vanellence. Explore the demo

THE VANELLENCE GLOSSARY

Understand the terms.
See the connections.

AI, sales and negotiation clearly explained. A starting point for better questions, informed decisions and your next practical application.

GLOSSARY

Observability

Observability helps investigate AI workflows. Identify evidence about errors, delays and cost while avoiding unnecessary logs of confidential customer content.

Explore the term

GLOSSARY

Open source AI

Open-source AI: understand the OSI definition and assess the actual rights, operating responsibilities and switching options behind an openness claim.

Explore the term

GLOSSARY

Open weight

Open model weights provide access to learned parameters. Explore the options they create and the licensing, quality and operating responsibilities that remain.

Explore the term

GLOSSARY

Orchestration

Orchestration connects AI, tools and approvals into a workflow. In sales, the quality of handovers determines whether it reduces effort or creates new failures.

Explore the term

GLOSSARY

Policy engine

A policy engine evaluates permissions and approvals. It makes sales rules testable, while the application must enforce decisions where each action occurs.

Explore the term

GLOSSARY

Probabilistic

Probabilistic AI uses probabilities. Understand variable answers, the limits of repeated agreement and the evidence needed for a responsible sales decision.

Explore the term

GLOSSARY

Prompt

A useful prompt specifies the task, context and requirements. Sales teams can improve drafts while keeping assumptions and unanswered questions visible.

Explore the term

GLOSSARY

Prompt injection

Prompt injection redirects AI through inputs or external content. Assess the risk in sales research and why limited permissions matter beyond prompt wording.

Explore the term

GLOSSARY

Quantization

Quantization can reduce AI model memory requirements. Business decisions need evidence about output quality, target hardware and the cost of accepted results.

Explore the term

GLOSSARY

Reasoning

Reasoning models process complex tasks through connected steps. Assess when extra processing improves negotiation options and what evidence is needed.

Explore the term

GLOSSARY

Red team

A red team deliberately tests weaknesses in AI applications. Clear scenarios, independent assessment and retesting turn findings into evidence of improvement.

Explore the term

GLOSSARY

Retention

Retention defines which information stays stored and for how long. Sales workflows need a clear purpose, appropriate limits and verified deletion processes.

Explore the term

GLOSSARY

Scalability

AI scalability handles more users and work at an agreed service quality. Capacity, approval bottlenecks and cost per useful result need joint assessment.

Explore the term

GLOSSARY

Self hosting

Self-hosting AI brings responsibility for infrastructure and applications. Assess control, operating capability and full cost for the intended sales workflow.

Explore the term

GLOSSARY

Semantic fidelity

Semantic fidelity preserves statements, conditions and uncertainty in AI summaries. It keeps customer interest, assumptions and commitments distinguishable.

Explore the term

GLOSSARY

Structured output

Structured AI outputs supply defined fields for CRM workflows. Valid formats can contain incorrect values, requiring schema checks and substantive review.

Explore the term

GLOSSARY

Synthetic data

Synthetic data creates targeted AI test cases. Assess how well it represents real work and whether it could reveal information about underlying records.

Explore the term

GLOSSARY

Temperature

Temperature influences output selection in supported AI models. Low settings do not ensure correctness; test configurations against actual task requirements.

Explore the term

GLOSSARY

Tenant isolation

Tenant isolation protects customer boundaries in shared AI applications. Controls must cover storage, search, background work and operational processes.

Explore the term

GLOSSARY

Throughput

Throughput measures work processed over time. Sales teams also need to understand waiting times, output quality and the treatment of urgent requests.

Explore the term

GLOSSARY

Token

Tokens are the processing units of language models. They affect capacity and cost; their price alone cannot establish the cost of a useful business result.

Explore the term

GLOSSARY

Token budget

Token budgets control the resources available to an AI request. Useful limits preserve essential information and make incomplete answers visible to the user.

Explore the term