Find the expert, not just the file
Ask "who handles APAC vendor contracts?" and get the person peers actually rely on, with availability and a suggested alternative. Not the org chart owner. The real one.
The Behavioral Context Layer for Enterprise AI
Your AI doesn't know
who to route to.
what to approve or dismiss.
who's the right alternative.
when to assign or escalate.
BehaviorGraph gives enterprise AI the organizational context to find the right person, choose a qualified backup, and escalate safely when documents are not enough.
The harder layer is interpretation, not capture. Who has practical authority. Where trust sits. What is delayed, avoided, or left unowned. That is what we surface.
BehaviorGraph is a pre-trained model, trained on behavioral patterns across hundreds of organizations and calibrated to yours. See how calibration works.
Analyzing organizational context...
Recommended routing
L7
LEG-07 Primary owner
Senior Counsel · Legal · APAC contracts
Relevance0.92
Load38h/wk
CentralityHigh
At capacity
Route to qualified backup
L2
LEG-02 Recommended
Counsel · Legal · Vendor agreements
Relevance0.78
Load22h/wk
Peer endorsedYes
Available
Behavioral context
Decision pathLEG-07 → LEG-02 → Dir. Legal
Escalation SLA24h to Director
Peer reliance0.87: peers consistently route this type of work here
Cross-teamProduct → Legal bridge active
GovernancePolicy-compliant
Related expertise
Audit trail logged · Metadata only · No content accessed
Integrates with your existing stack
Google
Microsoft 365
Slack
Jira
Salesforce
Box
ServiceNow
Workday
SAP
Find the right expert
Route work to the person who actually handles it, not just who the org chart says should.
Surface a qualified alternative
When the primary owner is unavailable, suggest someone with the relevant domain knowledge and capacity, with context on why they fit.
Escalate safely within policy
When documents are not enough, escalate to the right authority within approval bounds, with full context and a complete audit trail.
The missing layer
Enterprise AI is building on two tiers. The third is still missing.
Tier 1
Structural
Org charts, reporting lines, formal titles. Static. Reflects intended structure, not actual behavior.
HRISOrg chartJob titles
HRIS platforms · reporting tools · org chart systems
Tier 2
Informational
Documents, wikis, tickets, and meeting transcripts. Richer context, but still reflects what was written. Not how work actually moves between people.
DocumentsWikisTicketsTranscripts
Enterprise search · knowledge bases · ticketing systems · workflow platforms
Tier 3
Behavioral BehaviorGraph
The real story lives in the gaps: what gets delayed, what is avoided, what is informally escalated, or quietly left unowned. None of it is written down. Understanding this is the interpretation problem that separates AI that routes correctly from AI that guesses.
Tacit knowledgeCollaboration patternsPulse surveysWorkflow dataSilence signals
Research-backed. Privacy by design. Missing from every other AI stack.
Works with your stack
What BehaviorGraph adds to your existing AI.
Most agent deployments are hardcoded for one operator at a time. Cross-team workflows break when they cross reporting lines or informal authority structures. BehaviorGraph is the organizational context layer that lets agents coordinate across teams, people, and authority the way real work happens: with the judgment of someone who actually knows the org.
Multi-agent and human-agent routing
Route between agents, hand off to the right employee, and escalate safely across teams and authority structures. BehaviorGraph provides the organizational context graph for agent-to-agent and agent-to-human decisions via API or MCP.
Escalate with context, not just a ticket
When a decision needs a human, it lands with whoever is actually equipped to handle it, along with the reasoning, alternatives considered, and permission boundaries already attached. No follow-up required.
Every AI decision traceable by design
Every routing recommendation is logged: who was chosen, why, what was skipped, and whether it followed your approval structure. Behavioral governance, not just rule-based guardrails.
Use cases by what you're building
Whether you're building agents, improving enterprise search, or adding context to a search product.
🤖
Agent Builders
For individuals & teams
Give your agents the org context to route work, escalate decisions, and act across teams.
🔍
Enterprise Search
Using search internally
Your search finds documents. Help it surface the right person: the owner, the domain expert, the correct escalation path.
🧩
Search Platforms
Building search for customers
Embed organizational intelligence into your product. Differentiate with people context via API.
Organizational Context Layer
Your org, as it actually operates.
Capturing meetings, tickets, and operational signals is straightforward. What AI cannot do from documents and org charts alone is interpret them. It needs to understand who has practical authority, where trust sits, who should be involved, and where decisions actually stall. BehaviorGraph surfaces that organizational intelligence as runtime context, so agents act on how the org actually works, not how it is supposed to work.
BehaviorGraph is a pre-trained model built on behavioral patterns across hundreds of organizations. It is calibrated to each customer's specific org: their people, their authority structure, their coordination patterns. The output reflects how that organization actually behaves, not a generic benchmark. See how calibration works.
Example: An agent detects an invoice stuck past SLA. The org chart says route to the Finance lead. BehaviorGraph knows she is at 120h/wk, surfaces an available alternative who handles the same contract type, and escalates to the Director if unresolved in 24h.
Signals in
Collaboration patternsCalendars, presence, meetings
Relevance & expertiseWho peers rely on, by domain
Pulse surveysTeam health & readiness
Workflow & handoffsApprovals, escalations, routing
Metadata-only · privacy-first
ENG-01 DO NOT ROUTE
89h/wk · Bridge connector across 3 teams
Risk: if this person burns out, knowledge flow between Eng, Product, and Data breaks
Risk: if this person burns out, knowledge flow between Eng, Product, and Data breaks
SAL-14 OVERLOADED
92h/wk · 4 teams route through here
Action: redistribute to SAL-16 (available, same domain)
Action: redistribute to SAL-16 (available, same domain)
OPS-29 SAFE TO ROUTE
19h/wk · Peer-endorsed for cross-team handoffs
Suggested backup for OPS-25 (overloaded)
Suggested backup for OPS-25 (overloaded)
Large = high centrality
Bridge connector
Overloaded / bottleneck
Peer reliance (thick = strong)
Routing path
Relevance: who peers rely on for this type of work
Plugs into
Enterprise searchAdd people context to every result
Agent orchestrationRoute with real authority & fallback
AI governanceBehavioral context for smarter guardrails
Workflow systemsFix handoff gaps without a reorg
API-first · works with your existing stack
Permission-AwareRespects every access boundary at runtime
Policy-AlignedRoutes within your approval structure, never around it
Human-in-the-LoopKnows when to defer. Escalates to qualified people, not just managers
Full Audit TrailEvery routing decision is traceable and explainable
Metadata-OnlyReads patterns, not message content. Privacy by design.
The Enterprise AI Stack
Enterprise AI systems require three layers to operate inside real organizations.
Model Layer
Foundational reasoning, language, and code models.
Content & Data Layer
Enterprise documents, knowledge bases, and RAG systems.
Organizational Context Layer | BehaviorGraph
Behavioral mapping of networks, expertise, and workflow.
Designed by behavioral science and enterprise knowledge management researchers.
Built on years of research into how organizations actually coordinate, decide, and transfer knowledge.
Built on years of research into how organizations actually coordinate, decide, and transfer knowledge.
Open platform
Access organizational context from anywhere.
APIs, MCP, RAG enrichment, or direct chat. Plug behavioral context into your existing stack however it fits.
POST /v1/agents/route
// Agent requests routing decision
{
"route_to": "LEG-02",
"reason": "primary at capacity",
"escalation": "24h → Director"
}
Agents API + MCP
Expose routing and escalation context to agent platforms via REST API or Model Context Protocol.
GET /v1/search/enrich
// Enrich search with org context
{
"owner": "LEG-07",
"relevance": 0.92,
"backup": "LEG-02"
}
Context AI for search
Plug into your enterprise search to enrich every result with people context, relevance, and routing.
More integration options
POST /v1/chat/ask
// "Why is onboarding slow in Eng?"
{
"bottleneck": "ENG-03",
"load": "42h/wk, 6 teams depend",
"suggestion": "Redistribute to ENG-07"
}
Chat API
Embed conversational org intelligence into your apps. Ask questions, get answers grounded in behavioral context.
POST /v1/rag/enrich
// Inject into your RAG pipeline
{
"chunks": ["...doc results"],
"add_people": true,
"owner": "LEG-07",
"escalation": "Dir. Legal"
}
RAG enrichment
Inject people context into your retrieval pipeline. Re-rank by expertise, add owners, surface escalation paths.
Who should review the APAC contract if Legal is at capacity?
BehaviorGraph
Route to James Park (relevance 0.78, available). Escalate to Dir. Legal if unresolved in 24h.
Governance-compliant · audit trail logged
Governance-compliant · audit trail logged
Live chat
Ask your org questions directly. Real-time answers about who to route to, who's available, and how work flows.
FAQ
Common questions
What is BehaviorGraph?
BehaviorGraph is the behavioral context layer for enterprise AI. It turns collaboration patterns, relevance signals, pulse surveys, and workflow data into a queryable layer that enterprise search, AI agents, and governance platforms can use at runtime to route correctly, surface who actually owns a topic, and escalate with full context.
Why do enterprise AI platforms need behavioral context?
Because AI is not blind to content. It is blind to organizational behavior. It knows who the org chart says should handle a request, but not who colleagues actually trust, what is informally escalated, or where decisions quietly stall. That gap causes agents to misroute, search to surface the wrong people, and escalation to go nowhere.
Does BehaviorGraph read emails or message content?
No. BehaviorGraph operates on metadata only: collaboration patterns, calendar signals, presence data, and pulse survey responses. It never reads message content, emails, or documents. Privacy by design.
How does BehaviorGraph work with AI agents and orchestration platforms?
Your agents call our API or MCP tools before taking action. BehaviorGraph tells them who to route to (based on org chart authority combined with behavioral relevance), who's available as backup (when the primary is overloaded, with qualified alternatives suggested and a human in the loop for confirmation), and when to escalate (with the full decision path and approval chain). You can use this alongside agent platforms like LangChain or CrewAI, or directly through our API without any other platform. We cover the part others don't: the organizational context that makes routing decisions accurate.
Early access
Your AI already reads the docs. Give it the org.
BehaviorGraph works alongside your existing enterprise search, agent platform, or governance stack. If your AI is accurate on content but still misroutes decisions, stalls on escalation, or misses the right person, we should talk.