The AI in GTM Glossary

Artificial intelligence is the ability of software to perform tasks that normally require human judgment: reading a company's website and summarizing what they do, deciding whether an account fits your ICP, or writing a personalized first line for an email. In GTM, AI is most useful when it handles the research and prep work that eats up rep time, so humans can focus on the conversations that actually close deals.

A large language model is the type of AI behind tools like ChatGPT, Claude, and Gemini. It's trained on massive amounts of text and learns to generate human-sounding responses. In GTM, LLMs power everything from account research summaries to outreach copy. The model itself is a commodity. What matters is how it's trained on your specific GTM context: your ICP, your personas, your value proposition.

Generative AI refers to AI systems that create new content rather than just analyzing existing data. In sales, this means writing emails, generating account research summaries, building meeting prep briefs, or drafting call scripts. The output quality depends entirely on the context you feed it. Generic prompts produce generic output. Structured, persona-specific context produces outreach that sounds like a human who did their homework.

Natural language processing is how AI understands and generates human language. It's what allows an agent to read a company's 'About' page and extract that they serve mid-market SaaS companies, or to analyze a LinkedIn profile and determine someone's seniority and decision-making authority. NLP is the foundation that makes AI-driven research and qualification possible.

A prompt is the instruction you give an AI model. In consumer AI, you type a question and get an answer. In GTM, prompts are more structured: they include your ICP criteria, persona definitions, value proposition, and the specific account data the agent should analyze. The quality of the prompt determines the quality of the output. This is why training your agents with structured GTM context matters more than picking the 'best' model.

A hallucination is when AI generates information that looks real but isn't. In GTM, this is dangerous: a fabricated revenue figure in account research, a made-up job title for a contact, or a fictional product feature attributed to a competitor. Hallucinations happen when the model lacks context and fills the gap with plausible-sounding fiction. The primary defense is building fallback mechanisms directly into the agent's prompt: if the agent can't find the information or complete the task with confidence, it follows a fallback branch that says "I couldn't verify this" instead of inventing an answer. Grounding agents in verified data sources and adding human review layers help too, but the fallback design is what prevents fabrication at the source.

The context window is the amount of information an AI model can process in a single interaction, measured in tokens (roughly words). In an agentic workflow, upstream agents build the context window for downstream agents. Account Qualification and Account Research produce structured outputs (scores, research briefs, signals) that become the input context for Contact Finder, Play Copywriting, and other agents further down the chain. Each agent enriches the context window so the next one has more to work with. This is why purpose-built GTM workspaces outperform general-purpose AI: they structure and accumulate context across the entire workflow rather than starting from scratch at each step.

A token is the unit AI models use to measure text. Roughly, one token equals about three-quarters of a word. Tokens matter in GTM because they determine cost and capacity. Every agent run consumes tokens: reading account data, processing your ICP criteria, generating research output. Understanding token economics helps RevOps teams forecast agent costs and optimize workflows.

Fine-tuning is the process of training an existing AI model on specialized data to improve its performance in a specific domain. In GTM, this could mean training a model on thousands of successful outreach emails or qualification decisions. The alternative to fine-tuning is structured prompting with rich context, which is often more practical because it doesn't require a data science team and can be updated instantly when your ICP evolves.

RAG is a technique where an AI model retrieves relevant information from external sources before generating a response. Instead of relying only on what it learned during training, the model pulls fresh data in real time. In GTM, RAG is what allows agents to research a company's latest funding round, leadership changes, or product launches rather than working from stale training data.

An embedding is a numerical representation of text that captures its meaning. Embeddings allow AI systems to find semantically similar content: matching an account's description to your ICP criteria, or finding contacts whose LinkedIn summaries align with your buyer personas. They're the technical foundation behind intelligent search and matching in GTM tools.

A model provider is the company that builds and hosts the underlying AI model: OpenAI (GPT), Anthropic (Claude), Google (Gemini), and others. The model is the engine, but the engine alone doesn't do GTM work. What matters is the workspace built around it: the agent training, the CRM integration, the structured workflows, and the institutional memory. Your competitors have access to the same models. The differentiation is in the context layer, not the model.

An AI agent is software that can take autonomous actions to complete a task, not just answer questions. In GTM, an agent might qualify an account against your ICP criteria, research a company's tech stack, find persona-fit contacts, or write personalized outreach. Agents follow structured workflows with defined inputs and outputs. They're not chatbots waiting for instructions. They're digital colleagues that do real work.

An agentic workflow is a sequence of AI agents that execute tasks in a defined order, where the output of one agent becomes the input for the next. For example: Account Qualification scores the account, Account Research gathers intelligence, Contact Finder identifies the right people, and Play Copywriting generates personalized outreach. The chain runs autonomously. RevOps designs the workflow once; agents execute it at scale.

Agentic GTM is a go-to-market approach where specialized AI agents handle the research, qualification, enrichment, and outreach preparation that revenue teams traditionally do manually. Instead of reps spending hours on pre-call research or RevOps manually enriching lists, trained agents do it continuously and consistently. The key difference from 'AI-assisted GTM' is autonomy: agents don't just suggest actions, they complete them.

An AI SDR is an AI system designed to perform the work of a Sales Development Representative: prospecting, qualifying leads, writing outreach, and booking meetings. Some products fully automate the SDR role. Others augment human SDRs by handling the research and writing while reps focus on conversations. The risk with full automation is quality: if the AI doesn't understand your buyer, it sends the same shallow outreach at higher volume, which is worse, not better.

An AI assistant waits for you to ask a question and gives you an answer. An AI agent takes autonomous action to complete a task. ChatGPT is an assistant: you ask it to write an email, it writes one. A GTM agent is different: you give it an account list, and it qualifies, researches, finds contacts, and writes outreach without further input. The distinction matters because most 'AI sales tools' are assistants with a chatbot interface, not agents that do independent work.

A digital colleague is an AI agent that works alongside your team as a specialist, not a replacement. Each digital colleague has a specific job: one qualifies accounts, another researches companies, another finds contacts, another writes outreach. They share the same institutional knowledge (your ICP, personas, value props) and produce consistent, research-backed output that reps can review and use.

Human-in-the-loop means a human reviews and approves AI output before it reaches the prospect. In GTM, this typically means reps review agent-generated research and outreach before sending. HITL is the right default for most teams: it catches hallucinations, adds relationship context the AI doesn't have, and keeps reps connected to their accounts. Full automation without review is tempting but risky.

A multi-agent system is an architecture where multiple specialized AI agents work together, each handling a different part of a complex task. In GTM, this means separate agents for qualification, research, contact finding, enrichment, and copywriting, all connected in a workflow. The advantage over a single 'do everything' agent is specialization: each agent is optimized for one job and produces higher-quality output in its domain.

An agent chain is a connected sequence of agents where each one processes data and passes its output to the next. A typical GTM chain: Account Qualification → Account Research → Contact Finder → Contact Qualification → Play Copywriting. RevOps configures the chain once. It runs on every account automatically.

The orchestration layer is the system that coordinates multiple agents, managing the flow of data between them, handling errors, and deciding which agents to trigger based on conditions. Think of it as the project manager for your agent team. It decides: 'This account passed qualification, so trigger research. Research found a hiring signal, so route it to signal-based outreach.' Without orchestration, agents are isolated tools. With it, they're a connected workflow.

Tool use is an AI agent's ability to call external tools and APIs during its reasoning process. Instead of just generating text, the agent can search LinkedIn, query a CRM, check a company's tech stack, or verify an email address. This is what separates a chatbot from a functional agent: it doesn't just think, it acts. Tool use is the technical capability that makes agent workflows possible.

Guardrails are constraints built into AI agents to prevent unwanted behavior: limiting what data agents can access, ensuring outreach follows brand guidelines, preventing agents from making claims you can't substantiate, or stopping agents from contacting people who've opted out. Good guardrails are what make the difference between 'AI that helps' and 'AI that sends embarrassing emails to your CEO's prospect.'

Your ideal customer profile is a structured description of the companies most likely to buy and succeed with your product. It includes industry, company size, geography, tech stack, business model, and growth stage. In AI-driven GTM, your ICP isn't just a slide in a deck. It's the criteria agents use to qualify every account automatically. If your ICP is vague, your agents produce vague results.

A buyer persona is a detailed description of the person you sell to: their role, responsibilities, decision-making authority, daily challenges, and what they care about in a conversation. In context-driven GTM, personas go beyond job titles. A 'VP of Sales' at a 50-person startup and a 'VP of Sales' at a 5,000-person enterprise have completely different problems. Your personas should reflect that nuance because agents use them to find and qualify contacts.

A persona card is a structured document that defines a specific buyer persona in a format agents can use. It includes the persona's typical title patterns, responsibilities, decision-making role, pain points, communication preferences, and what a good conversation looks like. In Evergrowth's Agent Training Center, persona cards are one of the four pillars that every agent pulls from when qualifying contacts and writing outreach.

An ICP vertical is a specific market segment within your ideal customer profile, defined by industry, use case, or business model. If your ICP is 'mid-market B2B SaaS,' your verticals might be 'HR Tech,' 'FinTech,' and 'MarTech.' Verticals matter because agents use them to score accounts: a company might match your general ICP but score higher in one vertical than another based on specific qualification criteria.

A buying signal is an observable change at a company that suggests they might be ready to buy: a new VP of Sales hire, a funding round, expansion into a new market, a competitor evaluation, or a technology change. Signals are context-driven research findings, not database fields. 'New VP of Sales starts next month' is a signal. 'Hiring signal detected' is a data point. The difference is what your reps can actually use in a conversation.

Intent data measures online behavior that suggests a company is actively researching solutions in your category. This includes content consumption patterns, search queries, review site visits, and comparison page views. Intent data is useful as one signal among many, but it's not a silver bullet. An account showing 'high intent' in Bombora still needs qualification against your ICP, contact research, and personalized outreach to convert.

First-party intent comes from behavior on your own properties: website visits, content downloads, demo requests, pricing page views. Third-party intent comes from behavior tracked across the broader web by vendors like Bombora or G2. First-party intent is more reliable because you know exactly what they did. Third-party intent is broader but noisier. Most teams should prioritize first-party signals and use third-party as a supplementary layer.

Firmographic data describes a company's basic characteristics: industry, employee count, revenue, headquarters location, founding year, and business model. It's the minimum you need for account qualification. But firmographics alone are shallow. Knowing a company has 500 employees and is in FinTech tells you almost nothing about whether they need your product right now. That requires research, not just data fields.

Technographic data describes a company's technology stack: what tools they use, what they've recently adopted or dropped, and what integrations they run. It's valuable for competitive displacement plays ('they use Competitor X, here's why they should switch') and compatibility qualification ('they already use Salesforce, so our integration works'). Tech stack data is one of the strongest buying signals when combined with other research.

Signal-based selling is an outbound approach where you reach out based on specific changes at a company, not just static list criteria. Instead of emailing every FinTech company with 200+ employees, you reach out to the ones that just hired a new CRO, raised a Series B, or expanded into EMEA. Signal-based playbooks produce higher response rates because the timing is relevant and the outreach can reference the specific change.

The buying committee is the group of people involved in a purchase decision. In B2B, this typically includes a champion (who wants your product), an economic buyer (who controls budget), a technical evaluator (who vets the solution), and sometimes a legal or procurement gatekeeper. Deals stall when you're single-threaded with one contact. Account planning maps the full committee so reps can multi-thread effectively.

A champion is someone inside the target company who believes in your solution and actively advocates for it internally. They sell when you're not in the room. Champions are the most important contacts in any deal because they navigate internal politics, build consensus, and push procurement. When a champion leaves, your deal is at risk. Champion monitoring automates the detection of these departures.

Multi-threading means building relationships with multiple stakeholders in a target account, not just one contact. Single-threaded deals die when your one contact changes roles, goes on leave, or loses influence. Multi-threaded deals survive because you have multiple paths to the decision. In practice, this means using contact finder and research agents to identify and engage 3-5 people across the buying committee.

DISC is a behavioral assessment framework that categorizes communication styles into four types: Dominance, Influence, Steadiness, and Conscientiousness. In AI-driven GTM, agents can infer a contact's likely DISC profile from their LinkedIn activity, writing style, and role to calibrate outreach tone. A high-D executive wants bullet points and bottom-line impact. A high-S operations leader wants detailed process and low risk. Matching tone to personality increases response rates.

The Agent Training Center is where you teach AI agents your GTM intelligence: ICP verticals, buyer personas, value propositions, and key accounts. You define these once. Every agent uses them. Every rep benefits from them. It's the difference between generic AI that writes the same email for every prospect and trained agents that produce research and outreach calibrated to your specific market.

In traditional sales, a value proposition is a statement about why someone should buy from you. In AI-driven GTM, it's a structured input that agents use to write relevant outreach: your proven outcomes, competitive differentiators, objection responses, and use-case-specific messaging. The more specific your value prop, the better agents can tailor outreach to each persona and vertical.

Institutional memory is the accumulated GTM knowledge that persists across your organization regardless of who's on the team. When your best rep leaves, their research, qualification notes, and outreach strategies shouldn't leave with them. In an agentic workspace, institutional memory lives in the Agent Training Center and shared research outputs, not in individual reps' heads or personal spreadsheets.

Data enrichment is the process of adding missing information to your CRM records: filling in company revenue, employee count, tech stack, or contact email addresses. Traditional enrichment is field-level: you send in a domain, you get back 12 data points. Context-driven research goes deeper: you send in a domain, agents analyze the website, news, hiring patterns, and competitive landscape to produce a research brief reps can actually use.

Waterfall enrichment is a technique where a data request cascades through multiple vendors in sequence until it finds a match. If Vendor A doesn't have the email, try Vendor B, then C, then D. This maximizes coverage and accuracy while minimizing cost because you only pay for successful matches. In the Email & Phone Waterfall agent, requests cascade through 20+ providers automatically.

Contact verification is the process of confirming that a contact's information is accurate and deliverable. For email, this means checking that the address exists, isn't a catch-all, and won't bounce. For phone, it means confirming the number is active and reaches the right person. Verification should happen before outreach, not after. Sending to unverified contacts damages your domain reputation and wastes rep time.

Email deliverability is the percentage of your emails that actually reach the recipient's inbox instead of landing in spam or bouncing. It's affected by domain reputation, sending volume, content quality, authentication (SPF, DKIM, DMARC), and whether you're sending to verified addresses. High-quality, personalized outreach has better deliverability than mass templates because recipients engage with it instead of marking it as spam.

Bounce rate is the percentage of sent emails that fail to reach the recipient's mailbox. Hard bounces mean the address doesn't exist. Soft bounces mean the mailbox is temporarily unavailable. A bounce rate above 3-5% signals a data quality problem and will damage your sender reputation. This is why contact verification before sending matters: it catches bad addresses before they become bounces.

Data decay is the rate at which contact and company data becomes inaccurate over time. On average, 30% of B2B contact data goes stale every year: people change jobs, companies get acquired, phone numbers change. This is why enrichment isn't a one-time event. Scheduled research and contact requalification agents counter data decay by continuously validating your CRM records.

CRM hygiene is the practice of keeping your customer relationship management system accurate, complete, and free of duplicates. Poor hygiene means reps waste time on outdated contacts, deals get attributed to the wrong accounts, and reporting becomes unreliable. Agent-driven requalification can clean your CRM at scale: checking whether contacts still work at the company, updating firmographics, and flagging records that need attention.

A data vendor is a company that sells business contact and company information: ZoomInfo, Apollo, Cognism, Lusha, and others. They aggregate data from public sources, web scraping, surveys, and partnerships. The challenge with data vendors is that everyone has access to the same data. If you and your competitor both buy from ZoomInfo, you're sending the same emails to the same people. Context is the differentiator, not the data itself.

Credit-based pricing is a model where you pay per action (per enrichment, per research, per email found) rather than a flat subscription for unlimited access. It aligns cost with usage: you pay for what you actually consume. In GTM tools, credits typically have different costs per action. For example, an email lookup might cost 0.3 credits while a phone number costs 2 credits, reflecting the different vendor costs behind each lookup.

A catch-all domain is configured to accept emails sent to any address at that domain, whether the specific mailbox exists or not. This makes email verification difficult because the server won't bounce invalid addresses. Sending to catch-all domains carries risk: the email might arrive but go unread, or it might reach the wrong person. Most email verification tools flag catch-all addresses so you can handle them separately.

A direct dial is a phone number that reaches a specific person without going through a switchboard or gatekeeper. Direct dials are the most valuable phone numbers in outbound sales because they skip the 'Can I speak to...' step entirely. They're also harder to find than general company numbers, which is why phone waterfall enrichment cascades through multiple providers to maximize match rates.

Contact enrichment fills in data fields: email, phone, title, company. Contact research produces context: what this person cares about, what they've been posting on LinkedIn, their communication style, their likely priorities based on their role and company situation. Enrichment gives you the ability to reach someone. Research gives you something worth saying when you do.

A domain finder is an agent or tool that identifies the correct website domain for a company based on limited information like a company name. This sounds simple but it's surprisingly hard at scale: companies have multiple domains, subsidiaries use different URLs, and common names create ambiguity. The Domain Finder agent resolves this before the rest of the workflow runs, ensuring all downstream research starts from the right company.

Lead scoring assigns a numerical value to leads based on how likely they are to convert. Traditional lead scoring uses behavioral signals (website visits, email opens) and firmographic data (company size, industry). The problem is that most lead scoring models are formula-based and static. An account that visits your pricing page twice gets a high score even if they don't match your ICP at all.

Account scoring evaluates companies (not individual contacts) against your ICP criteria to determine fit and priority. Agent-driven account scoring is more nuanced than traditional models because agents can analyze qualitative factors: not just 'Is this a SaaS company with 200+ employees?' but 'Does their website suggest they're actively expanding their sales team, and is their tech stack compatible with our integrations?'

A qualification score is a research-backed assessment of how well an account or contact matches your ideal criteria. Unlike formula-based lead scores, a qualification score comes with evidence: the agent shows why it scored the account 75/100 by listing which ICP criteria matched and which didn't. In Evergrowth's Account Qualification, the score represents the percentage of research that came back positive.

A marketing qualified lead is a contact that marketing has determined is worth passing to sales based on engagement criteria: downloaded a whitepaper, attended a webinar, visited the pricing page. The MQL model is under pressure because engagement doesn't equal fit. Someone who downloads every ebook isn't necessarily a buyer. Context-driven qualification looks at both fit (ICP match) and timing (buying signals), not just engagement volume.

A sales qualified lead is a prospect that sales has confirmed meets the criteria for a real opportunity: they have budget, authority, need, and timeline (or some variation of those criteria). The SQL handoff from marketing to sales is where most GTM friction lives. Agent-driven qualification can pre-validate many SQL criteria before the lead even reaches a rep, reducing the 'I looked at this and it's not a fit' rejection rate.

A product qualified lead is a user who has demonstrated buying intent through product usage: they've hit a feature limit, used the product consistently for a certain period, or tried a premium feature during a trial. PQLs are most relevant for product-led growth companies where users try before they buy. The signal is behavioral ('they used the product enough to need more') rather than demographic ('they match our ICP on paper').

An ideal buyer persona is the contact-level equivalent of an ICP. While ICP defines the right company, IBP defines the right person within that company: their role, seniority, responsibilities, and decision-making authority. In agent workflows, IBP criteria determine which contacts the Contact Finder surfaces and which the Contact Qualification agent approves. An IBP-fit contact at an ICP-fit account is the highest-priority combination.

BANT stands for Budget, Authority, Need, and Timeline. It's one of the oldest qualification frameworks in B2B sales. A lead is 'BANT-qualified' when they have budget allocated, the authority to decide, a clear need your product solves, and a timeline to implement. BANT is simple and effective for transactional sales. For complex enterprise deals, teams often use more comprehensive frameworks like MEDDIC.

MEDDIC stands for Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, and Champion. MEDDPICC adds Paper Process and Competition. It's the gold standard qualification framework for enterprise sales because it forces reps to understand not just whether someone wants to buy, but how they buy. AI agents can pre-populate MEDDIC fields with research: identifying likely economic buyers, mapping decision processes from company structure, and finding competitive signals.

Account qualification is the process of evaluating whether a company matches your ICP before investing sales time. Agent-driven account qualification analyzes the company's website, financials, tech stack, hiring patterns, and market position against your ICP verticals and produces a scored assessment with evidence for each criterion. This replaces the manual 'let me Google this company for 10 minutes' step that reps do thousands of times.

Contact qualification validates whether a specific person matches your buyer persona criteria: Do they still work at the company? Does their role match the personas you sell to? Do they have decision-making authority? Contact qualification agents run this check against your persona cards and can requalify entire CRM databases periodically to catch job changes and role shifts.

Disqualification criteria define the conditions under which an account or contact should be excluded from outreach. Examples: company too small, wrong geography, already a customer, in an active deal, or in a regulated industry you don't serve. Good disqualification criteria are just as important as qualification criteria because they prevent reps from wasting time and protect your brand from irrelevant outreach.

Personalization at scale is the ability to send outreach that feels individually crafted to each recipient, even when reaching hundreds or thousands of contacts. The old approach was merge tags: 'Hi {first_name}, I see you work at {company}.' That's not personalization; it's mail merge. Real personalization references what the person cares about, what's changing at their company, and why the timing is relevant. AI makes this possible, but only when agents have enough context to personalize meaningfully.

A sequence is a pre-written series of template emails sent on a schedule: Day 1 email, Day 3 follow-up, Day 7 breakup. A play is a personalized outreach piece written for a specific contact based on their persona, DISC profile, account research, and buying signals. Sequences scale volume. Plays scale relevance. When agents write plays, each contact gets outreach calibrated to who they are and what's happening at their company.

A cadence is a structured series of touchpoints across multiple channels (email, phone, LinkedIn, video) designed to engage a prospect over time. Most sales engagement platforms organize outreach into cadences. The cadence defines the timing and channel mix. What fills the cadence, templates or context-driven plays, determines whether prospects feel spammed or understood.

A cold email is an outreach message sent to someone you have no prior relationship with. Cold email gets a bad reputation because most of it is terrible: generic templates blasted to purchased lists. Done well, cold email is a first impression backed by research. The difference between 'spam' and 'relevant outreach' isn't the channel; it's whether you actually understand the person you're writing to and have something worth saying.

A warm intro is a referral or introduction from a mutual connection. Warm intros convert at significantly higher rates than cold outreach because trust transfers from the connector to the seller. In practice, warm intros are hard to scale because they depend on relationships. The closest you can get at scale is outreach so well-researched and relevant that it feels like a warm conversation, not a cold pitch.

A sales engagement platform is software that automates multi-channel outreach sequences: Outreach, Salesloft, Apollo, and similar tools. SEPs handle the sending, tracking, and follow-up scheduling. They're the delivery mechanism. What they don't do is research accounts, qualify contacts, or write personalized outreach. That's the gap AI agents fill: they produce the content and context that SEPs then deliver.

Play copywriting is the AI-driven generation of personalized outreach for individual contacts. The Play Copywriting agent takes everything upstream agents have produced, account research, contact research, DISC profile, buying signals, and writes a message calibrated to that specific person. The output isn't a template with variables. It's a unique piece of writing that references the recipient's actual situation.

Template-driven outreach starts with a pre-written email and fills in variables: name, company, industry. Context-driven outreach starts with research and writes the email from scratch based on what agents found. The difference in prospect experience is obvious: one feels mass-produced, the other feels like someone did their homework. Context-driven outreach produces higher response rates because it is, by definition, relevant.

Email warmup is the process of gradually increasing sending volume on a new or underused email domain to build a positive sender reputation with email providers (Google, Microsoft, etc.). You start by sending a few emails per day to engaged contacts who will open and reply, then slowly scale up. Skipping warmup and blasting 500 cold emails from a new domain is the fastest way to land in spam permanently.

Domain reputation is how email providers (Gmail, Outlook) evaluate your sending domain based on historical behavior: bounce rates, spam complaints, engagement rates, and authentication. A good reputation means emails reach the inbox. A bad reputation means they go to spam, regardless of how good the content is. This is why many outbound teams use separate sending domains and warm them up carefully before scaling volume.

Multi-channel outreach is the practice of engaging prospects across multiple communication channels: email, phone, LinkedIn, video, and sometimes direct mail. The logic is simple: people have different channel preferences, and showing up in multiple places increases the odds of connecting. The challenge is maintaining personalization across channels. Sending a generic LinkedIn message that contradicts your email undermines the entire approach.

A/B testing in outreach means sending two variations of an email (different subject lines, different opening lines, different CTAs) to similar audiences and measuring which performs better. It's how you systematically improve messaging over time. In AI-driven outreach, A/B testing shifts from testing templates to testing prompt strategies: which context inputs produce higher-performing plays.

Account research is the process of gathering intelligence about a target company before reaching out or meeting with them. Traditional research means a rep spending 15 minutes on LinkedIn, the company website, and Google. Agent-driven account research does this in seconds, analyzing multiple sources and producing a structured brief covering business model, recent news, hiring signals, tech stack, and competitive landscape.

Pre-call research is the preparation work a rep does before a meeting or call: reviewing the company's website, checking LinkedIn for recent activity, reading news, and reviewing CRM notes. Most reps spend 15-30 minutes on this per meeting. Agents compress this to seconds by producing a structured brief that covers everything a rep needs to know: company context, recent signals, stakeholder profiles, and competitive landscape.

Account planning is the strategic process of mapping out how to win a specific account: identifying stakeholders, understanding the buying process, mapping competitive threats, and building an engagement strategy. The Account Planning agent automates the data-gathering portion: it maps the buying committee, identifies potential champions, surfaces relevant signals, and recommends a multi-threading approach so reps can focus on strategy instead of research.

Competitive intelligence is the ongoing collection and analysis of information about your competitors: their positioning, pricing, product changes, customer wins and losses, and market moves. In AI-driven GTM, agents can surface competitive signals during account research: 'This company uses Competitor X based on their tech stack' or 'Their job posting mentions evaluating alternatives to their current CRM.' This turns competitive intelligence from a quarterly deck into a real-time input for outreach.

Scheduled research is the automated, periodic re-research of accounts and contacts in your CRM. Instead of relying on stale data from when the account was first imported, scheduled research agents run on a cadence (weekly, monthly, quarterly) to surface new signals, validate existing data, and alert reps when something changes at an account they own.

A digital twin is an AI replica of a specific buyer persona that reps can practice conversations with. Based on the persona card, company context, and DISC profile, the digital twin responds the way the actual buyer might: raising objections, asking tough questions, and pushing back on positioning. It's practice before the real thing. Not a sequential workflow step, but an on-demand preparation tool.

AI roleplay is a practice environment where reps rehearse sales conversations with an AI-simulated buyer. Unlike generic roleplay bots, context-aware roleplay agents are trained on your buyer personas, value proposition, and common objections. The AI plays the role of a specific persona type and scores the rep's performance against your playbook. It's on-demand coaching without needing a manager's calendar.

AI sales coaching uses AI to provide feedback on rep performance: analyzing call recordings, scoring discovery questions, identifying missed objections, and recommending improvements. The best coaching AI is contextual: it knows your sales methodology, your personas, and your competitive positioning, so it can give specific feedback like 'You missed the budget question' rather than generic advice like 'Ask more open-ended questions.'

Buying committee mapping is the process of identifying all stakeholders involved in a purchase decision at a target account: who champions the deal, who controls budget, who evaluates the technology, and who can block it. Mapping the committee early prevents single-threading and late-stage surprises. Agents can accelerate this by analyzing the company's organizational structure and matching contacts to persona types.

Win/loss analysis is the practice of systematically reviewing closed deals (both won and lost) to identify patterns in why you win or lose. Common findings: lost deals often had single-threaded relationships, weak champion support, or misaligned use cases. AI can accelerate win/loss analysis by processing CRM data and call transcripts to surface patterns that would take a human analyst weeks to find.

A stakeholder map is a visual representation of all the people involved in a buying decision at a target account, showing their roles, relationships, influence levels, and stance toward your solution. Building a stakeholder map early in a deal prevents surprises in later stages. Agents can draft an initial map by analyzing the company's org structure and matching contacts to your buyer personas, which the rep then refines with real relationship intelligence.

A GTM engineer is an emerging role that sits between RevOps and data engineering. They design, build, and maintain AI-powered GTM workflows: configuring agent chains, connecting data sources, building enrichment logic, and optimizing automation. Think of them as the person who turns your GTM strategy into a working system. The risk with this role is the single-operator bottleneck: if all your GTM intelligence lives in one person's spreadsheet, you're one resignation away from starting over.

Revenue Operations is the function that aligns sales, marketing, and customer success around shared data, processes, and metrics. RevOps owns the GTM stack, manages data quality, designs workflows, and reports on pipeline and revenue. In AI-driven GTM, RevOps becomes the team that trains and manages agents, not just the team that maintains spreadsheets and dashboards.

A GTM motion is the repeatable way your company acquires and expands customers: product-led growth (PLG), sales-led growth (SLG), partner-led, community-led, or a hybrid. Your GTM motion determines which agents and workflows matter most. A PLG company needs inbound qualification and product usage signals. A sales-led company needs outbound research, contact finding, and personalized outreach at scale.

Your GTM stack is the collection of tools your revenue team uses to execute go-to-market: CRM, sales engagement platform, enrichment tools, conversational intelligence, analytics, and now AI agents. The average B2B company uses 10-15 tools in their GTM stack. The challenge isn't having tools; it's connecting them so data flows cleanly and agents can act on it. Disconnected tools create data silos that undermine everything else.

CRM integration is the ability of a tool to read from and write to your CRM (Salesforce, HubSpot, Pipedrive, etc.) automatically. In agentic GTM, CRM integration is what allows agents to pull account data for research, push qualification scores back to the CRM, and update contact records with enrichment results. Without CRM integration, agents produce output that sits in a separate system and never reaches the rep's workflow.

Bi-directional sync means data flows both ways between two systems. In GTM, this typically means your CRM and your agent workspace stay in sync: agents read account data from the CRM and write research results back. One-directional sync (CRM to tool only, or tool to CRM only) creates version conflicts and stale data. Bi-directional sync ensures reps always see the latest agent output in the system they already work in.

Lead routing is the process of assigning incoming leads to the right sales rep based on rules: geography, company size, industry, round-robin, or account ownership. Bad lead routing means high-value leads sit in a queue while reps cherry-pick the ones they recognize. Good lead routing combined with agent qualification means leads arrive pre-researched and pre-scored, so the assigned rep can act immediately instead of doing their own research first.

An SLA (service level agreement) between sales and marketing defines the commitments each team makes: marketing agrees to deliver a certain number of qualified leads per month, and sales agrees to follow up within a certain timeframe. SLAs reduce finger-pointing by making expectations explicit and measurable. When agents handle qualification and enrichment, the SLA can shift from 'number of MQLs' to 'number of research-backed, ICP-qualified accounts.'

Attribution is the process of determining which marketing and sales activities contributed to a conversion or deal. First-touch attribution credits the first interaction. Last-touch credits the final one. Multi-touch distributes credit across the journey. Attribution is important for budget allocation but notoriously difficult to get right, especially in B2B where buying cycles are long and involve many touchpoints across multiple channels.

A GTM playbook is a documented, repeatable workflow for a specific go-to-market scenario: how to handle inbound leads, how to follow up after an event, how to recycle closed-lost deals, how to respond to buying signals. In an agentic workspace, playbooks aren't just documents. They're configured agent workflows that execute automatically when triggered.

Pipeline velocity measures how fast deals move through your pipeline, calculated as: (number of opportunities x average deal value x win rate) / average sales cycle length. Increasing any of those four variables increases velocity. AI agents impact all four: better qualification increases win rates, deeper research shortens sales cycles, and automated outreach increases opportunity volume.

Closed-lost recycling is the practice of re-engaging deals that didn't close rather than letting them sit dormant in your CRM. An account that said 'not now' six months ago might be ready today if something changed: new leadership, budget cycle reset, competitor disappointment, or a buying signal. Recycling playbooks automate the re-research and re-qualification so reps only re-engage accounts where something actually shifted.

A dormant account is one in your CRM with no open opportunity, no recent activity, and no scheduled follow-up. These accounts represent untapped pipeline because someone once qualified them as worth pursuing. Reactivation workflows use agents to re-research dormant accounts, check for fresh buying signals, and surface the ones worth re-engaging, turning dead CRM weight into active pipeline.

Backfill mode is running agents on your existing CRM data to retroactively enrich, qualify, and research accounts that were imported before you had AI workflows. Instead of only applying agents to new accounts going forward, backfill mode processes your historical database so every account gets the same research depth. It's the fastest way to find hidden pipeline in data you already have.

Annual recurring revenue is the total value of recurring subscription revenue normalized to a one-year period. ARR is the foundational metric for SaaS businesses because it represents predictable, ongoing income. For GTM teams, ARR growth is the ultimate measure of whether your sales and marketing efforts are working. Every metric upstream (pipeline, meetings, qualification rates) ultimately ladders up to ARR.

Conversion rate measures the percentage of records that progress from one stage to the next in your funnel: website visitor to lead, lead to MQL, MQL to SQL, SQL to opportunity, opportunity to closed-won. Each conversion rate tells you where your funnel is healthy and where it leaks. Agent-driven qualification typically improves the MQL-to-SQL conversion rate by filtering out poor-fit leads before they reach sales.

Meeting booked rate is the percentage of outreach attempts that result in a scheduled sales meeting. For cold outbound, a 2-5% meeting booked rate is typical. Context-driven outreach tends to outperform template outreach on this metric because prospects are more likely to take a meeting when the outreach demonstrates understanding of their situation. It's the clearest single metric for measuring outbound quality.

Time-to-pipeline measures how long it takes from initial activity (a signal detected, a list uploaded, an event attended) to a qualified opportunity in the CRM. Shorter time-to-pipeline means your GTM engine is efficient. Agent workflows compress time-to-pipeline dramatically by running qualification, research, contact finding, and outreach generation in minutes rather than the days or weeks it takes manually.

Quota attainment is the percentage of a rep's or team's sales target that was actually achieved in a given period. Industry benchmarks suggest roughly 60% of B2B reps miss quota. The gap is often not effort or talent but preparation: reps spend too much time on research and admin and not enough time selling. Agents that handle the research and prep work give reps more selling hours per week, which is the most direct lever on quota attainment.

Annual contract value is the average annualized revenue per customer contract. ACV determines your GTM economics: high-ACV deals justify more personalized, research-intensive sales motions. Low-ACV deals require more automation and efficiency. Understanding your ACV helps RevOps decide how much agent processing to invest per account: a $200K ACV deal justifies the full agent chain, while a $5K deal might only need basic qualification.

The LTV/CAC ratio compares customer lifetime value (how much revenue a customer generates over their lifetime) to customer acquisition cost (how much you spent to acquire them). A healthy ratio is 3:1 or higher. AI agents improve this ratio from both sides: they reduce CAC by automating expensive research and prospecting work, and they can improve LTV by enabling better-fit customer acquisition (higher ICP scores correlate with longer retention).

Pipeline coverage is the ratio of total pipeline value to your revenue target, typically expressed as a multiple. If you need to close $1M this quarter and you have $3M in pipeline, your coverage ratio is 3x. Most B2B companies target 3-4x coverage. Low coverage means you need more top-of-funnel volume. High coverage with low win rates means you have a qualification problem, not a volume problem.

Reply rate is the percentage of outreach emails that receive a response (including negative responses). A healthy cold outreach reply rate is typically 5-15%, though context-driven outreach often exceeds this because it gives prospects something worth responding to. Reply rate is a better signal than open rate because opens are increasingly unreliable due to email privacy features.

Cost per meeting calculates the total cost of generating one qualified sales meeting, including tools, data, rep time, and AI credits. It's the metric that connects GTM investment to pipeline activity. If your reps spend 6 hours per week on research and agents can do it for $200/month in credits, the ROI math becomes straightforward. The ROI calculator helps estimate this for your specific team size and workflow.

Rep productivity measures the output a salesperson generates relative to their capacity: meetings booked, pipeline created, deals closed, or revenue per rep. AI agents increase rep productivity by eliminating the research and admin work that consumes 30-60% of a rep's week. The goal isn't to replace reps with agents. It's to give every rep the research support that only the top 10% currently do for themselves.

Ramp time is how long it takes a new sales hire to reach full productivity, typically measured as time to first deal or time to quota. Average B2B ramp time is 3-9 months. Agents compress ramp time by giving new reps instant access to institutional knowledge: ICP criteria, buyer personas, account research, and outreach templates calibrated to your market. A new rep with trained agents is productive faster than one who has to learn everything from scratch.

Credit economics is the practice of measuring the cost of agent runs (in credits) against the pipeline and revenue they generate. It answers the question: 'For every dollar we spend on AI agent credits, how much pipeline do we create?' Understanding credit economics helps RevOps optimize workflows, choosing when to run the full agent chain vs. a lighter workflow based on the expected value of the account. The credit simulator helps model this.

AI automation replaces human steps entirely: the AI sends emails, books meetings, and handles objections without rep involvement. AI augmentation keeps humans in the loop: agents do the research, qualification, and copywriting, but reps review, refine, and send. Most GTM teams need augmentation, not full automation, because deals still close on human judgment, relationship context, and the ability to read a room. Automation scales volume. Augmentation scales quality.

A point solution solves one problem: enrichment, or sequencing, or research, or coaching. A workspace connects multiple capabilities under shared context: your ICP, personas, and value proposition feed every agent and every workflow. The advantage of a workspace is consistency. When your Account Qualification agent and your Play Copywriting agent both pull from the same Agent Training Center, the outreach reflects the qualification. Point solutions create data silos.

Data-driven GTM relies on firmographic filters and behavioral scoring to build lists and send template-based outreach. Context-driven GTM uses AI agents to research accounts deeply, qualify them against specific ICP criteria with evidence, and write outreach based on what they found. The full comparison covers research, qualification, outreach, coaching, and economics across both approaches. Data is the commodity input. Context is what makes outreach feel human.

Systems of context is the philosophy that GTM competitive advantage comes from depth of understanding, not breadth of data. When agents don't just return data fields but produce research that explains why an account matters, what changed, and how to talk to the people there, you have a system of context. It's the difference between 'Company: 500 employees, SaaS, Series C' and a research brief that tells the rep exactly why this account is worth their time right now.

The GTM spreadsheet model (popularized by Clay) treats go-to-market operations like a spreadsheet where each row is an account and each column is an enrichment or AI action. It's powerful for technical operators who want full control. The limitation is that it creates a single-operator bottleneck: the GTM intelligence lives in one person's spreadsheet logic, not in shared institutional memory. When that person leaves, so does the system.

A single-operator bottleneck occurs when your entire GTM automation depends on one technical person who builds and maintains the workflows. If your GTM engineer built a brilliant Clay workflow and gets hired away, you're left with a spreadsheet nobody else can maintain. Shared workspaces solve this by making GTM intelligence accessible to the whole team: RevOps configures, reps use, and the knowledge persists regardless of who's on the team.

A sales intelligence platform provides data and insights about companies and contacts to help sales teams prospect and prepare for conversations. ZoomInfo, Apollo, Cognism, Lusha, and LinkedIn Sales Navigator all fall into this category. They vary in data coverage, accuracy, and pricing. The category is evolving as AI agents can now do real-time research that goes beyond static database lookups.

Conversational intelligence platforms (Gong, Chorus, Clari) record, transcribe, and analyze sales calls to surface patterns: talk-to-listen ratios, competitor mentions, objection handling, next steps. They're valuable for coaching and deal review. The next evolution is connecting conversational insights back to pre-call preparation: using what you learn from calls to improve how agents research accounts and write outreach.

Revenue intelligence aggregates signals from across the GTM stack, CRM, email, calls, calendar, and product usage, to provide visibility into deal health, pipeline risk, and forecast accuracy. Tools like Clari, Gong, and BoostUp sit in this category. Revenue intelligence tells you what's happening in your pipeline. Agentic GTM focuses on what happens before the pipeline: the research, qualification, and outreach that creates it.