Nodes is the People Decision Engine, built on the Internal Decision Engine above the systems your company already runs. It replays past decisions against measured outcomes, proposes a workflow with a signed Decision Trace, and leaves the final call with a named human.
Enterprise talent at one Fortune 500 insurance carrier is the only current production-proven domain. The evidence on this page covers recruitment screening and ramp.
Decision Replay first
The other fourteen decisions begin with a separate replay using past data only. The replay is read-only and does not touch a live decision.
Customer control
Data and model work stay inside a customer-controlled, single-tenant VPC. A named human approves, edits, or declines before action.
Talent decisions from applicant review through development and mobility.
Production evidence currently covers recruiting and ramp in one Fortune 500 insurance deployment. Each new company, role, and remaining talent decision requires its own historical validation.
Decision 01Production evidence · insurance talent
Recruitment & Screening
CHRO, talent acquisition, recruiting
Who should advance to the same structured evaluation?
Evaluation path
Nodes compares governed applicant evidence with measured post-hire outcomes, surfaces likelihood, uncertainty, and counterevidence, and records the review in a Decision Trace. A new company or role is validated on its own history first.
Named human
A named recruiter decides who advances.
Scoped production evidence
In an exact n=50 reference subset at one Fortune 500 insurance carrier, a retrospective counterfactual found that the industry-experience filter would have excluded 80% of eventual top performers. The rate does not transfer to another company or role.
Nodes compares governed ramp milestones with validated trajectories, proposes the next support workflow, and attaches the evidence for manager review. A new company or role is validated on its own history first.
Named human
A named manager approves, edits, or declines the intervention.
Scoped production evidence
At one Fortune 500 insurance carrier, median time to the first production milestone moved from 109 days to 62 days, a 47-day reduction.
Which internal role moves should an employee and manager review?
Evaluation path
Replay completed moves against measured ramp and performance outcomes to test whether approved skills, project, and learning evidence reveals useful adjacencies.
Named human
Employee opt-in and a named manager are required before a move is proposed.
Who is ready to enter succession review for a critical role?
Evaluation path
Replay prior transitions and development histories against measured role outcomes. Define role-specific evidence before any readiness view is proposed.
Named human
A named talent committee or executive owner makes the succession decision.
CHRO, HR business partners, learning and development
Where would manager coaching merit review?
Evaluation path
Compare team outcomes across similar operating contexts and test whether observed variation remains after relevant context is considered. Return evidence and counterevidence for review.
Named human
A named HR business partner decides whether and how to coach.
Test approved cohort-level signals against completed departures and stays. Define the permitted data, outcome, and review window before any shadow evaluation begins.
Named human
A named HR or business leader decides whether to start a conversation. Nodes does not issue an individual attrition verdict.
Operational decisions bounded by policy, permissions, and evidence.
These decision patterns are available for evaluation. Each starts with its own historical data, policy boundary, measured outcome, and named approval path.
Decision 07Decision Replay first
HR Service Delivery
CHRO, people operations, HR shared services
Which service requests can be drafted from approved policy?
Evaluation path
Replay resolved tickets against the policy version, employee context, escalation path, and resolution that applied. Identify which request classes support drafting and which require escalation.
Named human
A named HR owner approves the response and escalation rules.
What roles, capacity, and timing belong in the workforce plan?
Evaluation path
Backtest historical forecasts against actual hiring, attrition, ramp, and business demand. Measure whether additional inputs improve the current planning baseline.
Named human
A named finance or workforce-planning leader approves the plan.
Review completed pay decisions against approved policy, role bands, performance evidence, and group outcomes. Return exceptions with their source evidence and uncertainty.
Named human
A named total-rewards owner makes the compensation decision.
What evidence and approvals belong with an employee-relations action?
Evaluation path
Review closed cases to define the permitted evidence, policy version, documentation, and approval path. Test trace completeness before any production workflow is proposed.
Named human
A named employee-relations or legal owner authorizes every consequential action.
Replay completed hiring, promotion, and compensation decisions against group outcomes, sample adequacy, and approved validity criteria. Surface findings with their limitations.
Named human
Named HR and legal owners decide the review and any response.
Chief AI officer, CIO, compliance, general counsel
Which AI-assisted decisions lack traceable evidence or required approval?
Evaluation path
Inventory one completed decision type and reconstruct its inputs, model and policy versions, human input, and measured outcome. Return gaps in Decision Trace coverage.
Named human
A named governance owner sets the policy and approves any remediation.
Revenue decisions evaluated against customer-owned outcomes.
Insurance talent evidence does not transfer to sales, deal, or account decisions. Each program starts with a separate Decision Replay and stays read-only until the agreed evidence bar is met.
Decision 13Decision Replay first
Sales Productivity & Top Performer Prediction
CRO, sales operations, revenue operations
Which coaching action should a sales manager review?
Evaluation path
Compare completed rep histories with measured sales outcomes inside one role and segment. Return coaching hypotheses, uncertainty, and counterevidence for shadow review.
Replay closed and slipped deals against stage history and approved interaction metadata. Return candidate recovery triggers with evidence and limitations.
Named human
A named sales leader decides whether to use the recovery plan.
Which ramp support should a sales manager consider?
Evaluation path
Compare completed ramp cohorts with defined productivity milestones. Test candidate interventions against the current baseline before a live workflow is proposed.
Named human
A named manager approves, edits, or declines each intervention.
Agree on the decision, outcome, evidence bar, and named owner. Nodes runs a read-only replay. Your team decides whether the result justifies a production scope.