(Tentative)

Function Reference Model Input/Output Data
SubAIMs JSON Metadata Profiles

Function

The Prompt Creation AIM (PGM-PRC) receives Text input and augments it with structured semantic inputs from User Input, Spatial Reasoning AIMs (Audio and Visual), and, indirectly, Domain Access (PGM-DAC). Its primary function is to synthesise these multimodal and contextual signals into coherent natural language prompts that describe the User’s interaction, clarify intent, and confirm inferred goals.

Internally, PGM-PRC may perform the following operations:

  • Intent Parsing: Analyses raw Text input to extract communicative intent, temporal framing, and referential anchors.
  • Contextual Augmentation: Integrates spatial cues, salience maps, and domain-specific metadata to enrich the prompt’s semantic depth.
  • Goal Inference: Interprets interaction signals and system state to hypothesize the User’s underlying goals or desired outcomes.
  • Prompt Structuring: Assembles a natural language prompt that reflects the User’s intent, contextual constraints, and interaction history.
  • Expressive Alignment: Adapts prompt tone and framing to match the User’s expressive state and personality modulation profile.
  • Output Construction: Produces a Prompt Plan Output for A-User Control and a Prompt Dispatch for Personality Alignment or Rendering, depending on orchestration flow.

The resulting outputs enable A-User Control, Personality Alignment, and Rendering AIMs to operate with full awareness of the User’s communicative intent, supporting expressive coherence, goal-driven orchestration, and context-sensitive interaction.

Reference Model

Figure 6 gives Reference Model of Prompt Creation (PGM-PRC).

Figure 6 – The Reference Model of Prompt Creation (PGM-PRC)

The operation of Prompt Creation (PGM-PRC) may be structured as follows:

  1. Create Token Stream

Definition: A sequential stream of intentionally or un-intentionally user-generated tokens.

Origin: Derived from Text of Context Capture.

Function in PRC:

  • Serves as the raw linguistic substrate for prompt generation
  • Enables intent parsing, referent anchoring, and syntactic alignment
  • Supports continuity across multi-turn interactions
  1. Define Intent Candidates

Definition: A ranked list of possible user intents inferred from the TokenStream and interaction context.

Origin: From Context Capture.

Function in PRC:

  • Guides prompt framing by surfacing likely user goals.
  • Enables disambiguation and clarification prompt.
  • Supports behaviour suggestion and action confirmation.
  1. Develop User State

Definition: A snapshot of the User’s cognitive, emotional, and interactional state (e.g. engaged, confused, passive).

Origin: From Context Capture AIM.

Function in PRC:

  • Modulates prompt tone, pacing, and complexity
  • Enables adaptive phrasing (e.g. “Let’s take it step by step…”)
  • Supports escalation or simplification strategies
  1. Find Resolved Referents

Definition: Mappings from ambiguous tokens (e.g. “that”, “this one”) to specific Items in the scene.

Origin: From Spatial Reasoning (Spatial Guide).

Function in PRC:

  • Enables substitution of vague expressions with precise referents
  • Supports clarification logic if confidence is low
  • Anchors prompts to visual or spatial entities
  1. Create Spatial Primitives

Definition: Abstracted spatial relationships between entities (e.g. “left_of”, “near”, “inside”).

Origin: From Spatial Reasoning (Spatial Guide).

Function in PRC:

  • Enables spatially grounded phrasing (e.g. “the cube to your left”)
  • Supports layout-aware prompt generation
  • Modulates referent salience based on proximity or orientation
  1. Determine Goal Constraints

Definition: Inferred spatial goals derived from user Behaviour and scene context.

Origin: From Spatial Reasoning (Spatial Guide).

Function in PRC:

  • Enables action-oriented prompts (e.g. “Do you want to move the box?”)
  • Supports Behaviour arbitration and prompt sequencing
  • Triggers confirmation or suggestion prompts
  1. Develop Scene Context Tags

Definition: High-level descriptors of the User’s spatial situation (e.g. posture, gaze, gesture alignment).

Origin: From Spatial Reasoning (Spatial Guide).

Function in PC:

  • Modulates prompt relevance and focus
  • Enables adaptive phrasing (e.g. “Since you’re facing zone_2…”)
  • Supports context-sensitive prompt timing

Input/Output Data

Table 15 – Input/Output Data of PGM-PRC

Input Description
Context A structured and time-stamped snapshot representing the initial understanding that the A-User achieves of the environment and of the User posture.
Audio Spatial Guide A Data Type that conveys spatially grounded semantic audio data from the Spatial Reasoning AIM that enables the Prompt Construction AIM to generate context-aware and referentially precise prompts.
Visual Spatial Guide A Data Type that conveys spatially grounded semantic visual data from the Spatial Reasoning AIM that enables the Prompt Construction AIM to generate context-aware and referentially precise prompts.
Visual Spatial Guide Trigger to initiates prompt generation or refinement from PGM-AUC.
Output Description
PC-Prompt Prompt to Basic Knowledge
Prompt Plan Status Prompt readiness, alignment status, and semantic goal framing to PGM-AUA.

SubAIMs

No SubAIMs.

JSON Metadata

https://schemas.mpai.community/PGM1/V1.0/AIMs/PromptCreation.json

Profiles

No Profiles.