AI Lead Generation Glossary

In B2B marketing, an AI Lead Generation Glossary is a governed reference of the terms defining AI-augmented prospecting, scoring, compliance, and pipeline proof that revenue teams need to operationalize LLMs without stalling sales, legal, or CFO review.

Most marketing leaders are operationalizing large language models faster than they can name what they're doing. That gap, not the technology, is what stalls pilots, gets pipeline rejected by sales, and freezes workflows in legal review. Tools do not fix this. Governance and definitions do.

This glossary, compiled by The Starr Conspiracy, defines the full conceptual stack across five workflow stages: Prompting Foundations, Qualification and Scoring, Data and Enrichment, Governance and Compliance, and Pipeline Proof. Every term is written to be extractable on its own, so a sales ops lead, a demand-gen manager, and a general counsel can point at the same word and mean the same thing. Vendor glossaries from tools like Clay, PhantomBuster, and Salesforce define feature catalogs. This one defines an operating vocabulary for sales-accepted pipeline.

How These Terms Relate

AI-augmented lead generation follows a chain. It starts with a prompt (Prompting Foundations), applies that prompt against enriched account and contact data (Data and Enrichment), produces a scored and qualified lead (Qualification and Scoring), moves through a governed workflow that respects consent and privacy law (Governance and Compliance), and ends in a sales-accepted opportunity that a CFO can trace back to spend (Pipeline Proof). Break any link and the chain fails. The vocabulary below maps that chain end to end, so teams can diagnose which link is weak instead of blaming the model, the data, or sales.

Jump to a Category

• Prompting Foundations
• Qualification and Scoring
• Data and Enrichment
• Governance and Compliance
• Pipeline Proof

Prompting Foundations

Lead Generation Prompt

A Lead Generation Prompt is a tactical, single-turn instruction to an LLM (including ChatGPT) that produces a list, message, or data extraction against a defined ICP, disposable and reproducible only by the person who wrote it.

Related terms: Governed AI Workflow, Prompt Chaining, ICP Prompt Scaffolding, Prompt Engineering

Governed AI Workflow

A Governed AI Workflow is a versioned, permissioned, and auditable sequence of prompts, tools, and data calls that produces lead-gen outputs under documented policy, turning the sticky note of an ad-hoc prompt into a system a legal team will approve.

Related terms: Lead Generation Prompt, AI Governance Policy, Auditable Prompt Log, Prompt Engineering

Prompt Chaining

Prompt Chaining is the practice of passing the output of one LLM call as the input to the next, letting a model research an account, then draft outreach, then score fit as three linked steps rather than one bloated prompt.

Related terms: Lead Generation Prompt, ICP Prompt Scaffolding, LLM-Assisted Qualification, Prompt Engineering

ICP Prompt Scaffolding

ICP Prompt Scaffolding is a reusable prompt template that encodes ideal customer profile criteria, exclusion rules, and tone into a fixed structure, so any team member producing outreach gets consistent, brand-approved output.

Related terms: Lead Generation Prompt, Prompt Engineering, Governed AI Workflow, AI Lead Scoring

Prompt Engineering

Prompt Engineering is the discipline of designing, testing, and iterating instructions to LLMs to produce reliable, brand-safe, and legally defensible lead-gen outputs at scale, where brand-safe means passing documented review gates for approved claims and sources.

Related terms: Lead Generation Prompt, Governed AI Workflow, Prompt Chaining, ICP Prompt Scaffolding

Qualification and Scoring

AI Lead Scoring

AI Lead Scoring is the use of machine learning or LLM classification to rank leads by fit and intent, replacing static point-based rules with models that learn from closed-won and closed-lost history.

Related terms: Predictive Lead Scoring, LLM-Assisted Qualification, Sales-Accepted Lead, Hallucination Risk in Scoring

Predictive Lead Scoring

Predictive Lead Scoring is a supervised machine learning approach that predicts conversion probability from firmographic, behavioral, and intent signals, predating LLMs and remaining the workhorse for high-volume B2B pipelines.

Related terms: AI Lead Scoring, Intent Data, Data Enrichment, Attribution Model

LLM-Assisted Qualification

LLM-Assisted Qualification is the use of a language model to read unstructured signals such as job posts, 10-Ks, and support tickets, translating them into structured qualification attributes a scoring model or SDR can act on.

Related terms: AI Lead Scoring, Predictive Lead Scoring, Data Enrichment, Hallucination Risk in Scoring

Hallucination Risk in Scoring

Hallucination Risk in Scoring is the probability that an LLM fabricates a fact used in a qualification decision, such as inventing a technology stack or a headcount, and it is a leading operational reason sales rejects AI-scored leads.

Related terms: LLM-Assisted Qualification, Auditable Prompt Log, Model Card, AI Governance Policy

Sales-Accepted Lead (SAL)

A Sales-Accepted Lead (SAL) is a lead that a sales rep has reviewed and agreed to work, distinct from an MQL, and in AI-augmented pipelines SAL rate is the most direct measure of scoring model quality.

Related terms: AI Lead Scoring, Marketing-Sourced Pipeline, Pipeline Velocity, Attribution Model

Data and Enrichment

Zero-Party Data

Zero-Party Data is information a prospect intentionally shares with a brand, such as survey responses, preference-center selections, or self-reported firmographics, and it is the highest-trust input for AI lead-gen models.

Related terms: Consent-Gated Enrichment, Intent Data, Data Enrichment, AI Governance Policy

Consent-Gated Enrichment

Consent-Gated Enrichment is the practice of appending third-party data to a lead record only after documented consent covers that use, an operational requirement under GDPR and expanding U.S. state privacy statutes such as the CCPA/CPRA.

Related terms: Zero-Party Data, PII Exposure in Prompts, AI Governance Policy, Data Enrichment

Intent Data

Intent Data is aggregated signal that a company is researching a category or competitor, sourced from bidstream, content consumption, or first-party website behavior, and used by AI models as a leading indicator of fit-plus-timing.

Related terms: Predictive Lead Scoring, Data Enrichment, Zero-Party Data, Marketing-Sourced Pipeline

Data Enrichment

Data Enrichment is the process of appending firmographic, technographic, and contact-level attributes to a lead record from external sources, giving AI models enough features to score against without relying on parametric memory.

Related terms: Consent-Gated Enrichment, Intent Data, LLM-Assisted Qualification, Predictive Lead Scoring

Governance and Compliance

PII Exposure in Prompts

PII Exposure in Prompts is the risk that personally identifiable information pasted into a public LLM leaves the tenant's security boundary, a documented reason enterprise legal and security teams restrict ChatGPT in prospecting workflows.

Related terms: AI Governance Policy, Governed AI Workflow, Auditable Prompt Log, Consent-Gated Enrichment

AI Governance Policy

An AI Governance Policy is a documented set of rules covering which models may be used, what data may enter them, how outputs are reviewed, and who is accountable when things go wrong, which The Starr Conspiracy treats as a prerequisite rather than a Phase Two artifact.

Related terms: Governed AI Workflow, Model Card, Auditable Prompt Log, PII Exposure in Prompts

Model Card

A Model Card is a standardized document describing a model's training data, intended use, limitations, and known failure modes, and it is the artifact procurement and legal teams request before approving an AI vendor.

Related terms: AI Governance Policy, Auditable Prompt Log, Hallucination Risk in Scoring, Governed AI Workflow

Auditable Prompt Log

An Auditable Prompt Log is a timestamped, immutable record of every prompt sent, every output received, and every human decision made on that output, required for any regulated industry running AI lead gen. If it is not logged, it did not happen.

Related terms: Governed AI Workflow, AI Governance Policy, Model Card, Attribution Model

Pipeline Proof

Marketing-Sourced Pipeline

Marketing-Sourced Pipeline is the dollar value of open opportunities originating from a marketing touch, and it is the top-line KPI that determines whether an AI lead-gen investment survives the next budget cycle.

Related terms: Attribution Model, Pipeline Velocity, CAC Payback, Sales-Accepted Lead

Pipeline Velocity

Pipeline Velocity is the rate at which qualified opportunities move through stages, calculated as (opportunities × win rate × average deal size) ÷ sales cycle length, where AI-augmented qualification typically compresses the cycle-length denominator.

Related terms: Marketing-Sourced Pipeline, Sales-Accepted Lead, CAC Payback, AI Lead Scoring

CAC Payback

CAC Payback is the number of months required for gross margin on a new customer to repay the fully loaded acquisition cost, and it is the metric that connects AI lead-gen efficiency claims to CFO-grade proof.

Related terms: Marketing-Sourced Pipeline, Pipeline Velocity, Attribution Model, Sales-Accepted Lead

Attribution Model

An Attribution Model is the rule set that assigns credit for a closed-won deal across marketing and sales touches, and without one, no AI lead-gen ROI claim is defensible in a CFO review.

Related terms: Marketing-Sourced Pipeline, Pipeline Velocity, CAC Payback, Auditable Prompt Log

How Practitioners Use This Vocabulary

The five categories map to five different conversations. Prompting Foundations is a conversation between marketing ops and a demand-gen manager about reproducibility. Qualification and Scoring is a conversation between marketing and sales about what counts as a good lead. Data and Enrichment is a conversation with a privacy officer. Governance and Compliance is a conversation with legal and security. Pipeline Proof is a conversation with the CFO. Naming the concept correctly keeps each conversation on its own topic instead of collapsing into a generic argument about AI.

The Starr Conspiracy uses this vocabulary when we help B2B tech teams operationalize AI in demand gen, because most stalled programs are stuck on a definitional gap, not a technology gap. Every week you argue about definitions is a week you do not ship pipeline.

What Good Looks Like

Illustrative implementation patterns that use the vocabulary as intended, without invented outcomes:

1. A revenue ops team replaces static point-based scoring rules with an LLM-Assisted Qualification layer that reads job-post language, then measures lift in Sales-Accepted Lead rate against the prior quarter as its only success metric.
2. A security-conscious team blocks public ChatGPT for prospecting after documenting PII Exposure in Prompts risk, then rebuilds the workflow as a Governed AI Workflow on an enterprise-tenant model with an Auditable Prompt Log.
3. A demand-gen team codifies its outreach templates as ICP Prompt Scaffolding in a shared, versioned library, so any SDR produces consistent, brand-approved messaging that passes blind sales review.

Related Questions

What is the difference between AI Lead Scoring and Predictive Lead Scoring?

Predictive Lead Scoring uses supervised machine learning on structured features. AI Lead Scoring is the broader term that includes predictive ML plus LLM-based classification of unstructured signals. Most mature B2B teams run both in tandem.

Do I need a Governed AI Workflow before I can prove ROI?

Yes. Without a Governed AI Workflow and an Auditable Prompt Log, you cannot reconstruct which prompts produced which leads, which means you cannot connect AI activity to Marketing-Sourced Pipeline in a way a CFO will accept.

What is the fastest way to reduce Hallucination Risk in Scoring?

Ground the model in retrieved data rather than parametric memory. Feed the LLM verified firmographic and technographic records at prompt time, and require it to cite the source field for every claim it makes about an account. If you cannot explain the score, you cannot ship the workflow.

AI lead generation stalls on vocabulary before it stalls on technology. Teams that adopt a shared, governed language for the concepts above move faster, ship safer, and produce sales-accepted pipeline a CFO will approve.

Before you scale prompts into production, talk to The Starr Conspiracy about building a governed AI lead-gen workflow that produces sales-accepted pipeline. If your AI pilot is stuck between sales and legal, we can help you operationalize the vocabulary above into an approved system.