Who they are
Telkomsel is Indonesia’s largest telecommunications company, serving more than 190 million subscribers on the nation’s widest 4G and 5G network with a large, complex organization.
All case studies
02 / Case study · Telkomsel
AI-Powered Competency Mapping Platform
Making competency mapping reviewable.
What they set out to do
To speed up an organizational restructuring, Telkomsel needed 19,000 job positions mapped to standardized competency catalogs, with confidence scores and validation rules instead of manual interpretation.
The problem they were living with
Telkomsel’s teams had to interpret organizational hierarchies and repeatedly cross-reference catalog rules to map job positions. Manual classification created bottlenecks and inconsistent mappings, especially whenever a catalog changed.
What it was costing them. Manual mapping of 19,000 roles meant interpreting complex hierarchies by hand and cross-referencing catalog rules again and again. Every catalog update brought inconsistent classifications and cascading rework.
What changed
- Mapping time reported down from 50 working days to 3
The portfolio records the mapping cycle shrinking from 50 working days to 3 days.
- Batch processing reported five times faster, with live monitoring
Teams follow each run on a processing dashboard and receive export-ready Excel outputs.
- Pre-validated, catalog-compliant outputs ready for implementation
Results stay inside the approved catalog, so teams move to implementation with fewer corrections.
- Confidence and reasoning attached to every match
Reviewers see why a role was matched, not only what it was matched to.
Outcomes as recorded in the QuantumByte enterprise case-study portfolio (2025). The 50-to-3-day figure is the engagement’s own reporting, not an independent audit. The portfolio’s separate percentage claim is omitted because it does not reconcile with those numbers.
Designed solutions
What we designed for Telkomsel’s people.
Validated spreadsheet ingestion
Teams upload the Excel template; the system validates the data and extracts the role information needed for matching.
Job-function analysis
AI extracts job functions and relevant context, while predefined rules adjust the matching weights.
Catalog-only matching
Outputs are limited to entries in the existing competency catalog, so results use the organization’s approved vocabulary.
Confidence and reasoning
Each match carries confidence information and reasoning, so reviewers can judge it quickly.
Batch processing and progress visibility
A processing dashboard shows job status and results while a mapping run is underway.
Review and reporting
Reviewable results and Excel exports feed technical-competence planning and implementation.
The workflow
How the work flows now.
Prepare the input
The team starts with the template and uploads the role spreadsheet.
Validate and interpret
The system checks inputs and extracts the job-function context.
Apply AI and rules
Roles are matched against the approved catalog using predefined constraints and weights.
Review the mapping
Reviewers inspect confidence, reasoning and processing results.
Export the result
The team downloads the mapping for organizational planning and implementation.
The interfaces
A closer look at what the team uses.
Where else this applies
The same approach, in other industries.
| Industry | Fit | Why |
|---|---|---|
| Telecommunications | High | Restructuring across large, evolving role families in network, IT and digital needs rapid, catalog-compliant mapping. |
| Financial services | High | Governance-heavy environments with audit trails and policy rules benefit from constrained, explainable matching. |
| Healthcare | Medium | Standardized clinical and administrative competencies; rule-based constraints reduce compliance risk. |
| Retail and e-commerce | Medium | Large frontline workforces with seasonal hiring benefit from consistent role-to-competency mapping. |
| Government and public sector | Medium | Civil-service reforms need catalog-compliant mapping across many agencies at once. |
Kevin’s part
Kevin led delivery with the Quantum team: the ingestion and review workflow, the catalog-only matching design and the reporting outputs.
How I think about work like this
Working on something like Telkomsel’s?
Tell me where the workflow breaks today and who carries it. I’ll tell you what the smallest useful slice would be, and whether I’m the right person to build it.