Skip to main content

Process-Centric Routing

Process-centric routing is the core innovation of the Naas Ontology system, enabling intelligent AI agent selection based on cognitive processes rather than specific AI models. This approach is grounded in BFO's systematic categorization of processes and realizable entities.

Conceptual Foundation​

Traditional AI routing focuses on brand-based selection:

  • "I want to use Claude"
  • "Switch to GPT-4"
  • "Ask Gemini"

Process-centric routing focuses on cognitive needs:

  • "I need ethical reasoning"
  • "I need truth-seeking analysis"
  • "I need multimodal processing"

BFO Process Framework​

The routing system is built on BFO's process ontology and the 7 Buckets framework:

Cognitive Process Taxonomy​

The Naas Ontology defines a systematic taxonomy of cognitive processes:

@prefix abi: <http://ontology.naas.ai/abi/> .
@prefix bfo: <http://purl.obolibrary.org/obo/> .

# Core cognitive process types
abi:CognitiveProcess a owl:Class ;
    rdfs:subClassOf bfo:BFO_0000015 ; # Process
    rdfs:label "Cognitive Process"@en ;
    skos:definition "A process that involves mental activities such as reasoning, analysis, or decision-making."@en .

# Analytical processes
abi:AnalyticalProcess a owl:Class ;
    rdfs:subClassOf abi:CognitiveProcess ;
    rdfs:label "Analytical Process"@en ;
    skos:definition "A cognitive process focused on breaking down complex information into components for understanding."@en .

abi:TruthSeekingAnalysis a owl:Class ;
    rdfs:subClassOf abi:AnalyticalProcess ;
    rdfs:label "Truth-Seeking Analysis"@en ;
    skos:definition "An analytical process focused on identifying truth, detecting biases, and providing contrarian perspectives."@en .

abi:EthicalReasoning a owl:Class ;
    rdfs:subClassOf abi:AnalyticalProcess ;
    rdfs:label "Ethical Reasoning"@en ;
    skos:definition "An analytical process focused on moral and ethical considerations in decision-making."@en .

# Creative processes
abi:CreativeProcess a owl:Class ;
    rdfs:subClassOf abi:CognitiveProcess ;
    rdfs:label "Creative Process"@en ;
    skos:definition "A cognitive process focused on generating novel ideas, solutions, or content."@en .

abi:CodeGeneration a owl:Class ;
    rdfs:subClassOf abi:CreativeProcess ;
    rdfs:label "Code Generation"@en ;
    skos:definition "A creative process focused on generating computer code and programming solutions."@en .

abi:ContentCreation a owl:Class ;
    rdfs:subClassOf abi:CreativeProcess ;
    rdfs:label "Content Creation"@en ;
    skos:definition "A creative process focused on generating written, visual, or multimedia content."@en .

# Integration processes
abi:IntegrationProcess a owl:Class ;
    rdfs:subClassOf abi:CognitiveProcess ;
    rdfs:label "Integration Process"@en ;
    skos:definition "A cognitive process focused on combining multiple sources of information or modalities."@en .

abi:MultimodalProcessing a owl:Class ;
    rdfs:subClassOf abi:IntegrationProcess ;
    rdfs:label "Multimodal Processing"@en ;
    skos:definition "An integration process that combines information from multiple modalities (text, image, audio, video)."@en .

Agent Capability Mapping​

Each AI agent is characterized by its cognitive capabilities, grounded in BFO's realizable entity framework:

# Agent capability profiles
abi:grok_agent a abi:AIAgent ;
    rdfs:label "Grok Agent"@en ;
    abi:hasCapability abi:truth_seeking_capability ;
    abi:hasPerformanceScore abi:grok_intelligence_score .

abi:truth_seeking_capability a abi:CognitiveCapability ;
    rdfs:subClassOf bfo:BFO_0000017 ; # Realizable Entity
    rdfs:label "Truth-Seeking Capability"@en ;
    abi:realizesProcess abi:TruthSeekingAnalysis ;
    abi:hasCapabilityLevel abi:high_level .

abi:claude_agent a abi:AIAgent ;
    rdfs:label "Claude Agent"@en ;
    abi:hasCapability abi:ethical_reasoning_capability ;
    abi:hasPerformanceScore abi:claude_intelligence_score .

abi:ethical_reasoning_capability a abi:CognitiveCapability ;
    rdfs:label "Ethical Reasoning Capability"@en ;
    abi:realizesProcess abi:EthicalReasoning ;
    abi:hasCapabilityLevel abi:high_level .

abi:gemini_agent a abi:AIAgent ;
    rdfs:label "Gemini Agent"@en ;
    abi:hasCapability abi:multimodal_capability ;
    abi:hasPerformanceScore abi:gemini_intelligence_score .

abi:multimodal_capability a abi:CognitiveCapability ;
    rdfs:label "Multimodal Processing Capability"@en ;
    abi:realizesProcess abi:MultimodalProcessing ;
    abi:hasCapabilityLevel abi:very_high_level .

Routing Algorithm​

The process-centric routing algorithm operates through systematic BFO-based analysis:

1. Request Analysis Phase​

2. Capability Matching Phase​

# Routing rules based on process-capability matching
abi:RoutingRule a owl:Class ;
    rdfs:subClassOf bfo:BFO_0000031 ; # Generically Dependent Continuant
    rdfs:label "Routing Rule"@en ;
    skos:definition "An information entity that specifies how to route requests based on cognitive process requirements."@en .

# Specific routing rules
abi:truth_seeking_routing_rule a abi:RoutingRule ;
    rdfs:label "Truth-Seeking Routing Rule"@en ;
    abi:matchesProcess abi:TruthSeekingAnalysis ;
    abi:recommendsAgent abi:grok_agent ;
    abi:hasConfidenceScore "0.85"^^xsd:decimal .

abi:ethical_reasoning_routing_rule a abi:RoutingRule ;
    rdfs:label "Ethical Reasoning Routing Rule"@en ;
    abi:matchesProcess abi:EthicalReasoning ;
    abi:recommendsAgent abi:claude_agent ;
    abi:hasConfidenceScore "0.90"^^xsd:decimal .

abi:multimodal_routing_rule a abi:RoutingRule ;
    rdfs:label "Multimodal Routing Rule"@en ;
    abi:matchesProcess abi:MultimodalProcessing ;
    abi:recommendsAgent abi:gemini_agent ;
    abi:hasConfidenceScore "0.88"^^xsd:decimal .

3. Performance Optimization Phase​

The routing system considers multiple factors for optimal selection:

# Performance factors
abi:PerformanceFactor a owl:Class ;
    rdfs:subClassOf bfo:BFO_0000019 ; # Quality
    rdfs:label "Performance Factor"@en ;
    skos:definition "A quality that influences the performance of AI agent routing decisions."@en .

# Specific performance factors
abi:ResponseSpeed a owl:Class ;
    rdfs:subClassOf abi:PerformanceFactor ;
    rdfs:label "Response Speed"@en ;
    skos:definition "The speed at which an AI agent can process and respond to requests."@en .

abi:CostEfficiency a owl:Class ;
    rdfs:subClassOf abi:PerformanceFactor ;
    rdfs:label "Cost Efficiency"@en ;
    skos:definition "The cost-effectiveness of using a particular AI agent for a given task."@en .

abi:Availability a owl:Class ;
    rdfs:subClassOf abi:PerformanceFactor ;
    rdfs:label "Availability"@en ;
    skos:definition "The availability and reliability of an AI agent at the time of request."@en .

# Agent performance profiles
abi:grok_agent abi:hasPerformanceFactor [
    a abi:ResponseSpeed ;
    abi:hasValue "medium"^^xsd:string
] , [
    a abi:CostEfficiency ;
    abi:hasValue "high"^^xsd:string
] , [
    a abi:Availability ;
    abi:hasValue "high"^^xsd:string
] .

Routing Decision Process​

The complete routing decision follows this BFO-grounded process:

Implementation Example​

Here's a complete routing scenario implemented in the Naas Ontology:

# User request instance
abi:request_001 a abi:UserRequest ;
    rdfs:label "Bias Analysis Request"@en ;
    abi:hasContent "I need to analyze this claim for potential bias and provide a contrarian perspective" ;
    abi:submittedBy abi:user_john ;
    bfo:BFO_0000108 abi:timestamp_2024_01_15 . # exists at

# Process analysis result
abi:process_analysis_001 a abi:ProcessAnalysis ;
    rdfs:label "Process Analysis for Request 001"@en ;
    abi:analyzesRequest abi:request_001 ;
    abi:identifiesProcess abi:TruthSeekingAnalysis ;
    abi:hasConfidenceScore "0.92"^^xsd:decimal .

# Routing decision
abi:routing_decision_001 a abi:RoutingDecision ;
    rdfs:label "Routing Decision for Request 001"@en ;
    abi:basedOnAnalysis abi:process_analysis_001 ;
    abi:selectsAgent abi:grok_agent ;
    abi:hasRationale "Grok agent has highest truth-seeking capability and contrarian analysis strength" .

# Process execution
abi:truth_seeking_process_001 a abi:TruthSeekingAnalysis ;
    rdfs:label "Truth-Seeking Analysis Process 001"@en ;
    bfo:BFO_0000057 abi:grok_agent ; # has participant
    bfo:BFO_0000066 abi:truth_seeking_capability ; # realizes
    abi:processesRequest abi:request_001 .

Advanced Routing Patterns​

Multi-Agent Collaboration​

For complex requests requiring multiple cognitive processes:

# Complex request requiring multiple processes
abi:complex_request_001 a abi:UserRequest ;
    rdfs:label "Multi-Process Request"@en ;
    abi:hasContent "Analyze the ethical implications of this AI system design and provide technical implementation recommendations" .

# Multiple process identification
abi:process_analysis_002 a abi:ProcessAnalysis ;
    abi:identifiesProcess abi:EthicalReasoning, abi:CodeGeneration ;
    abi:requiresCollaboration "true"^^xsd:boolean .

# Multi-agent routing decision
abi:routing_decision_002 a abi:RoutingDecision ;
    abi:selectsAgent abi:claude_agent, abi:mistral_agent ;
    abi:definesWorkflow abi:sequential_workflow_001 .

abi:sequential_workflow_001 a abi:SequentialWorkflow ;
    abi:firstAgent abi:claude_agent ; # Ethical analysis first
    abi:secondAgent abi:mistral_agent ; # Technical implementation second
    abi:hasHandoffCriteria abi:ethical_analysis_complete .

Context-Aware Routing​

Routing decisions can consider contextual factors:

# Context-aware routing
abi:routing_context_001 a abi:RoutingContext ;
    rdfs:label "Enterprise Context"@en ;
    abi:hasOrganization abi:acme_corp ;
    abi:hasSecurityLevel abi:high_security ;
    abi:hasFoundryAccess "enterprise_management_foundry" .

# Context-influenced routing rule
abi:enterprise_routing_rule a abi:RoutingRule ;
    abi:appliesInContext abi:routing_context_001 ;
    abi:prefersAgent abi:claude_agent ; # Preferred for enterprise contexts
    abi:hasContextualBonus "0.15"^^xsd:decimal .

Benefits of Process-Centric Routing​

1. Cognitive Alignment​

Users think in terms of what they need to accomplish, not which tool to use.

2. Optimal Performance​

Automatic selection of the best-suited agent for each cognitive task.

3. Adaptability​

New agents can be integrated by defining their capabilities, not rewriting routing logic.

4. Transparency​

Clear ontological grounding makes routing decisions explainable and auditable.

5. Consistency​

BFO-based framework ensures systematic and consistent routing across all scenarios.

Process-centric routing transforms AI interaction from tool selection to cognitive need specification, enabling more natural and effective human-AI collaboration grounded in formal ontological principles.