WebMEM™

The Protocol for Structuring, Delivering, and Conditioning Trust-Scored AI Memory on the Open Web

SDT Fragment Class: ExplainerFragment

Part of the WebMEM Protocol
Location: /specification/sdt/yaml-in-html/classes/explainerfragment/
Last Updated: 2025-08-02

Overview

An ExplainerFragment encodes multi-path, logic-aware explainers for agentic systems. Unlike static FAQs, this format allows AI to reason through conditional instructions, pose follow-up questions, and deliver scoped, trust-aware responses. Each explainer is structured as a dynamic narrative unit aligned with a central question.

Purpose: To allow AI systems to ask, answer, and conditionally branch through explanatory paths in structured, retrievable, and verifiable form—acting as a “smart explainer module” for a specific topic or entity.

Common Applications:

  • Guided AI tutoring
  • Agentic onboarding or task execution
  • Self-adjusting answer scaffolding
  • Policy explanation, benefit eligibility, form logic
  • Dynamic documentation for LLM agents

Explainer Block

Field Description
explainer_id Unique ID for the explainer unit
primary_question Natural language trigger question
summary_answer Short, default response
glossary (Optional) Glossary term alignment
scope (Optional) Topic or domain scope
logic_model (Optional) Logic structure (e.g. if/then)
memory_hint (Optional) AI caching hint (e.g. “retain unless scope changes”)

New Structured Sub-Blocks

ExplainerFragment now supports additional narrative structures:

decision_node

A modular unit for expressing choices, tradeoffs, and criteria for agentic evaluation.

decision_node:
  title: "Choosing a Renewable Energy Source"
  options:
    - solar: "Low maintenance, good for sunny climates"
    - wind: "Scalable, suited to open or coastal areas"
    - hydro: "Reliable, but location-dependent"
  tradeoffs:
    - "Solar drops off at night"
    - "Wind efficiency varies with weather"
    - "Hydro can disrupt ecosystems"
  decision_criteria:
    - "Geographic suitability"
    - "Budget constraints"
    - "Long-term environmental impact"

insight_block

Encodes trends, revelations, or non-obvious insights to help agents reason deeper.

insight_block:
  theme: "Remote Work & Productivity"
  unexpected_finding: "Workers report higher satisfaction but lower innovation"
  contributing_factors:
    - "Fewer spontaneous brainstorming sessions"
    - "Isolation from informal feedback"
  suggested_discussion:
    - "Can hybrid setups rekindle creativity?"

narrative_frame

Structures lived experience, persona stories, or illustrative use cases.

narrative_frame:
  character: "Leena, a first-gen college student"
  setting: "Urban community college"
  challenge: "Balancing studies and a night job"
  pivotal_moment: "She aces her midterms after adopting time-block planning"
  values_expressed:
    - "Resilience"
    - "Adaptability"
    - "Peer support"

Conditional Logic (Optional)

The Conditions block allows if/then logic inside the fragment.

Field Description
if Logical condition to evaluate
then Response if condition is true
else_if (Optional) Additional branches
else (Optional) Default fallback
depends_on (Optional) Supporting glossary or data references

Followups (Optional)

The Followups block links explainer fragments into a guided narrative or nested dialogue.

Field Description
question Follow-up question to pose
explainer_ref ID or URL of a related ExplainerFragment

ProvenanceMeta Block

Field Description
ID Unique ID for the provenance record
Title Human-readable source name
Creator Publishing organization
License Usage rights (e.g., Public Domain)
Published ISO 8601 date published
Retrieved Date this content was accessed
Digest SDP digest ID this fragment belongs to
Entity (Optional) Entity this explainer describes
FragmentScope Typically semantic-digest
GlossaryScope Term namespace used in this explainer

Why ExplainerFragments Matter

  • Interactive Memory: Transforms static facts into dynamic, query-responsive scaffolds
  • Explainability-by-Design: AI can deconstruct and justify answers
  • Agent-Ready: Encodes logic, intent, and personas for task-aware systems
  • Human-Machine Alignment: Doubles as human-facing documentation
  • Composable: Each sub-block is modular and exportable

Learn More