Decision intelligence for HR

Ask your workforce anything.
Get answers you can defend.

IntRpret answers on a knowledge graph that understands HR. An agent builds and watches your KPIs, explains every change by its cause and models what you could do about it — with every number traceable to its source.

  • Built on a knowledge graph of HR
  • Every number traceable to its source
  • Connects to Keka and Workday

Connected to the systems you already run

  • Keka
  • Workday
  • SAP SuccessFactors
  • Darwinbox
  • Oracle HCM
  • BambooHR
  • greytHR
  • Greenhouse
  • Lever
  • Culture Amp
  • PDFs
  • Excel & CSV
  • SharePoint
  • Google Drive
  • Claude · MCP
How it works

One decision, end to end.

Follow the trail an HR leader walks in IntRpret — from a number that moved to a decision that can be defended. Each station is a part of the product.

01—Understand

It understands HR, not just tables.

IntRpret models your workforce as a knowledge graph: one common data model over every system you run, with the meaning of HR built in. People, positions, org units, pay, performance and events are joined by the relationships that matter, and unstructured records — policies, reviews, exit notes — are linked to the same people and roles, so every answer carries its context.

  • Semantic relationships: reports to, holds, backfills, was promoted into — meaning, not foreign keys
  • Structured and unstructured together: HRIS records beside policies, reviews and exit notes, tied to the same entities
  • One common data model across systems: Keka, Workday, ATS, payroll and surveys resolve to one employee, one position
  • Context-aware: every question is answered knowing which org, period, policy and role apply
02—Measure

KPIs built by an agent, watched and explained.

IntRpret proposes the questions your data can already answer and builds the KPI — definition, population, breakdowns, time basis — checked against your catalogue and your data. Every KPI is then projected forward with its band, and when a number moves the alert arrives with its cause, traced through the graph — and the follow-up, from a ticket to the quarterly review, writes itself.

  • Discovery: every proposed question is proven on your data before you see it
  • Build, modify, delete: when an answer should live on as a KPI, the agent builds it into the system — defined once, checked for duplicates and coverage — and only you can change or retire it, with every version kept
  • Forecast: where each KPI is heading, with its band — alerts fire on the forecast, not just the actual
  • Root cause: contribution by segment, tenure, pay position, manager change — and what was ruled out
  • Counterfactual: replay a past period with a different choice and see the KPI that would have followed — what the delay cost
  • Act and share: raise a ticket to track the fix, share a KPI with its owner, and get the weekly, monthly and quarterly reviews — WBR, MBR, QBR — written automatically from what moved and why
03—Decide

Decisions with their what-ifs, and their receipts.

Turn a cause into options and compare them before you commit — do nothing, correct the band, run a promotion cycle — projected on the same KPI with cost and headcount side by side. Then keep the whole trail: the KPI and its version, the definition on the graph, the cause, the scenario chosen, who saw what under which role, and the source behind every number.

  • What-ifs: scenarios on the KPI you already watch, with the levers HR actually pulls — pay, promotion, hiring, structure
  • Auditability: every decision carries its KPI version, cause analysis, chosen scenario and approver
  • Traceability: any number leads back to its definition on the graph and the source system it came from
  • Role-clamped: one question, the right answer for each person — history and feedback stay with the thread
04—Advise

Twenty years of HR practice, built in.

IntRpret ships with Skills written from two decades of HR work — succession planning, attrition prevention, pay equity, workforce planning, spans and layers, critical roles. They shape how it measures, explains and recommends. Your team can write its own.

  • Playbooks, not prompts: each Skill carries rules, thresholds and worked examples
  • Applied to every answer they’re switched on for — measuring, explaining and recommending the way you do
  • Build your own: write a Skill from a description or a CSV of examples, test it on a real question, and version it as your rules change
  • Share: publish to your tenant’s marketplace, switch Skills on per team, and retire one without losing what it decided

Your next people decision, with the why attached.

Sign in with your work account, or talk to us about connecting your HRIS.