Agentic AI
AI systems that can plan and take multi-step actions toward a goal rather than just answering a single prompt.
AI roleplay & coaching
The vocabulary of AI-powered practice: personas, scenarios, scorecards and the models behind them.
41 terms
AI agent
An AI system that can plan and take actions toward a goal, such as looking up data or completing tasks, instead of only answering questions.
AI analytics
Using artificial intelligence and machine learning to analyze data, find patterns and make predictions or recommendations. It relies on algorithms and statistical models to surface insights that would be difficult or impossible for people to detect by hand, with the goal of improving business outcomes and informing strategic decisions.
AI coach
An AI system that gives individualized, on-demand feedback on skills such as speaking, pitching or handling objections.
AI continuous coaching
Ongoing AI feedback on practice and real conversations over time, so skills keep improving after a training program ends.
AI grading
Automated scoring of a practice session against a rubric by an AI model, often with written feedback on each criterion.
AI guardrails
Rules and controls that keep an AI system's outputs on-topic, accurate, safe and aligned with company policy.
AI persona
A simulated conversation partner, such as a skeptical CFO, a busy IT director or a frustrated customer, with its own role, goals, objections and personality, used for realistic practice.
AI roleplay
Practicing a real-world conversation, such as a discovery call, a negotiation or a tough feedback talk, with an AI persona that responds dynamically, followed by feedback.
AI roleplay platform
Software that lets teams practice real conversations with AI personas and get scored feedback, with admin tools to build scenarios, assign practice and track results.
AI sales roleplay
Sales roleplay where an AI persona plays the buyer, responds in real time and scores the rep against a rubric. Reps can practice on demand, as often as they need, without waiting for a manager or peer to play the other side.
AI tutor
An AI that teaches a concept, answers questions and checks understanding before a learner moves on to practice.
Benchmarking
Comparing an individual's or team's performance against a standard, top performers or peers.
Build vs. buy
The decision between building a capability in-house, for example an AI roleplay tool on top of an LLM, and buying a purpose-built product from a vendor.
Conversational AI
Technology that lets computers understand and respond to human language in natural back-and-forth dialogue, by text or voice.
Custom rubric
Evaluation criteria written for your own methodology, messaging or role, so AI feedback grades what your organization cares about.
Feedback loop
The cycle of practice, feedback and repetition that turns knowledge into skill.
Generative AI
AI that creates new content (text, audio, images and conversations) based on patterns learned from data.
Hallucination
When an AI model produces information that sounds plausible but is false or unsupported.
Human-in-the-loop
A design where people review, correct or approve AI outputs, keeping final judgment with a person.
Large language model (LLM)
An AI model trained on large amounts of text to understand and generate language. LLMs power most modern conversational AI.
Model Context Protocol (MCP)
An open standard that lets AI assistants connect to external tools and data sources, so an AI can read from and act in the apps a team already uses.
Multi-party roleplay
A roleplay with more than one AI persona in the conversation, such as a buying committee with a champion, a CFO and an IT lead.
Natural language processing (NLP)
The branch of AI that helps computers understand, interpret and generate human language.
Persona difficulty
A setting that controls how challenging an AI persona is, from cooperative to highly skeptical, so practice can scale with the learner's skill.
Practice scenario
A defined situation for roleplay, including context, persona, goal and success criteria, that mirrors a real conversation.
Prompt
The instructions or input given to an AI model that shape its response or behavior.
Prompt engineering
Writing and refining instructions to an AI model to get more accurate, useful output.
Pronunciation dictionary
A list of custom pronunciations for company names, product names and acronyms so an AI voice says them correctly.
Real-time feedback
Coaching delivered during or immediately after a practice session, while the moment is still fresh.
Responsible AI
Practices for developing and using AI that is fair, transparent, secure and accountable, including human oversight and data protection.
Retrieval-augmented generation (RAG)
A technique where an AI model looks up relevant documents, such as your playbook or product docs, before answering, so responses reflect your own content.
Scorecard
A structured evaluation that grades a conversation against defined criteria, giving learners and managers a consistent view of performance.
Sentiment analysis
Using AI to detect emotional tone (positive, negative or neutral) in text or speech.
Simulation-based training
Learning through realistic simulated situations where people can practice and fail safely before it counts.
Skills assessment
A structured evaluation of how well someone can perform a skill, often run before and after training to measure growth.
Speech analytics
Analyzing spoken language for signals such as pacing, filler words, talk time, sentiment and keywords.
Speech-to-text (STT)
Technology that converts spoken audio into written text. Also called automatic speech recognition (ASR).
Text-to-speech (TTS)
Technology that converts written text into natural-sounding spoken audio, used to give AI personas a voice.
Video coaching
Recording yourself practicing and getting feedback on delivery: words, voice and body language.
Voice AI
AI that listens and speaks in a natural-sounding voice, enabling spoken rather than typed practice conversations.