Use Cases Context layer Demo Pricing Blog Join the waitlist
Integrations

A pre-trained model. Calibrated to your org.

BehaviorGraph is pre-trained on behavioral patterns across hundreds of organizations. To calibrate it to yours, we need to see how your people actually work. That is what the integrations give us — the signal the model uses to tell your org apart from everyone else's.

Read the academic paper (SSRN) Patent pending

Read-only. Metadata only. Revocable at any time. We never read message content, email body, document text, or anything your teams say to each other. Only timestamps, presence signals, and workflow events.

What calibration needs

The pre-trained model already knows what behavioral patterns look like. Your data tells it which ones are true for your org.

Usually enough to calibrate For most orgs, these two sources give the model enough signal to produce meaningful results. We'll tell you after the first run if more would sharpen the picture.
Google / Outlook Calendar
Calibrates: coordination and network edges
  • Who actually coordinates with whom, and how often
  • Which cross-team relationships are real vs. nominal
  • Availability patterns and working rhythms
Workday / HRIS
Calibrates: structure and authority context
  • Who people are, what they do, who they report to
  • Formal vs. informal authority structures
  • How the org has changed over time
More signal means a more calibrated model.
These are optional — we'll tell you what's worth connecting after we see your first results
See what else we can use
Enrichment — validates expertise and trust If you already use one of these platforms, connecting it lets us validate who actually knows what, and who people trust — not just who they meet with.
AWS Azure Google Cloud Snowflake
Already have the data in your cloud?
If your org already pipes behavioral data into a cloud data platform or warehouse, connect us there directly — no additional tool authorizations needed. We read the same metadata signals from wherever they already live.
e.g. AWS S3 / Redshift, Azure Data Lake / Synapse, Google BigQuery, Snowflake
Slack Microsoft Teams
Team communication platform
Channel membership, message timestamps, and presence signals. Adds network edge data between colleagues.
e.g. Slack, Microsoft Teams
ServiceNow
Workflow or ticketing platform
Ticket assignment, status changes, and escalation events. Shows where work actually moves and where it stalls.
e.g. Jira, Linear, ServiceNow
Culture Amp Lattice 15Five Glint
Engagement & feedback platform
Peer feedback, survey responses, engagement scores, and performance patterns — whatever your team already uses to measure how people are doing.
e.g. Culture Amp, Lattice, 15Five, Glint, Viva Insights
Viva Engage Staffbase LumApps
Internal social or comms platform
Community participation, cross-team engagement, and topic-level activity signals from your internal network.
e.g. Viva Engage, Staffbase, LumApps
What we look for in enrichment data

We use these signals to validate two things that the network alone cannot prove: subject matter expertise (who actually knows what, confirmed by peers and managers, not just by job title) and trust signals (who people rely on to give them honest feedback, escalate to, or bring into a sensitive decision). Together these let us make routing recommendations with confidence, not just correlation.

Don't have any of these? No problem. We can collect the signals directly.
LEAD
LEAD ONA Survey

We bring the survey. Your team takes five minutes.

If your org does not use an engagement or feedback platform, LEAD deploys a targeted organizational network analysis survey directly to your team. We ask specific questions designed to surface expertise and trust signals that no passive data source can provide.

  • Who do you go to when you need an expert on a topic?
  • Who do you trust to give you honest feedback on your work?
  • Who would you loop in on a sensitive decision?
  • Who do you think is underutilized relative to their actual knowledge?

These responses are combined with your core integration data to produce a behavioral map that reflects both how people actually coordinate and who they actually trust. Responses are anonymous. Processing is metadata-only.

LEAD ONA Survey · Q3 2026
Who do you go to when you need an expert in Legal or compliance decisions?
Select up to 3 people
J. Park Legal · Senior Counsel
M. Chen Legal · Counsel
R. Okafor Compliance · Director
3 of 5 questions

How calibration works

Join the waitlist. We assess your data, calibrate the model to your org, and give you real results before you decide anything.

1

Tell us your setup

Join the waitlist and we reach out to understand your org's tool stack. We tell you what data the model needs and whether what you have is sufficient to calibrate it.

2

We calibrate and run

We take the pre-trained model and calibrate it to your org's specific patterns — your people, your coordination structure, your authority map. Then we run the first results.

3

First two batches, free

You get the first two rounds of results at no cost. See if the calibration reflects how your org actually works. Decide whether to continue before you commit.

What a calibrated model surfaces

Once calibrated to your org, the model tells you things your org chart and documents cannot. These are specific to how your organization actually operates, not industry benchmarks.

🔴

Overload and bottleneck detection

See which people are carrying disproportionate routing load and where work is stalling before it becomes a problem.

Example: LEG-07 is at 94h/wk. 6 teams route APAC contracts through here. Suggested backup: LEG-02 (22h/wk, same domain).
🌉

Bridge connector risk

People who connect otherwise disconnected teams. If they leave or burn out, collaboration across those teams breaks. Surface them before it happens.

Example: ENG-01 bridges Engineering, Product, and Data. No formal backup. Knowledge flow risk is high.
🧭

Routing accuracy for AI agents

When your agents ask "who should handle this?", BehaviorGraph returns the person peers actually rely on, not just the org chart owner, with availability and a qualified alternative.

Example: Agent routes vendor contract to LEG-07. BehaviorGraph flags overload, returns LEG-02 with 0.78 relevance score.

Read-only access

We never write to your tools. Permissions are scoped to the minimum required.

Metadata only

Timestamps, patterns, and signals. Never message content, documents, or what people say.

Your data, your terms

You own your data. We are a processor, not an owner. Full data processing agreement included.

Revoke any time

Disconnect any integration in one step. Your data is deleted within 30 days of request.

See what the model finds in your org — before you pay.

Join the waitlist. We assess your data, calibrate the model to your org, and deliver the first two result batches free. You evaluate whether the calibration is right before you commit to anything.

Join the waitlist →