Architecture
RolegacyAI Architecture
RolegacyAI is an operational continuity layer — built to sit between the people who hold institutional knowledge and the organisations that depend on it. When a role holder leaves, retires, or transitions, their accumulated knowledge doesn't have to leave with them.
System Flow
From role activity to successor readiness
Five architectural stages transform everyday role activity into a structured, transferable institutional memory — ready for the next person before the transition happens.
Core Capabilities
Seven capabilities, one continuity outcome
Each capability solves a distinct part of the knowledge continuity problem. Together they produce a role memory that a successor can actually use.
Capture Engine
Ingests role knowledge from documents, meeting notes, workflows, and structured inputs. Converts unstructured activity into a normalised, machine-readable format that downstream layers can classify and store.
Classification Engine
Applies a structured taxonomy — decisions, lessons, workarounds, processes, relationships, risk flags — to every captured memory entry. Makes the institutional record navigable and queryable across any role or team.
Sanitisation Engine
Separates institutional knowledge from personal information before anything is stored. Removes sensitive content and ensures the role memory is safe to retain, share, and audit — without compromising the departing employee's privacy.
Successor Brief Generator
Generates the structured handoff document a successor inherits when they take over a role. Not a generic AI summary — a role-specific brief drawn from validated memory, scored for completeness, and tailored to the incoming person's context.
Confidence & Coverage Scoring
Every memory entry carries a confidence score based on source reliability, corroboration, recency, and human validation. Coverage scoring measures how completely a role's knowledge domains are documented — and surfaces the gaps that need filling.
Human Validation Loop
Every AI-extracted memory entry passes through a human review stage before entering the validated record. Role holders and subject-matter experts can approve, correct, or reject entries — keeping the institutional memory trustworthy at scale.
AI Uplift Measurement
Captures and transfers the AI-enhanced workflows a role holder develops over time — prompt libraries, agent configurations, and productivity patterns. When they leave, their AI uplift is preserved for the successor, not lost with the person.
Architecture Deep Dive
All 20 architecture components
The complete technical reference — every layer of the RolegacyAI architecture, from vector embeddings to agentic workflows.
Capture Engine
How RolegacyAI captures role knowledge from documents, meetings, workflows, and structured inputs — and turns it into structured institutional memory.
Knowledge CaptureClassification Engine
How RolegacyAI classifies captured role knowledge into a structured taxonomy — decisions, lessons, workarounds, processes, relationships, and more.
ArchitectureSanitisation Engine
How RolegacyAI separates institutional knowledge from personal information, removes sensitive content, and ensures role memory is safe to store and share.
SecurityPersonal Role Layer
The technical and conceptual boundary between a role holder's personal knowledge and the institutional knowledge that belongs to the role — and how RolegacyAI manages it.
ArchitectureRAG Engine
How RolegacyAI uses Retrieval Augmented Generation to surface relevant role memories in context — powering the successor brief, AI assistant, and onboarding tools.
ArchitectureConfidence Scoring
How RolegacyAI scores confidence in each memory entry based on source reliability, corroboration, recency, and human validation — and how that score drives review and retrieval.
ScoringCoverage Scoring
How RolegacyAI measures how completely a role's knowledge domains are documented — and uses gap identification to improve successor readiness.
ScoringSuccessor Brief Generator
How RolegacyAI generates the structured handoff document that a successor inherits when they take over a role — and what makes it different from a generic AI summary.
Successor ReadinessRole Risk Score
How RolegacyAI calculates the knowledge continuity risk for any role — and how organisations use it to prioritise where to invest in memory capture before it's too late.
ScoringAI Uplift Measurement
How RolegacyAI measures the efficiency and capability improvements that AI tools create at the role level — making the invisible visible and preserving the improvement for successors.
AI UpliftMulti-Tenant Security
How RolegacyAI isolates role memory between organisations, enforces access controls, and maintains the security posture required for enterprise institutional knowledge.
SecurityHuman Validation Loop
How RolegacyAI involves role holders and subject-matter experts in validating, correcting, and approving AI-extracted memories — keeping the institutional record trustworthy.
ArchitectureKnowledge Graph
How RolegacyAI connects role memories through a structured knowledge graph — enabling relationship traversal, pattern discovery, and richer context for successor briefing.
ArchitectureAgentic Workflows
How autonomous agents can use RolegacyAI role memory to complete onboarding tasks, answer successor questions, coordinate transitions, and fill knowledge gaps.
Workflow IntelligenceDeduplication Engine
How RolegacyAI detects, merges, and reconciles duplicate or near-duplicate role memory entries to keep the institutional memory clean, coherent, and trustworthy.
ArchitectureAPI Integrations
How RolegacyAI connects to the enterprise tools where role knowledge lives — enabling capture, trigger-based workflows, and role memory consumption across the stack.
IntegrationsVector Embeddings
How RolegacyAI encodes role memories as vector embeddings to enable semantic search, similarity detection, and context-aware retrieval across the institutional memory store.
ArchitectureMemory Timeline
How RolegacyAI organises role memories chronologically — providing a navigable history of decisions, transitions, and institutional learning across every holder of a role.
Knowledge CaptureSuccessor Readiness Score
The composite score that measures how ready a role is for transition — combining coverage, confidence, brief completeness, and knowledge transfer progress into a single operational metric.
Successor ReadinessPrompt Engineering
How RolegacyAI approaches prompt engineering for role memory tasks — from capture and classification to successor brief generation — and how prompt libraries are captured and transferred.
AI UpliftEnterprise Grade
Built for institutional trust
Enterprise knowledge is sensitive. Every architecture decision in RolegacyAI reflects that.
Each organisation's role memory is fully isolated. No cross-tenant data access at any layer of the architecture.
Personal information is stripped at the sanitisation layer before any memory entry is stored. The institutional record contains no personal data.
No AI extraction enters the validated record without human review. Role holders and subject-matter experts approve every memory entry.
The core architecture — including the personal role layer boundary and successor brief generation process — is patent pending.
Where It Applies
Built for operationally complex industries
RolegacyAI was designed for environments where role knowledge is hard to replace, transitions are high-risk, and institutional memory sits in people rather than systems.
Asset management roles where years of system knowledge, workaround patterns, and vendor relationships are locked in the outgoing person.
ITSM and workflow configuration teams whose platform knowledge is undocumented and impossible to transfer through a standard handover.
Long-tenure operational roles in rail, ports, and utilities — where a single departure can strand years of project context and vendor relationships.
Client-embedded practitioners who carry relationship history, delivery patterns, and institutional context that doesn't survive rotation.
Site operations, safety procedures, and equipment-specific knowledge held by experienced operators approaching retirement.
Programme and delivery leads whose context — stakeholder maps, risk history, decision rationale — determines the next project's success.
Ready to Build
If your best person left tomorrow, what would remain?
RolegacyAI answers that question — by making role knowledge visible, structured, and transferable before the transition happens. We're working with a small number of design partners.