The AI sales vocabulary, defined

An AI sales agent is software that autonomously runs sales conversations, answering questions, qualifying buyers, and booking meetings, rather than following a scripted decision tree.

Unlike chatbots, AI sales agents ground their answers in a company's approved knowledge, adapt to what each buyer says, and complete multi-step jobs like scheduling and CRM updates. They typically operate across website chat, WhatsApp, and email.

AI sales role-play is practice conversation between a salesperson and an AI simulating a realistic buyer, with configurable personas, industries, and objection styles, scored against a framework like SPIN or MEDDIC.

It lets new reps make their early mistakes in rehearsal instead of on live pipeline, and gives managers comparable scores across a cohort.

Answer engine optimization is the practice of structuring content so AI assistants and answer engines can find, understand, and cite it when generating answers.

It favors direct-answer formatting, clean HTML that crawlers can actually read (a server-rendered page, not an empty JavaScript shell), schema markup, and genuine definitional clarity. AEO matters because a growing share of buyer research now happens inside AI assistants rather than on search results pages.

Call scoring is the structured evaluation of a sales call against defined criteria, discovery quality, playbook adherence, objection handling, next-step clarity.

AI-based scoring reviews every call rather than the sample a manager has time for, turning coaching from memory-based feedback into evidence-based feedback.

A chatbot is a program that responds to messages using predefined rules, scripts, or decision trees.

Chatbots handle predictable, repetitive queries well but fail when a conversation leaves the script, which is why they route so many conversations to "let me connect you with an agent." The distinction between chatbots and AI sales agents is the difference between following a flowchart and conducting a conversation.

Conversational commerce is selling inside a messaging or chat conversation, answering product questions, recommending items, and completing purchase actions like add-to-cart without leaving the thread.

It compresses the distance between question and purchase, and it's most advanced in markets where messaging apps are the default way customers contact businesses.

Conversational qualification is the practice of qualifying a lead inside a natural dialogue instead of through a form.

The system asks discovery questions, adapts follow-ups based on answers, and captures needs, budget, authority, and timeline as structured data. It typically produces higher completion rates than forms because prospects experience it as help rather than data entry.

A grounded answer is an AI response generated strictly from an approved source of information, company documents, product catalogs, policies, rather than from the model's general training data.

Grounding is what keeps a business AI from inventing prices, features, or policies. Systems achieve it by retrieving relevant approved content first and constraining the answer to it, escalating to a human when confidence is low.

Human handoff is the transfer of a conversation from an AI agent to a person, triggered by low confidence, sensitive topics, high deal value, or an explicit customer request.

A good handoff transfers the full conversation context so the customer never repeats themselves. Escalation design is a primary trust factor in evaluating AI sales tools.

An intent signal is a behavior indicating that a person or company is actively moving toward a purchase, repeat pricing-page visits, comparison searches, demo requests, or high-engagement conversations.

Sales teams use intent signals to time and prioritize outreach toward accounts already leaning in, rather than treating all leads as equal.

In the AI-agent context, a knowledge base is the curated set of company information, documents, web pages, FAQs, catalogs, playbooks, that an agent is permitted to answer from.

The quality of an AI sales agent tracks the quality of its knowledge base, which is why mature deployments include gap detection: flagging questions the agent couldn't confidently answer so the knowledge base improves from real conversations.

Lead scoring is the practice of ranking leads by their likelihood to buy, using signals like stated needs, budget, engagement, and on-site behavior.

Scores are commonly bucketed as hot, warm, or cold to set follow-up priority. Modern scoring combines what a lead says in conversation with what they did on the website before and after identifying themselves.

MEDDIC is a sales qualification framework covering Metrics, Economic buyer, Decision criteria, Decision process, Identify pain, and Champion.

It's used to judge how winnable and how advanced a B2B deal is, especially in mid-market and enterprise sales. Teams embed MEDDIC in playbooks, call scoring, and, increasingly, in AI qualification and coaching tools.

A Meta Business Partner is a company vetted and approved by Meta to provide services on its platforms, including official WhatsApp Business Platform integrations.

Partner status matters in WhatsApp selling because unofficial gateways violate WhatsApp's terms and risk number bans, a material operational risk for teams whose revenue runs through the channel.

A multi-agent AI system divides work among specialized AI agents coordinated toward one outcome, for example, an orchestrator or router agent that classifies intent, plus specialists for knowledge retrieval, lead qualification, appointment booking, and human handoff.

The architecture outperforms single general-purpose bots on complex jobs because each agent is narrower, more controllable, and easier to evaluate.

Speed-to-lead is the time between a prospect's first expression of interest and a company's first meaningful response.

Research consistently shows conversion odds collapse as response time stretches from minutes to hours, because intent decays and competitors answer first. It's the metric AI sales agents most directly improve, since they respond in seconds at any hour.

SPIN is a questioning framework for sales discovery built on four question types: Situation, Problem, Implication, and Need-payoff.

Developed by Neil Rackham from research on 35,000 sales calls, it guides sellers from understanding a buyer's context to making the cost of inaction explicit. AI sales agents like LIA use SPIN to structure discovery questions in live conversations.

Visitor de-anonymization is the process of connecting anonymous website activity to an identified person or company.

Person-level de-anonymization happens when a visitor shares their details and their prior browsing history joins their profile; company-level de-anonymization infers which business a visitor works for from network signals. Consent and disclosure practices are central to doing this responsibly.

Visitor intelligence is the tracking and interpretation of website visitor behavior, pages viewed, time spent, repeat visits, content engaged, to identify buying interest before and after a visitor identifies themselves.

When a visitor later shares contact details, their anonymous history connects to their profile, so lead prioritization reflects demonstrated behavior rather than form answers alone.

The WhatsApp Business API (via the WhatsApp Business Platform) is Meta's official interface for companies to send and receive WhatsApp messages programmatically at scale.

It's how businesses run verified, multi-agent, system-integrated WhatsApp operations, as opposed to the WhatsApp Business app, which is designed for a single small business on a phone. Access runs through Meta Business Partners and official onboarding like Embedded Signup.