The AI in GTM glossary

Clear definitions and practical examples for the language of AI, sales and revenue operations.

157 terms · 10 categoriesReviewed by Evergrowth ·
157 terms. Search by name, acronym or concept.

Definitions explain the general concept. “In Evergrowth” notes describe the product's application. Maintained by Evergrowth through reviews of product pages and documentation. See the current Evergrowth product reference.

Foundations

A large language model is a machine-learning model that processes and generates language based on patterns learned during training. Applications can supply instructions, retrieved information and tools to help it perform tasks. Model choice, task design and evidence all affect the result.

Example
A model drafts a meeting brief from supplied account research.

Read: AI and sales

In Evergrowth · Orchestration

Generative AI

BeginnerEveryone

Generative AI produces content such as text, images or audio. In GTM, it can draft a message or summarize research. Output quality depends on the model, evidence, instructions and review. Generative capability and agentic execution can be used together.

Example
A seller generates a first draft of an email and checks its claims before sending.

Read: AI and sales

In Evergrowth · Orchestration

Prompt

BeginnerEveryone

A prompt is input that guides an AI model, including instructions, a question or relevant context. GTM prompts may specify qualification criteria or a messaging task. Clear prompts help, but accurate evidence, suitable tools and output review also matter.

Example
A research instruction specifies the question, evidence required and how to report uncertainty.

Read: AI and sales

In Evergrowth · Orchestration

Hallucination

BeginnerEveryone

A hallucination is generated content that is inaccurate or unsupported while appearing plausible. In GTM, it can create a false company fact or an invented contact detail. Source grounding, explicit uncertainty and human review reduce risk; none guarantees error-free output.

Example
A generated brief invents a funding round that no cited source supports.

OpenAI: why language models hallucinate

In Evergrowth · Orchestration

Context Window

IntermediateRevOps & GTM Engineers

A context window is the maximum amount of input and output a model can handle for a request, measured in tokens. Workflow results and retrieved information may be supplied within that limit. Persistent CRM data or saved knowledge is not automatically present in every request.

Example
A long research report may need selective retrieval before a model can use it.

Read: AI and sales

In Evergrowth · Orchestration

Token

IntermediateGTM Engineers

A token is a unit used by a model to represent text or other input and output. Tokenization varies by model, language and content. Token usage can affect model cost and request capacity, while a product may bill customers using a separate credit model.

Example
A model counts the tokens in both the supplied research and its generated answer.

Read: AI and sales

In Evergrowth · Orchestration

RAG (Retrieval-Augmented Generation)

IntermediateGTM Engineers & RevOps

RAG combines retrieval from an external information source with generation using the retrieved material. It can help an answer use relevant documents or facts beyond model training, but the quality still depends on retrieval, evidence and the generated interpretation.

Example
An application retrieves relevant product documents before drafting an answer.

OpenAI retrieval guide

In Evergrowth · Orchestration

Grounding

IntermediateRevenue teams

Connecting an AI answer to relevant evidence or supplied context rather than relying only on model-generated assertions. Grounding improves traceability but does not guarantee that an answer is correct.

Example
A meeting brief cites account research and distinguishes sourced facts from suggestions.

Read: AI and sales

In Evergrowth · Orchestration

AI Architecture in GTM

Agentic GTM

IntermediateEveryone

Agentic GTM is an approach in which AI agents perform parts of go-to-market work, such as research, qualification and outreach preparation. People define objectives and remain responsible for decisions and customer relationships.

Example
Agents prepare account context while sellers review it and lead the conversations.

Read: The future of GTM software

In Evergrowth · Workflows

AI Assistant vs. AI Agent

IntermediateEveryone

An assistant helps a person through an interactive interface; an agent can carry out actions using tools toward a goal. One application can support both patterns. The useful distinction is its enabled tools, permissions, autonomy and review controls.

Example
An assistant may summarize research and also invoke a permitted tool to start another task.

OpenAI: ChatGPT agent

In Evergrowth · Workflows

Digital Colleague

IntermediateEveryone

Digital colleague is a metaphor for an AI system with a defined role in a team's workflow. It describes how people work with the system, rather than a separate technical category.

Example
A team treats its research agent as a specialist whose output sellers review.

Read: The future of GTM software

In Evergrowth · Workflows

GTM Intelligence

Persona Card

IntermediateRevOps

A persona card is a structured description of a buyer role for a defined segment, including responsibilities, problems, decision role and relevant messaging. It guides contact selection and communication; it does not establish an individual person's personality or purchasing authority.

Example
A team records the responsibilities and role criteria its Contact Finder should use.

Read: Qualification frameworks in practice

In Evergrowth · Agent Training Center

ICP Vertical

IntermediateRevOps

An ICP vertical is a target segment with its own company-level qualification criteria. It may be defined by industry, business model or use case. Segment membership and qualification fit should be distinguished from timing signals used to prioritize accounts.

Example
A team separates financial services and healthcare because each needs different qualification criteria.

Read: Qualification frameworks in practice

In Evergrowth · Agent Training Center

DISC Profile

IntermediateSales Reps

DISC is a framework commonly used to describe behavioral and communication styles. A style inferred from public professional material is a hypothesis, not the result of a validated assessment. Sales teams should use it cautiously and adapt to the person's actual communication.

Example
A seller treats a suggested communication style as a tentative aid rather than a verified personality assessment.

Read: Qualification frameworks in practice

In Evergrowth · Accounts and Contacts

Agent Training Center

IntermediateRevOps

The Agent Training Center is Evergrowth's configuration surface for shared GTM intelligence. Relevant agents use its inputs to perform their tasks; this configures context and instructions rather than fine-tuning a foundation model.

Example
RevOps publishes a value proposition and defines ICP verticals and persona cards.

Read: Qualification frameworks in practice

In Evergrowth
Evergrowth stores value proposition, ICP verticals and persona cards here. Agents read the inputs relevant to their task; configuration is not model fine-tuning.
Agent Training Center

Data & Enrichment

Waterfall Enrichment

IntermediateRevOps & GTM Engineers

Waterfall enrichment queries data providers in a configured sequence until a usable result is found or the search ends. It can improve coverage across providers, but results still require verification. Billing and stopping rules depend on the service and workflow.

Example
A lookup tries another provider when the previous provider cannot return a usable result.

Read: How contact data is changing

In Evergrowth · Email and phone waterfall

Contact Verification

BeginnerSales Ops

Contact verification checks selected aspects of contact information, such as email status, phone validity or current employment. These checks answer different questions and may become stale. A verified address does not guarantee inbox placement, and a valid number does not guarantee that the intended person answers.

Example
A team distinguishes a verified email result from a catch-all result before sending.

Read: How contact data is changing

In Evergrowth · Email and phone waterfall

Qualification & Scoring

Account Scoring

IntermediateRevOps

Account scoring assigns a numerical value to a company using defined criteria such as fit, activity or timing signals. The score's meaning depends on its inputs and calculation. A research signal score should not be interpreted as a probability of purchase or a qualification verdict.

Example
A team ranks qualified accounts by relevant research findings rather than treating the rank as a fit verdict.

Read: Qualification and list building

In Evergrowth · Account Qualification

Qualification Score

IntermediateRevOps

Qualification can produce a decision, a numerical score or both, depending on the system. In Evergrowth, account fit is a verdict that shows its reason; the research signal score is a separate measure of positive configured signals.

Example
An account has ICP status Yes and a separate Signal Score of 75.

Read: Qualification and list building

In Evergrowth
In Evergrowth, ICP and IBP fit are verdicts that show their reason. The separate Signal Score summarizes research findings.
Account Qualification

MEDDIC / MEDDPICC

IntermediateSales Reps & Leaders

MEDDIC and MEDDPICC are sales qualification frameworks for understanding an opportunity. MEDDIC covers Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain and Champion; MEDDPICC also covers Paper Process and Competition.

Example
A seller records the approval process and validates the champion's influence during discovery.

Read: Qualification and list building

Account Qualification

IntermediateRevOps

Account qualification evaluates whether a company meets defined company-level criteria before sales effort is invested. Research supplies the evidence behind the verdict. Account fit differs from whether the company currently has a reason or intention to buy.

Example
A company fails a required geography criterion and receives a No verdict with a reason.

Read: Luzmo qualification story

In Evergrowth · Account Qualification

Contact Qualification

IntermediateRevOps

Contact qualification evaluates whether a person matches the buyer role needed for a defined account or segment. It may check current employer, responsibility, seniority and decision role. Passing qualification does not confirm a buying decision or replace human relationship knowledge.

Example
A contact's current employer and role are checked against the target persona.

Read: Luzmo qualification story

In Evergrowth · Contact Qualification

Signal Score

IntermediateRevenue teams

A signal score summarizes relevant research findings for prioritizing accounts. Its formula depends on the system and should not be treated as a probability of buying.

Example
An account fits the ICP but has few positive timing signals, so fit and priority remain separate.

Read: how signal research works

In Evergrowth
In Evergrowth, Signal Score is the share of configured signal research with positive findings, shown from 0 to 100 for qualified accounts. It is separate from ICP fit.
Accounts and Contacts

Outreach & Personalization

Plays vs. Sequences

IntermediateSales Reps & RevOps

A play defines a purposeful sales action or approach for a situation. A sequence organizes outreach steps and timing, often across channels. A play can supply personalized messages or talk tracks to a sequence; the terms describe different aspects of execution.

Example
A seller places a researched play into a sequence with agreed timing and channels.

Read: A research-led calling playbook

In Evergrowth · Play Copywriting

Sales Engagement Platform (SEP)

BeginnerSales Reps & Ops

A sales engagement platform helps organize and execute outreach activities, sequences and follow-up. Capabilities vary, and some products also include AI research or writing. Compare the actual workflow, integration and governance needs rather than assuming fixed category boundaries.

Example
A team exports reviewed plays to its existing sequencer for outreach execution.

Read: A research-led calling playbook

In Evergrowth · Play Copywriting

Email Warmup

BeginnerSales Ops

Email warmup describes a gradual increase in sending activity intended to establish a sending pattern. It does not guarantee inbox placement, and artificial engagement is not a substitute for legitimate sending practices.

Example
A team introduces a new sending domain cautiously and monitors actual delivery outcomes.

Read: A research-led calling playbook

Research & Account Intelligence

Scheduled Research

IntermediateRevOps

Scheduled research refreshes selected account or contact information at recurring intervals. Its value depends on the audience, research questions and follow-up rules. A completed research run does not necessarily send a change alert or immediately detect every event.

Example
A research agent refreshes records in a saved view each week.

Read: High-velocity and enterprise sales

In Evergrowth · Scheduled Research

AI Roleplay

IntermediateSales Reps & Enablement

AI roleplay lets sellers rehearse a conversation with a simulated buyer. Scenarios can use buyer-role and account context. A separate evaluation process may assess the transcript against a rubric; simulation and coaching are distinct jobs.

Example
A simulated buyer challenges a seller's pitch, then a separate coach reviews the session.

Read: ARIS adoption story

In Evergrowth · Voice Roleplay and Roleplay Coach

Champion Monitoring

IntermediateRevenue teams

Periodically checking whether important contacts remain in their roles and accounts. A detected move can identify relationship risk or a potential warm introduction, but requires review before acting.

Example
Re-check an open deal's champion and investigate a detected employer change.

Read: IGEN account research story

In Evergrowth
In Evergrowth, a job change is detected on the next configured re-check, rather than through an instant job-change alert.
Champion monitoring

GTM Operations & Architecture

Bi-directional Sync

IntermediateRevOps

Bi-directional sync transfers selected changes in both directions between systems. Actual coverage depends on supported objects, field mapping, timing, filters and conflict rules. Two-way integration does not mean every field is continuously mirrored.

Example
Account updates flow out while CRM changes are imported on the configured schedule.

Read: How the revenue stack evolves

In Evergrowth · CRM integrations

Pipeline Velocity

IntermediateRevOps & Sales Leaders

Pipeline velocity estimates expected revenue throughput as opportunity count multiplied by average deal value and win rate, divided by average sales cycle length. More eligible opportunities, higher value or win rate, or a shorter cycle increases the calculated velocity.

Example
10 opportunities at €10,000, a 20% win rate and a 40-day cycle imply €500 of expected revenue per day.

Read: How the revenue stack evolves

Agent Hub

IntermediateRevenue teams

Evergrowth's surface for finding, creating and managing configured agents. It is distinct from the Agent Training Center, which holds shared GTM intelligence, and from workflows, which coordinate execution.

Example
RevOps manages a research agent, then includes it in a scheduled workflow.

Read: How the revenue stack evolves

In Evergrowth
Configured agent instances are different from the specialist agent types in the catalog.
Agent Hub

New-business Gap / Whitespace

IntermediateRevenue teams

Eligible target accounts outside the team's current CRM coverage or active selling work. The meaning of whitespace should state whether it refers to missing companies, unworked accounts or expansion opportunities.

Example
Identify qualified companies in a target region that are not yet represented in the CRM.

Read: How the revenue stack evolves

In Evergrowth · New-business gap

Metrics & Measurement

Metric examples are illustrative arithmetic, not industry benchmarks. Define the cohort, denominator and period before comparing results.

Competitive & Category Terms

Automation performs tasks with limited manual execution. Augmentation helps people do their work, including through automated preparation. A team can automate research while keeping sellers responsible for review and conversations. Neither pattern alone determines output quality.

Example
Research runs automatically while a seller reviews the output and leads the buyer conversation.

Read: The future of GTM software

In Evergrowth · Workspace overview