AIAgentic AI takes independent action to achieve goals rather than simply generating text responses. In customer service, agentic AI resolves issues end-to-end by accessing systems, executing workflows, and verifying outcomes.
- Agentic AI for Customer SupportAgentic AI for Customer Support represents autonomous artificial intelligence systems that perceive customer intent, reason through complex scenarios, and execute complete end-to-end resolutions across integrated enterprise systems without human intervention.
- Agentic DeterminismAgentic Determinism means the platform decides which data, actions, and limits apply before a Maven AI agent reasons, not after it responds.
- Agentic WorkflowAgentic workflows let AI agents plan, execute, and verify multi-step tasks independently. An agentic workflow might check order status, process a refund, and send confirmation without human intervention.
- AI Accuracy RateAI accuracy rate measures the percentage of AI responses that are factually correct and helpful. Enterprise AI platforms target 85-95% accuracy through grounding, guardrails, and continuous quality monitoring.
- AI AgentAn AI agent perceives context, reasons over information, and takes autonomous action to achieve goals. Unlike chatbots that follow scripts, AI agents dynamically solve problems using tools and knowledge.
- AI Agent OrchestrationAI agent orchestration is the coordinated management of multiple specialized AI agents working together on complex workflows, enabling enterprises to handle multi-step customer service processes that exceed single-agent capabilities.
- AI Audit TrailAI audit trails record every decision, action, and data access made by AI agents for compliance and accountability. Regulated industries require audit trails to prove AI operates within approved parameters.
- AI Chatbot vs AI AgentAI chatbots handle simple questions with pre-defined responses while AI agents resolve complex issues by reasoning and taking actions. Agents access backend systems, execute workflows, and maintain full context.
- AI Confidence ScoreAI confidence scores quantify how certain a model is about its response, typically on a 0-1 scale. Support teams use confidence thresholds to trigger human escalation when AI certainty drops below set levels.
- AI CopilotAn AI copilot works alongside human agents, suggesting responses, surfacing relevant knowledge, and automating repetitive tasks in real time. Copilots help agents respond faster without replacing human judgment.
- AI Customer Experience (CX)AI CX applies artificial intelligence to every customer touchpoint, from first contact through ongoing support. AI-driven CX improves satisfaction and loyalty by reducing wait times and personalizing interactions.
- AI Customer ServiceAI customer service uses artificial intelligence to understand, respond to, and resolve customer inquiries autonomously. AI handles routine issues instantly while routing complex cases to human agents with full context.
- AI Deployment TimelineAI deployment timelines range from 2 weeks for basic setups to 3-6 months for enterprise-grade AI agents. Timeline depends on integrations, compliance requirements, and knowledge base complexity.
- AI Email ResponseAI email response reads, classifies, and replies to customer emails automatically. AI processes incoming emails in seconds rather than hours, handling routine requests while flagging complex issues for human review.
- AI EmbeddingsAI embeddings convert text into numerical vectors that capture semantic meaning. Embedding quality directly determines how well AI retrieves relevant knowledge base content for customer queries.
- AI EscalationAI escalation transfers conversations from AI to human agents when issues require empathy, judgment, or authorization. Effective escalation passes full conversation context so customers never repeat themselves.
- AI for Banking SupportAI for banking automates account inquiries, fraud alerts, and dispute resolution while maintaining regulatory compliance. Banks use AI support to reduce handle time and cost per interaction significantly.
- AI for E-commerce SupportAI for e-commerce automates order tracking, return processing, and product inquiries at scale. Online retailers use AI agents to handle high-volume, repetitive support during sales events and holiday seasons.
- AI for Education SupportAI for education handles enrollment questions, billing inquiries, and IT support for students and parents at scale. Education providers use AI to manage seasonal enrollment spikes without temporary staffing.
- AI for FinTech SupportAI for FinTech handles transaction inquiries, account management, and fraud alerts while maintaining regulatory compliance. FinTech companies use AI to deliver instant support without compromising data security.
- AI for Healthcare SupportAI for healthcare delivers instant patient support while maintaining HIPAA compliance. Healthcare organizations use AI agents for appointment scheduling, billing inquiries, and care navigation.
- AI for Insurance SupportAI for insurance automates claims intake, coverage questions, and policy management while meeting compliance standards. Insurers use AI to reduce claims processing time and improve policyholder experience.
- AI for Marketplace SupportAI for marketplaces handles buyer and seller interactions simultaneously while enforcing platform policies. Two-sided support requires AI that understands both perspectives of every transaction.
- AI for Real Estate SupportAI for real estate handles property inquiries, lease management, and tenant communications across portfolios. PropTech companies use AI to provide instant responses to high-volume property questions.
- AI for SaaS SupportAI for SaaS handles product questions, troubleshooting, and account management using product documentation as its knowledge base. SaaS companies use AI to scale support alongside user growth without proportional headcount.
- AI for Telecom SupportAI for telecom automates billing disputes, plan modifications, and device troubleshooting at scale. Telecom providers handle millions of routine inquiries monthly, making AI automation especially impactful.
- AI for Travel SupportAI for travel manages reservations, cancellations, and disruption recovery in real time. Travel companies use AI to handle demand spikes during weather events and peak booking seasons.
- AI GovernanceAI governance ensures AI systems operate within ethical, legal, and organizational boundaries. Frameworks like ISO 42001 provide structured approaches to managing AI risk in enterprise environments.
- AI GroundingAI grounding transforms generic language models into reliable, enterprise-ready systems by anchoring their responses to your organization's verified data sources and eliminating the risk of AI hallucinations.
- AI GuardrailsAI guardrails enforce boundaries on what AI agents can say and do in customer interactions. Guardrails prevent off-topic responses, unauthorized actions, and brand-damaging outputs in production.
- AI HallucinationAI hallucination occurs when models produce confident but factually incorrect responses. Enterprise AI platforms prevent hallucination through grounding, retrieval augmented generation, and confidence scoring.
- AI Help DeskAn AI help desk uses artificial intelligence to resolve support tickets, categorize issues, and route complex cases. AI help desks reduce ticket backlog by automatically handling repetitive, well-documented inquiries.
- AI Live ChatAI live chat delivers real-time customer support through messaging interfaces on websites and apps. AI-powered chat provides instant responses while maintaining natural, conversational tone and accuracy.
- AI ObservabilityAI observability provides real-time visibility into how AI agents perform, fail, and improve over time. Observable AI systems surface accuracy metrics, latency, error patterns, and satisfaction signals continuously.
- AI OrchestrationAI orchestration coordinates multiple AI agents, models, and systems to execute complex workflows seamlessly with reliable task routing and governance.
- AI PersonalizationAI personalization adapts support interactions based on customer history, preferences, and context. Personalized AI delivers more relevant answers and faster resolution by understanding each customer's situation.
- AI Pilot ProgramAn AI pilot program validates AI customer service performance at limited scale before enterprise deployment. Well-designed pilots run 4-8 weeks with clear success criteria and measurable KPIs.
- AI Proof of Concept (POC)An AI POC validates that a solution works with your specific data, workflows, and customer base. A good POC runs 2-4 weeks with predefined success metrics and realistic support scenarios.
- AI ROI (Return on Investment)AI ROI measures the financial return of deploying AI in customer service against total investment. Key ROI drivers include reduced cost per ticket, faster resolution, and improved customer retention.
- AI Service Level Agreement (SLA)AI SLAs define uptime, accuracy, and resolution commitments for AI customer service platforms. Setting the right SLA benchmarks protects your organization from underperforming AI deployments.
- AI Ticket RoutingAI ticket routing uses machine learning to classify and assign support tickets to the right team or agent. Intelligent routing considers issue type, customer tier, agent expertise, and current workload.
- AI Training DataAI training data teaches models how to handle customer interactions accurately. Data quality is the single biggest factor in AI performance, directly impacting accuracy and response relevance.
- AI Vendor EvaluationAI vendor evaluation assesses platforms on accuracy, security, integrations, and total cost of ownership. A structured evaluation framework prevents costly mistakes when selecting AI customer service technology.
- AI vs Human SupportEffective customer service combines AI automation for routine issues with human agents for complex situations. The key is designing smart handoffs so each interaction reaches the right resource at the right time.
- AI Workflow AutomationAI workflow automation combines artificial intelligence with business process orchestration to create adaptive, self-improving systems that handle complex, multi-step customer service workflows with intelligent decision-making capabilities.
- API Integration (Customer Service)API integration connects AI agents to CRMs, helpdesks, and business systems so they can take real actions. Integration depth determines whether AI can actually resolve issues or only provide information.
- App StudioApp Studio is Maven AGI's developer platform for building, testing, and deploying custom Maven Apps with the Maven TypeScript SDK.
- Asynchronous MessagingHow async messaging works for support: reply on your own time, persistent threads, and sync vs live chat—plus why AI CX teams care. Maven AGI glossary.
- Autonomous ResolutionAutonomous resolution means AI fully resolves a customer issue without human intervention. Unlike deflection, autonomous resolution actually solves the problem rather than redirecting the customer elsewhere.
- Average Handle Time (AHT)AHT measures the average time spent on each customer support interaction from start to resolution. AI reduces AHT by giving agents instant knowledge access and automating repetitive steps within conversations.
- Average Wait TimeAverage wait time measures how long customers spend waiting before connecting with support, serving as a critical gauge of service responsiveness and operational efficiency.
B
- BPO (Business Process Outsourcing)BPO is the practice of contracting customer support operations to external service providers. AI is transforming BPO economics by automating routine interactions while BPO agents handle complex escalations.
- Build vs Buy (AI Agents)The build-or-buy decision for AI agents depends on your team's ML expertise, timeline, and maintenance capacity. Most enterprises find buying accelerates time-to-value while building increases long-term control.
C
- CapabilitiesCapabilities is where Maven teams build, test, and manage what their AI agent can do: Actions, Triggers, and Intelligent Fields.
- Chain of Thought ReasoningChain of thought reasoning lets AI break complex problems into logical steps before answering. This approach improves accuracy on multi-step customer issues like billing disputes and account changes.
- Charter TemplatesCharter Templates are pre-built starting points for common Maven Charters, like brand foundations, escalation handling, and voice behavior.
- ChartersCharters are natural-language playbooks that tell a Maven AI agent how to behave, what it can use, and what it can do in each customer situation.
- Chatbot vs Virtual AgentChatbots handle basic questions with scripted responses while virtual agents resolve complex issues autonomously. Virtual agents access backend systems, take actions, and maintain context across long conversations.
- Contact Center AIContact center AI automates voice and digital customer interactions within call center environments. AI handles routine calls and chats while routing complex issues to human agents with full context.
- Contact Center as a Service (CCaaS)CCaaS delivers contact center technology through the cloud, replacing on-premise infrastructure. Modern CCaaS platforms integrate AI agents for automated resolution alongside traditional human agent tools.
- Context EngineeringContext engineering is the systematic practice of designing and optimizing how AI agents access, organize, and utilize information from multiple enterprise so
- Context WindowThe context window is the maximum amount of text an AI model can consider in a single interaction. Larger context windows let AI reference more conversation history and knowledge during customer conversations.
- Conversational AIConversational AI enables multi-turn, natural language interactions between customers and AI systems. Unlike scripted chatbots, conversational AI maintains context across a full conversation and adapts to user intent.
- Conversational IVRConversational IVR replaces button-press phone menus with natural language understanding. Callers speak their request instead of navigating numbered options, reducing call time and abandonment rates.
- Cost Per TicketCost per ticket measures the average expense of resolving a single support interaction, including labor, technology, and overhead. AI reduces cost per ticket by automating routine issues that previously required human agents.
- Customer Churn PreventionCustomer churn prevention uses AI to identify and re-engage at-risk customers before they leave. AI detects dissatisfaction signals in support interactions and triggers proactive outreach to retain accounts.
- Customer Effort Score (CES)CES measures the effort customers must exert to get their issue resolved on a scale of 1-7. Lower effort scores correlate with customer loyalty, making CES a leading indicator of retention.
- Customer Journey MappingCustomer journey mapping traces every interaction a customer has with your support organization. Journey maps reveal where AI agents can create the greatest impact on customer experience and efficiency.
- Customer Onboarding SupportCustomer onboarding support guides new users through product setup, configuration, and first-time tasks. AI-powered onboarding delivers personalized guidance at scale without requiring additional headcount.
- Customer Retention RateCustomer retention rate measures how well a business keeps its existing customers over a specific time period, providing crucial insight into customer loyalty, product satisfaction, and the long-term health of customer relationships.
- Customer Satisfaction Score (CSAT)CSAT measures how satisfied customers are with a specific support interaction, typically on a 1-5 scale. Consistent AI-powered support helps maintain high CSAT by delivering reliable quality across every interaction.
- Customer Service AutomationCustomer service automation uses AI to handle support requests without manual intervention. Modern automation resolves issues end-to-end rather than simply deflecting customers to static FAQ pages.
- Customer Support AICustomer support AI uses artificial intelligence to resolve requests, assist human agents, and surface operational insights. Modern support AI takes real actions within business systems, going well beyond chatbot capabilities.
D
- Data Residency for AIData residency determines the geographic location where AI systems store and process customer data. GDPR, CCPA, and industry-specific regulations often mandate specific data residency requirements.
- Deflection RateDeflection rate measures the percentage of support requests handled through self-service before reaching a human agent. High deflection alone doesn't indicate whether customer issues were actually resolved.
E
- Enterprise AI SupportEnterprise AI support meets the security, compliance, and scale requirements of large organizations. Enterprise-grade platforms include SOC 2, HIPAA, and PCI certifications alongside SSO and role-based access.
- Escalation ManagementEscalation management governs when and how AI transfers conversations to human agents. Effective escalation preserves full conversation context so customers never need to repeat themselves after handoff.
F
- Fallback BehaviorFallback behavior defines how AI responds when it lacks confidence in an answer. Well-designed fallback strategies route to humans gracefully rather than providing inaccurate or fabricated responses.
- Fallback IntentA fallback intent is a safety mechanism in conversational AI that activates when customer inputs don't match any predefined response categories, ensuring graceful handling of unexpected queries rather than providing incorrect information or conversation breakdown.
- Fine-Tuning (for Customer Service)Fine-tuning adapts pre-trained AI models to your specific customer service domain and brand voice. Unlike prompt engineering alone, fine-tuning embeds domain knowledge directly into model weights.
- First Contact ResolutionFCR measures the percentage of customer issues resolved in a single interaction without follow-up contact. High FCR rates correlate directly with customer satisfaction and lower overall support costs.
G
- Generative AI for SupportGenerative AI uses large language models to create helpful, natural responses to customer questions. Unlike rule-based systems, generative AI understands nuance and generates contextually appropriate answers.
- Graph of RecordGraph of Record connects Maven AI agents to the systems you already run and maps how their data relates, so answers come from current source-of-truth data.
- Grounding (in AI)Grounding constrains AI responses to verified source material rather than generated content. Grounded AI cites specific knowledge base articles, dramatically reducing hallucination in customer interactions.
H
- HIPAA Compliance for AIHIPAA compliance for AI requires protecting patient health information (PHI) throughout AI processing. AI platforms handling healthcare data must implement encryption, access controls, and BAAs.
- Human-in-the-LoopHuman-in-the-loop (HITL) keeps human agents involved in AI workflows for oversight, correction, and approval. HITL is essential for high-stakes interactions where errors carry financial or compliance risk.
I
- Intelligent FieldsIntelligent Fields turn conversations into structured data. The Maven agent fills in fields like intent, sentiment, or churn risk from plain-English prompts.
- Intelligent Virtual Assistant (IVA)IVAs go beyond basic chatbots by understanding context, managing multi-turn conversations, and taking real actions. IVAs integrate with backend systems to resolve issues rather than just providing information.
- Intent RecognitionIntent recognition classifies what a customer is trying to accomplish from their message or query. Accurate intent detection determines whether a request needs billing help, technical support, or account changes.
- Intercom AlternativeIntercom alternatives offer different approaches to customer support, from AI-first resolution to traditional messaging. AI-native platforms focus on resolving issues autonomously rather than routing to human agents.
- IVR (Interactive Voice Response)IVR is phone system technology that routes callers through pre-recorded menu options using keypad or voice inputs. While still widely deployed, IVR is increasingly being replaced by conversational voice AI.
- IVR ReplacementIVR replacement swaps rigid phone tree menus for AI voice agents that understand natural speech. Organizations replacing IVR typically see lower abandonment rates and faster average call resolution.
K
- Knowledge Base AIKnowledge base AI delivers precise answers from help articles, documentation, and internal wikis using AI. AI-powered knowledge bases understand questions semantically rather than relying on keyword matching.
- Knowledge GraphKnowledge graphs organize information into interconnected entities and relationships that AI can reason over. In customer service, knowledge graphs help AI understand how products, policies, and processes relate.
- Knowledge GroundingKnowledge grounding is the process of anchoring AI responses to verifiable, real-world data sources, ensuring every answer traces back to specific enterprise documents, databases, or systems rather than relying on potentially outdated training data.
- Knowledge InboxMaven's Inbox flags knowledge issues like duplicates, conflicts, outdated content, and gaps, with suggested fixes for teams to review.
- Knowledge IngestionKnowledge ingestion is the process of importing, parsing, and structuring content so AI agents can use it. Effective ingestion handles PDFs, help centers, Confluence wikis, and internal docs automatically.
L
M
- MagiMagi is the AI assistant built into Agent Designer. Ask why your agent did something or how it’s performing, and get a grounded answer on the page you’re on.
- ManualsA Manual is the natural-language instruction text inside a Maven Charter that tells the AI agent how to behave when that Charter applies.
- Maven AppsMaven Apps connect the Maven AGI platform to external systems as knowledge sources, actions and triggers, and customer-facing channels.
- Maven CommerceMaven Commerce lets a Maven AI agent sell products and services within guardrails on what it can offer, at what price, and to whom.
- Maven CopilotMaven Copilot is an AI assistant inside your help desk that summarizes tickets, drafts cited replies, and answers questions for human agents.
- Maven VoiceMaven Voice is an enterprise AI voice agent that resolves support calls in real time and completes workflows like refunds and account updates.
- Model Context Protocol (MCP)MCP is an open protocol that lets AI models connect to external tools, databases, and APIs in a standardized way. MCP enables AI agents to take real actions like issuing refunds or updating accounts.
- Multi-Agent SystemsMulti-agent systems use multiple specialized AI agents that collaborate to handle complex problems. Each agent manages a specific domain while an orchestrator coordinates handoffs between billing, technical, and policy tasks.
- Multilingual Customer SupportMultilingual support uses AI to serve customers in their preferred language without hiring native speakers. Modern AI agents handle 100+ languages with near-native fluency for common support scenarios.
N
- Natural Language Processing (NLP)NLP enables AI to parse, interpret, and respond to human language in customer interactions. Modern NLP powers intent classification, entity extraction, and sentiment analysis in customer service automation.
- Net Promoter Score (NPS)NPS gauges customer loyalty by asking how likely they are to recommend your company on a 0-10 scale. Support quality is one of the strongest drivers of NPS, directly impacting promoter and detractor ratios.
O
P
- PCI-DSS for Customer Service AIPCI-DSS compliance ensures AI systems handling payment card data meet strict security standards. AI agents processing transactions must isolate cardholder data and maintain Level 1 certification.
- PII RedactionPII redaction automatically removes or masks personally identifiable information from AI interactions. Effective redaction prevents sensitive data like SSNs, credit cards, and health info from being stored.
- Proactive Customer SupportProactive support identifies and resolves issues before customers reach out. AI enables proactive support by detecting patterns in usage data, error logs, and customer sentiment signals.
- Prompt EngineeringPrompt engineering is the practice of crafting precise instructions that guide AI model behavior and output. In customer service, prompt design determines AI tone, accuracy boundaries, and escalation triggers.
- Prompt InjectionPrompt injection is an attack where users manipulate AI through crafted inputs to bypass safety controls. Customer-facing AI must implement input validation and guardrails to prevent prompt injection exploits.
Q
- Quality Assurance (QA) for SupportQA for support evaluates interaction quality across both human and AI agents. Modern QA programs must assess AI accuracy, tone, and compliance alongside traditional human agent performance metrics.
- Queue ManagementQueue management is the strategic orchestration of customer flow and wait times across service touchpoints, optimizing how customers move through support interactions to reduce abandonment and maximize satisfaction.
R
- RAG (Retrieval Augmented Generation)RAG combines information retrieval with AI text generation to produce accurate, source-backed responses. RAG prevents hallucination by requiring AI to cite actual knowledge base content in every answer it generates.
- Real-Time Voice AIReal-time voice AI enables AI agents to hold natural phone conversations with customers at human-like speed. Sub-500ms latency is critical for voice AI to feel natural and conversational.
- Resolution RateResolution rate tracks the percentage of customer issues fully resolved, not just responded to or deflected. Resolution rate measures actual outcomes, making it a more meaningful metric than deflection rate.
- Responsible AIResponsible AI is the comprehensive framework of policies, technical controls, and governance mechanisms that ensure AI systems operate safely, transparently, and ethically while maintaining regulatory compliance and customer trust.
- Rule-Based vs AI ChatbotRule-based chatbots follow predefined decision trees while AI chatbots use natural language understanding. Rule-based bots work for simple FAQs but fail when customers phrase questions unexpectedly.
S
- SegmentsSegments are reusable definitions of users or conversations that decide which knowledge, actions, and Charters a Maven AI agent can use.
- Self-RAGSelf-RAG (Self-Reflective Retrieval-Augmented Generation) is an advanced AI framework that enables language models to autonomously decide when to retrieve external information and continuously evaluate their own outputs for accuracy and relevance.
- Self-Service PortalSelf-service portals give customers access to knowledge bases, account tools, and AI assistance on demand. Effective self-service handles the majority of routine inquiries, freeing human agents for complex issues.
- Semantic SearchSemantic search understands the intent behind queries rather than matching exact keywords. AI support platforms use semantic search to surface relevant answers even when customers phrase questions unexpectedly.
- Sentiment AnalysisSentiment analysis uses AI to identify emotional tone in customer messages, from frustration to satisfaction. Support teams use sentiment data to prioritize urgent cases and surface systemic experience issues.
- Service Level Agreement (SLA)A service level agreement (SLA) is a formal contract between a service provider and customer that defines specific performance standards, measurable metrics, and consequences for not meeting agreed-upon service levels.
- Show ReasoningShow Reasoning reveals the context, knowledge sources, and actions behind any Maven AI agent response, so teams can see why it answered as it did.
- SOC 2 Compliance for AISOC 2 certification validates that an AI platform meets rigorous security, availability, and confidentiality standards. Enterprise buyers typically require SOC 2 Type II reports before deploying AI in production.
- Speech-to-Text (ASR)ASR converts spoken language into text that AI systems can process and act on. Accurate speech recognition is the foundation of voice AI, with modern systems achieving over 95% word accuracy.
- Strict Return ChartersStrict Return Charters return a fixed, pre-approved response word for word, for answers like legal disclosures that must be exact every time.
- Support Automation RateSupport automation rate measures the percentage of customer issues resolved by AI without human involvement. This metric captures AI's true operational impact more accurately than deflection rate alone.
- Support ROISupport ROI quantifies the financial return of customer service investments including technology, staffing, and training. Calculating ROI requires tracking both direct cost savings and revenue retained through better service.
T
- TelephonyTelephony is the technology infrastructure enabling voice communication in customer service operations. Modern telephony includes SIP trunking, VoIP, and PSTN integration for AI voice agent deployment.
- Text-to-Speech (TTS)TTS converts written text into natural-sounding speech for AI voice agents. Modern neural TTS produces voices nearly indistinguishable from humans, supporting dozens of languages and regional accents.
- Ticket DeflectionTicket deflection measures support requests handled through self-service before reaching a human agent. While useful as a volume metric, deflection alone doesn't indicate whether customers actually got their issues resolved.
- Ticket VolumeTicket volume tracks the number of support requests an organization receives over a given period. AI handles volume spikes from product launches, outages, and seasonal events without requiring temporary staffing.
- Ticketing SystemA ticketing system tracks customer support requests from submission through resolution in a centralized workflow. Modern ticketing systems integrate with AI to auto-classify, route, and resolve tickets.
- Tiered Support ModelTiered support organizes customer service into escalation levels, from basic inquiries (L1) to complex issues (L3). AI now handles most L1 interactions, letting human agents focus on high-value work.
- Time to ResolutionTime to resolution measures the duration from when a customer reports an issue to when it's fully solved. AI reduces resolution time for routine issues by providing instant, accurate responses around the clock.
- Tool Use (Function Calling)Tool use enables AI agents to call external APIs, query databases, and execute workflows in real time. Function calling is what separates AI agents that resolve issues from chatbots that only provide information.
- Total Cost of Ownership (AI Support)TCO for AI support includes licensing, integration, training, maintenance, and ongoing optimization costs. Many organizations underestimate TCO by focusing on subscription fees while ignoring implementation and upkeep.
- Tree of Thoughts (ToT)Tree of Thoughts (ToT) is an advanced AI reasoning framework that enables large language models to explore multiple branching paths of thought simultaneously, evaluate options, and backtrack from dead ends to solve complex problems more effectively.
- TriggersTriggers are Maven automations that run on events or a schedule, like syncing conversation data to a warehouse, without a customer asking.
V
- Vector EmbeddingVector embeddings transform unstructured data—text, images, audio, and customer interactions—into numerical arrays that capture semantic meaning and relationships, enabling AI systems to understand and process information with human-like comprehension.
- Voice AIVoice AI enables AI agents to handle live phone calls with natural speech understanding and generation. Modern voice AI achieves sub-second response latency, making phone interactions feel genuinely conversational.
- Voice BotVoice bots handle customer phone calls using speech recognition and natural language understanding. Unlike rigid IVR menus, voice bots engage in natural conversation to resolve issues over the phone.