Patent Position Mapping
Intelligent Creator Ecosystem & AI-Driven Cross-Platform Orchestration Architecture
Company mapping — Agent infrastructure and platform primitives (not a creator ecosystem)

Amazon Web Services

66.1 / 100
weighted correspondence — hypothesis 94

Publicly documents AgentCore Memory maintaining short- and long-term memory so agents retain historical interactions, preferences and knowledge across sessions, and documents multi-agent collaboration under a supervising agent.

Overview

Public technology position and identified correspondence

AWS is best characterised as infrastructure and platform oriented rather than as a creator ecosystem. Its public documentation nonetheless corresponds closely with several proposed pillars at the primitive level: Bedrock AgentCore documents runtime, memory, identity, gateway and observability services for autonomous agents; AgentCore Memory documents short-term and long-term memory retaining historical interactions, preferences and knowledge across sessions; and Bedrock documents multi-agent collaboration in which specialised agents operate under a supervisor.

Bedrock Knowledge Bases document retrieval grounding including structured and graph-backed retrieval, corresponding with knowledge-graph and semantic-retrieval pillars. Step Functions, EventBridge and Lambda document durable, event-driven workflow orchestration corresponding with autonomous workflow execution and enterprise/API orchestration.

Creator-ecosystem functions — creator intelligence, viewer agents, cross-platform audience intelligence, sponsorship orchestration and content rights lineage — are not identified in reviewed public material for these services. AWS therefore records high correspondence on infrastructure pillars and low correspondence on domain pillars. The strategic relevance is that AWS demonstrates that the proposed architecture's primitives are available as commodity building blocks, which increases the importance of claiming the coordinated domain architecture rather than the primitives.

Evidence register

Primary public evidence supporting each entry

Sources are prioritised in the order: official technical and developer documentation, official architecture and product documentation, official announcements and research, published patent material, then reputable secondary sources. All links open in a new tab.

Sections B, A and E

What they have / what we have

Where the reviewed company material demonstrates a capability more developed than the proposed architecture, that is recorded plainly. Where an element of the proposed architecture cannot be located, the finding is expressed as 'not identified in reviewed public material' — not as an assertion that the company lacks it.

What the company publicly demonstratesWhat the proposed patent architecture containsKey difference / potential patent position
Production-grade agent runtime, cross-session memory, multi-agent collaboration, knowledge grounding and durable workflow orchestration, at global infrastructure scale.The same primitives assembled into a creator/viewer ecosystem with behavioural twins, cross-platform audience intelligence and rights-aware commercial orchestration.AWS demonstrates the primitives and materially exceeds the proposed architecture in operational scale and reliability engineering. The proposed position must be the coordinated domain architecture. Requires further technical/legal analysis.
Security, isolation, compliance and distributed deployment capability well beyond the proposed specification's present description.Privacy-preserving learning described at architectural level.AWS is materially stronger here; this is not an area of proposed differentiation.
Long- and short-term agent memory across sessions.Behavioural memory coupled to a Digital Twin that produces predictions and is corrected by measured outcomes.Coupling of memory to a predictive behavioural model is not identified in reviewed public material.
No creator, viewer, sponsorship, audience-intelligence or content-rights functionality identified in reviewed public material for these services.The whole creator-ecosystem domain layer.Largest domain gap of the seven companies reviewed. Strategic relevance is as a potential platform/licensing counterparty rather than as an architectural analogue.
Architecture correspondence

Proposed stack mapped against Amazon Web Services's reviewed public architecture

01 · External Digital Platforms / Enterprise Systems
Any external system reachable through AgentCore Gateway, API integrations and event sources.
02 · Intelligent AI Orchestration Layer
AgentCore Runtime and Step Functions provide orchestration primitives; a domain orchestration layer is not identified in reviewed public material.
03 · Autonomous AI Agents
Bedrock agents and multi-agent collaboration with a supervising agent.
04 · Digital Twins
Not identified in reviewed public material as Digital Twins in the behavioural sense.
05 · Persistent Behavioural Memory
AgentCore Memory documents short-term and long-term memory across sessions.
06 · Semantic Knowledge Graph
Bedrock Knowledge Bases including graph-backed retrieval.
07 · Predictive / Commercial / Audience / Content / Rights Intelligence
General ML and forecasting services; creator/audience prediction not identified in reviewed public material.
08 · Autonomous or Assisted Workflow Decisions
Agent decisions with tool invocation; human approval patterns documented.
09 · Execution Across External Platforms
Execution through gateways, tools and APIs.
10 · Observed Outcomes
Observability and evaluation tooling documented.
11 · Continual Learning and Model Evolution
Evaluation-driven improvement documented; automatic outcome write-back into a behavioural twin is not identified in reviewed public material.
Correspondence ratings

Matrix entries for this company

AI orchestrationStrong public correspondence
Creator AI AgentNot identified in reviewed public material
Viewer AI AgentNot identified in reviewed public material
Specialised agentsStrong public correspondence
Agent-to-agent communicationModerate public correspondence
Digital TwinsDifferent implementation / objective
Behavioural Digital TwinsNot identified in reviewed public material
Digital Content TwinsNot identified in reviewed public material
Persistent memoryStrong public correspondence
Knowledge graphsStrong public correspondence
Predictive intelligenceModerate public correspondence
Content lifecyclePartial correspondence
AI content transformationPartial correspondence
Multimodal semantic searchStrong public correspondence
Cross-platform audience intelligencePartial correspondence
Commercial orchestrationPartial correspondence
Sponsorship orchestrationNot identified in reviewed public material
Rights managementNot identified in reviewed public material
ProvenancePartial correspondence
AuthenticityPartial correspondence
Intelligent notificationsPartial correspondence
Enterprise / API orchestrationStrong public correspondence
Distributed AIStrong public correspondence
Privacy-preserving AIStrong public correspondence
Explainable AIModerate public correspondence
Continual learningModerate public correspondence
Autonomous workflow executionStrong public correspondence
Feedback-based optimisationModerate public correspondence
Public patent research

Publicly available patent activity indicating technological direction

Patent references are evidence of technological direction only. The existence of another party's patent does not determine infringement or validity, and no such conclusion is drawn here. Searches below are live public queries for counsel to review and refine.

Agent memory and orchestration

Public filings concerning agent execution, memory management and tool invocation.

Open public patent search (Google Patents)
Recommendation and personalisation

Public filings concerning personalisation and recommendation infrastructure.

Open public patent search (Google Patents)
Distributed and privacy-preserving computation

Public filings concerning secure distributed inference and data isolation.

Open public patent search (Google Patents)
Section F — Where the architecture could evolve

Patent evolution opportunities

Drafting considerations for patent counsel arising from this mapping. These are not draft claims and no final claim language is proposed.

High Priority

Explicitly claim the twin as distinct from the memory store

Technology development observed
Managed agent memory is now a standard cloud primitive.
Demonstrated by
AWS (AgentCore Memory), Google (Memory Bank)
Relation to existing architecture
Pillars 4 and 5.
Addressed by present proposed specification?
The distinction is implicit and should be made explicit.
Additional embodiment that may be useful
Describe the Digital Twin as a derived, queryable predictive model maintained from the memory store, capable of answering forward-looking questions the raw memory cannot.
Potential use case
The twin answers 'what will this audience do next quarter', which a memory store cannot.
Possible future claim family
Potential claim family: predictive behavioural model derived from and continuously reconciled with an interaction memory store.
Medium Priority

Federated behavioural learning across independent operators

Technology development observed
Privacy and isolation are strong, but cross-operator federated behavioural learning is not identified in reviewed public material.
Demonstrated by
AWS (by contrast)
Relation to existing architecture
Pillar 16.
Addressed by present proposed specification?
Named but under-specified.
Additional embodiment that may be useful
Describe secure aggregation of behavioural gradients across platform boundaries so ecosystem models improve without centralising raw viewer data.
Potential use case
Cross-platform audience models improve without any platform disclosing viewer-level data.
Possible future claim family
Potential claim family: federated cross-platform behavioural model improvement.
Medium Priority

Behaviour-generated workflows

Technology development observed
Durable workflow orchestration assumes a pre-defined workflow graph.
Demonstrated by
AWS (Step Functions)
Relation to existing architecture
Pillars 1, 7 and 18.
Addressed by present proposed specification?
Partially.
Additional embodiment that may be useful
Describe generation of the workflow definition itself from twin state and prediction, then execution and measurement of that generated workflow.
Potential use case
The orchestrator composes a new distribution workflow for an unusual audience condition.
Possible future claim family
Potential claim family: predictive generation of executable orchestration workflows.
Watch / Future Continuation

Edge and hybrid execution of viewer agents

Technology development observed
Hybrid and edge deployment is well supported at infrastructure level.
Demonstrated by
AWS
Relation to existing architecture
Pillar 16.
Addressed by present proposed specification?
Mentioned.
Additional embodiment that may be useful
Describe viewer-agent inference executing on device with only aggregated updates returned.
Potential use case
Viewer behavioural state never leaves the device.
Possible future claim family
Potential future continuation: on-device viewer twin with aggregated update return.
Scoring

Score derivation for this company

CategoryWeightScoreWeighted
Core architectural correspondence30%8024.0
AI Agent correspondence15%8512.8
Digital Twin / behavioural modelling10%303.0
Memory / knowledge architecture10%909.0
Creator / content / audience correspondence15%203.0
Enterprise / API orchestration10%929.2
Commercial / rights / provenance correspondence5%402.0
Continual learning / predictive intelligence5%623.1
Weighted total66.1

AWS records strong infrastructure correspondence and the lowest creator/content/audience correspondence of the seven, reflecting its platform orientation. The resulting score is below the preliminary hypothesis and the ranking has been adjusted accordingly.

Continue

Other company mappings

AdobeMicrosoftMetaGoogleSalesforceNVIDIA