Glossary of terms
Platform terms
Agent (formerly called "Twin"): a personalised artificial intelligence entity created on the AIsuru platform, capable of interacting conversationally.
Contents: the structured information that makes up an Agent's knowledge base. Each content item contains a question paired with an answer.
Knowledge import: the process of loading and integrating external data into an Agent's knowledge base.
Sharing layout: a customised interface for sharing and interacting with an Agent.
Context: variables that influence the Agent's responses by restricting access to certain contents or simply prioritising some over others.
Roles (previously called "Personas"): in AIsuru, these represent roles or categories of users that you can use to personalise the Agent's responses. The management tab was called "Personas" in earlier versions of the platform.
Author: a user with permissions to create and edit an Agent's contents.
Dynamic intents: a mechanism that lets the Agent recognise the user's intent and trigger specific actions in response.
Invocation: the process of activating an intent or a specific Agent function.
Slot: variable elements within an intent that can be extracted from the user's text.
Functions: specific actions the Agent can perform in response to certain intents.
MCP: Model Context Protocol (MCP) is an open standard for communication between AI assistants and external systems.
Parameters: configurable values that influence the behaviour of functions.
Expert group: an AIsuru feature that lets you create a group of Agents specialised in different areas of expertise, able to collaborate and provide more complete and accurate responses.
Deep thinking: an AIsuru feature that lets the AI remember and recall past interactions with users, improving response personalisation over time.
Open badge: digital certifications that can be issued by the Agent to recognise skills or achievements reached by users.
Timeout: the time interval after which the Agent can take a specific action if it receives no input from the user.
Blob: a default drop-shaped avatar used to visually represent an Agent when no custom image is uploaded.
Virtual space: an interactive 3D environment where users can explore and interact with Agents in virtual or augmented reality.
PaaS (Platform as a Service): the enterprise-focused version of AIsuru, offering greater customisation and control.
Known facts: information the Agent has stored about a specific user, used to personalise future interactions.
Artificial Intelligence terms
Generative AI: a type of artificial intelligence capable of creating original content based on training data.
LLM (Large Language Model): a large language model, such as those used by AIsuru to generate responses.
Fine-tuning: the process of refining a pre-trained language model to adapt it to specific tasks or domains.
Function calling: the AI's ability to invoke external functions to perform specific actions or retrieve information.
RAG (Retrieval-Augmented Generation): a technique that combines information retrieval with text generation to produce more accurate and contextualised responses.
System prompt: baseline instructions given to the AI that define its general behaviour and the context in which it operates.
NLP (Natural Language Processing): the field of AI that deals with interaction between computers and human language.
Embeddings: numerical representations (vectors) of text that allow the system to understand the semantic meaning of questions and contents.
Semantic matching: the process of finding the most similar content by comparing vectors.
Token: a unit of text used by language models to process and generate text.
Technical terms
Webhook: a mechanism that lets the Agent send data in real time to external systems when certain events occur (such as a dynamic intent being called by AIsuru or a function being called by AIsuru with LLM support).
API (Application Programming Interface): a set of protocols and tools that allow AIsuru to integrate with other systems and applications.
VR (Virtual Reality): an immersive virtual environment where you can interact with AIsuru Agents.
AR (Augmented Reality): technology that overlays digital elements, such as Agents, onto the real world through devices like smartphones or smart glasses.
Embed: code used to embed an Agent into a website or external application.
JSONL: a file format used for exporting and importing data in AIsuru, particularly useful for fine-tuning.
Confidence threshold: a parameter that determines the level of certainty the Agent requires before using a specific content item to answer a user's question.
Last updated