Internal mobility · Novo Nordisk job architecture

Your people already fit.
Find where.

Redeploy matches every employee to the open position they're best qualified for — with a deterministic, auditable engine that never guesses. Layoffs are expensive. Finding the right role is not.

Drop your roster → see matches How the engine works
  • 100% reproducible — same input, same answer, every run
  • Works on messy data: typos, Danish headers, whatever HR exports
  • Every match ships with a written reason
0% of benchmark people land in the exact right posting
0% find the correct role in the top 3 — even on messy, typo-ridden rosters
0 Median external time-to-hire — avoided entirely
0 Net savings per redeployed employee*

The business case

Firing is the most expensive option.
It just looks like the fastest.

Lay off & rehire

  • ~75%of annual salary — replacement cost
  • 6 mosearch + notice + onboarding lag
  • 30%productivity dip during ramp‑up
  • ≠institutional knowledge walks out the door

≈ DKK 553k per person

Redeploy with Redeploy

  • 45kDKK direct cost per internal move
  • 0 moexternal search — role already open
  • ~0%knowledge loss — same person, new role
  • ✓objective, explainable, works-council ready

≈ DKK 45k per person

*Planning figures based on standard replacement-cost benchmarks (75% of annual salary to replace, 6-month search, 30% ramp-up dip, DKK 45k direct cost per internal move). Put in your own numbers with the savings calculator.

How it works

From a messy export to a matched roster in four steps.

  1. 01

    Drop the file

    CSV or Excel. Any column names — Emp_ID, Medarbejder-ID, Current Position, Stilling… nobody cares what HR called it.

  2. 02

    AI maps your columns

    An LLM reads headers + a few sample rows and proposes the mapping onto the canonical schema. You see and approve it — nothing is guessed silently.

  3. 03

    The engine matches — deterministically

    Hard constraints first (capability, sub‑family, location), then an explainable score: keyword profile, title overlap, description coverage. Same input, same result, every run. No LLM in the matching path.

  4. 04

    Get the roster — with reasons

    Results CSV + Excel, unmatched report, per-person explanations, overrides honoured. Ready for your redeployment committee.

Live demo

Try it with your own file.

Drop a people roster below. We map the columns, run the engine against the real Novo Nordisk job universe, and show you every match — with reasons.

Job universe: … open positions · careers.novonordisk.com · checking…

Drop Excel / CSV here

or click to browse — headers can be anything

Why deterministic matters

An AI that suggests. A rulebook that decides.

Works-council ready

Every assignment is the output of a documented rulebook — constraints, weights, thresholds — not a model's mood. Defensible in front of any committee, union or regulator.

Explainable per person

Every person gets a written breakdown: which candidates passed the constraints, each score component, and why the winner won. No black boxes in people decisions.

Reproducible to the byte

Same inputs in — same outputs out, every single run. Auditors can replay any decision and get the identical answer. Try that with a prompt.

Human overrides, on record

When the committee decides differently, the override is applied — and every exception is flagged in the output. The rulebook stays visible, the exception stays logged.