Loading
One moment.
Platform · release 2026.4
One calibrated model. Every decision. Inside your VPC.
NODES is the intelligence layer and system of action for enterprise decisions, starting with hiring. Thirteen agents across three pillars (Hire & Develop, Operate & Run, Sell & Grow), all on one calibrated model, calibrated against your own production data, not a vendor benchmark. Same model, same proof primitive, different decision.
overview · 3 pillars · 13 agents
01
Hire & Develop
11 agentsScreening · Interview · Sourcing · Ramp · Retention · Attrition · Internal Mobility · Succession · Onboarding · Training · Promotion Readiness
See all eleven
03
Sell & Grow
extensionThe same intelligence layer extends to revenue systems as the substrate learns from each deployment. Sell & Grow runs as pilot-ready decision surfaces on the same substrate: rep fit, account health, deal slippage, and sales coaching. Production proof to date comes from the hiring chain. Named agents appear only where a production deployment exists behind them.
Read more Three outputs
Three outputs. Every decision. Every pillar.
One calibrated model runs every agent. Whatever the decision (a candidate, an HR ticket, a slipping deal, a comp recommendation), the brain produces the same three outputs. Same substrate, same proof primitive, same audit chain. Your team approves.
01 · Fit Score
A 0 to 100 calibrated score on every decision.
The substrate scores every decision against the calibrated pattern for that decision type at your company. A candidate against the top-performer pattern. An HR ticket against the auto-resolvable pattern. A deal against the closed-deal pattern. Same score range, same calibration math, different decision.
02 · Decision Trace
A signed, reproducible explanation on every output.
Every score, every recommendation, every drafted action ships with a Decision Trace. The trace shows what the brain read, which signals drove the result, which dimensions weighted highest, and the bias and adverse-action floor checks that passed before the output surfaced. Sha256-immutable. Replayable on quarterly retrain. Built to withstand challenge: signed, reproducible, and paired with the adverse-action floor, so no adverse action ever executes without a human approver. Last reviewed: July 2026.
download a sample trace03 · Drafted Action
The next step, ready for human approval.
The brain does not stop at scoring. It drafts what comes next, prices the wait, and holds the draft at the gate. Your team approves in the UI they already use. The Decision Trace logs the approval.
proposed_action · pa_8871.json
draftAdvance M. Rojas to structured interview · scheduler hold placed
writes backWorkday · req rv-2412-08831 · native fields · blocked until sign-off
act this weekinterview and licensing check complete by day 9
ramp value$54.35 per day faster to first milestone · published coefficient
illustrative artifact · mirrors live structure · no action runs before approval
Same model. Same three outputs. Sixteen decisions, run by thirteen agents. The brain proposes, your team approves, the trace signs.
Thirteen agents. Three pillars. One substrate.
Every agent is a thin orchestration layer over the same calibrated model. The substrate is shared, the calibration is per-decision. A signal seen in sourcing compounds into the retention stream ninety days later. A signal seen in an HR ticket compounds into the manager intelligence stream a quarter later.
Model spec
nodes-substrate-2026.04. The one model under every agent.
A fine-tuned open-source foundation model, calibrated inside the customer boundary. Weights owned by the customer. No third-party AI in the chain. The same substrate runs every agent across every pillar. Per-pillar calibration changes the weights, not the model.
model.id · nodes-substrate-2026.04
nodes-substrate-2026.04
Fine-tuned on 10,765 hire traces across 215+ locations, validated on 747 real-time-scored agents with 9+ months tenure. Retrained quarterly against post-hire production. Expanding per pillar.
Calibration corpus
4-yr retrospective · p75+ sustained cohort
Dimensions
28+
behavioral · skill · cultural
Inference p50
74ms
in-cluster · customer-vpc · gpu
Fine-tune cadence
Quarterly
Calibrated on rolling 18-mo production
Evidence · post-deployment
What the model did after it shipped.
-
RPA rate gap between sales-keyword + high-scored hires and no-sales + low-scored hires. segment analysis within the 747-agent cohort (n=507 with classifiable keyword and score data) · p=0.0007
-
Keywords that predict production after Bonferroni correction, across 8,181 tested. 30 anti-predictive · median OR 0.749
-
Production lift per day of ramp acceleration, every day faster to first milestone. $54.35/day/person · 47-day median acceleration
-
Of RPA-winning hires who would have been eliminated by applying all 6 standard ATS keyword filters. 80% eliminated by insurance-experience filter alone (n=50 RPA winners)
Reads from the systems you already run.
ATS · HRIS · CRM · assessment · identity
SSO · SCIM · audit stream · terraform
SSO · SCIM · audit stream · terraform
ATSWorkday · SAP SuccessFactors · Greenhouse · iCIMS · Avature · Taleo
HRISUKG · ADP · Workday HCM · Rippling
CRMSalesforce · Microsoft Dynamics · HubSpot
AssessmentsAssessio · Criteria · SHL · Pymetrics
IdentityOkta · Azure AD · Ping · Google Workspace
DataSnowflake · Databricks · BigQuery · redshift
TelemetryDatadog · Splunk · audit stream · siem export
DeployTerraform · Helm · AWS · Azure · GCP
TicketingServiceNow · Zendesk · Jira Service Desk · Freshservice
Calendar / Email metadataGoogle Workspace · Microsoft 365 · Outlook
Comparisons
Why legacy point tools miss what NODES catches.
Keyword ATSes score resumes against strings. Gen-AI chatbots score candidates against a generic corpus. HR service tools route tickets without context. RevOps tools dashboard the past without drafting the next action. NODES scores every decision against your own production data, drafts the action, and writes a Decision Trace your team can audit.
Dimension
Keyword ATS
Gen-AI chatbot
NODES
Calibrated against
Recruiter-authored filters
Public internet corpus
Your 4-yr production data10,765 agents · p75+ cohort
Decision evidence
String-match log
LLM rationale, non-auditable
Signed decision tracereproducible · customer-owned
Data residency
Vendor cloud
Third-party AI in the chain
Customer VPC · zero egresssingle-tenant · SOC 2 Type II
Post-hire coverage
None
None
13 agents · same model · 3 pillars
Multi-pillar coverage
Hiring only · siloed
Hiring only · siloed
Hire, operate, run, sell · same model · same trace
Legal review median
Varies
6+ months at regulated enterprises
17 daysat the reference customer · 6 prior rejections
30 minutes · Your data · No pitch deck
See the platform on your data.
A 30-minute working session. NODES backtests against a sample of your production data in your environment, on whichever pillar you want to test. Bring hiring data, HR ticket data, deal data, or comp data. NODES shows the score, the trace, and the drafted action on three real decisions you have already made.
01
Scope a de-identified sample of your data on the pillar you want to test
02
NODES backtests against it in your environment, not in any vendor cloud
03
You see what worked, what did not, and the counterfactual the Decision Trace draws
04
If the evidence holds, scope the complete workflow from trigger through approved action and evidence. One Fortune 500 insurance deployment ran from contract to production in 34 days