The ROI Calculation Method
How the Skills AI savings model works: process-step decomposition, before-and-after comparison, phased rollout, and key assumptions.
Methodology Overview
The ROI calculation follows a rigorous, bottom-up approach. Every savings figure is derived from the same fundamental method: decompose a process into discrete steps, measure each step in cost-per-unit terms, and compare the totals before and after Skills AI implementation.
There are no estimates based on intangible value, employee satisfaction surveys, or theoretical productivity multipliers. The model counts work — person-days, placements, campaigns, cycles — and prices it.
Process-Step Decomposition
Each of the seven modules breaks its process into 10–34 individual steps. Every step is defined by five attributes:
| Attribute | Description | Example |
|---|---|---|
| Type | Process category the step belongs to | Recruiting Planning, Data Integration, Overhead |
| Process Step | Name of the specific activity | Job Description Drafting & Skill Definition |
| Description | Detailed narrative of what happens, who is involved, and why it takes time | "Managers outline outcomes, recruiters rewrite JDs, align skills/levels, and loop through reviews..." |
| Unit of Measurement | The countable work unit | HR Recruiter + Manager blended person-days |
| Total Units per Year | Volume of that work unit annually | 875 person-days |
| Cost per Unit | Fully-loaded cost of one work unit | EUR 500 |
The product of units and cost-per-unit gives the annual operational expenditure for that step. Summing all steps gives the module total.
Before vs. After Comparison
Each module contains two complete process maps:
- Before Skills AI — How the process runs with manual methods, fragmented data, and traditional tooling
- After Skills AI — How the same process runs with AI-generated architectures, automated workflows, and continuous skill signals
The "After" descriptions are specific about what AI does and what humans still do. This is not a generic "AI improves everything" claim — each step explains the mechanism of change: clustering algorithms that replace manual de-duplication, pre-filled review packets that replace evidence gathering, AI-ranked candidate slates that replace full-funnel screening.
Phased Rollout Model
Savings do not materialize overnight. The model accounts for three phases:
Phase 1: Implementation and Setup
Duration: 1 month (configurable). During this phase, the platform is deployed, HRIS connections are established, and initial data is imported. No savings accrue yet.
Phase 2: Architecture Generation and SBO Roll-Out
Duration: 12 months (configurable). The Skills AI generates the job and skill architecture, rolls it out across business units, and activates downstream processes. Savings ramp as more employees are affected by the new operating model. The ramp follows an S-curve: slow initial adoption, accelerating mid-phase, tapering as saturation approaches.
Phase 3: Full Go-Live
After Phase 2, the skills-based organization is fully active. Savings reach their steady-state annual rate. The model calculates cumulative savings through Year 4 to show the time-to-value trajectory.
Adoption Curve
Each module tracks the number of employees affected month by month. For the recruiting module in a 20,000-employee organization, the ramp looks approximately like this:
- Month 1–12 (setup + initial rollout): 0 employees affected
- Month 13: ~1,800 employees affected
- Month 14: ~3,400 employees affected
- Month 24: ~12,345 employees affected
- Month 36: ~16,400 employees affected
- Month 48: ~17,600 employees affected (approaching saturation)
This ramp determines when savings begin and how quickly they accumulate.
Key Assumptions
The reference model uses these configurable parameters:
| Parameter | Default Value | Notes |
|---|---|---|
| Company size | 20,000 employees | All volumes scale with this |
| Number of job role profiles | 600 | Unique role definitions |
| Number of job level profiles | 1,800 | Roles × levels |
| Total job profiles | 2,400 | Sum of role + level profiles |
| Number of courses | 20,000 | Learning portfolio size |
| Number of job requisitions/year | 3,500 | Annual hiring volume |
| Supervisors conducting reviews | 2,500 | Performance review coverage |
| Architecture cycle | 6 years | Full refresh interval |
Person-Day Cost Rates
| Role Type | Cost per Day (EUR) | Annualized (250 days) |
|---|---|---|
| HR / Talent Manager / L&D Manager | 400 | 100,000 |
| Operations / Data Ops | 250 | 62,500 |
| IT Employee | 400 | 100,000 |
| Recruiter | 400 | 100,000 |
| Marketing Employee | 400 | 100,000 |
| Senior Manager / Executive | 600 | 150,000 |
| General Employee | 400 | 100,000 |
| External Consultant | 1,200 | N/A |
| Agency placement (per hire) | 25,000 | N/A |
| Marketing campaign (per block) | 80,000 | N/A |
All parameters are configurable. The model is designed for customer-specific calibration — plug in actual headcount, cost rates, and process volumes for a tailored projection.
How to Read the Module Pages
Each of the seven module pages in this documentation follows a consistent structure:
- Executive Summary — The one-paragraph insight from the study
- Key Numbers — Before/after totals and the savings percentage
- Process Steps Before Skills AI — Full table of every step, with description and cost
- Process Steps After Skills AI — Matching table showing the AI-transformed process
- What Changes — Narrative analysis of the 3–5 biggest transformation levers
- Industry Context — External research and analyst quotes that corroborate the findings
- Bottom Line — The headline savings figure and what it means
Start with the ROI Overview for the aggregate picture, then drill into whichever module is most relevant to your current priorities.
Need help building your Job & Skill Architecture? Talk to our team.
Talk to Hanns