The Process

From claim data to confident decisions

Toduba takes your existing HRIS export and two years of claim history, then surfaces a defensible benefit configuration recommendation your leadership can approve.

Step 1

Connect and ingest your workforce data

You upload a standard HRIS export alongside 18 months of anonymised claim records. Toduba parses plan structure, employee cohorts, and utilisation patterns without storing personally identifiable information on its servers.

HRIS Export Headcount, roles, benefit elections Toduba Parse, anonymise, structure cohorts Claim records (18 mo.) Cohort Profiles Anonymised utilisation segments
1a

Upload standard export

Drop your HRIS CSV or connect via Personio or HiBob. No custom IT work required.

1b

Anonymisation pass

All personal identifiers are stripped on ingestion. Toduba operates on cohort-level aggregates, never individual records.

1c

Segment formation

The parser groups employees by age band, family status, role level, and claim frequency to form utilisation cohorts.

Step 2

Model employee preferences and current utilisation gaps

Toduba runs a preference inference model across your cohorts, comparing what each segment actually claimed against what the current plan covers. The output is a ranked list of under-served and over-priced benefit categories, specific to your workforce.

What the model reads

  • Claim frequency by category (dental, vision, mental health, physio)
  • Average claim value vs. plan ceiling by cohort
  • Enrollment rate in optional modules (EAP, flexible allowance, sports)
  • Lapse rates: benefits offered but never claimed across two consecutive years
  • Budget vs. realised spend ratio per plan layer

What the model outputs

Ranked gap list

Categories where a significant cohort exceeds plan ceiling, ordered by total unreimbursed claim value.

Over-investment flags

Benefits consuming budget but claimed by fewer than 8% of eligible employees in both measured years.

Cohort sensitivity map

Which employee segments would be most affected by each potential configuration change.

Step 3

Receive a costed recommendation inside your budget

Toduba generates a concrete benefit mix proposal that stays within your declared per-employee budget. Each category adjustment is shown as a line item with projected claim coverage and expected spend shift.

Current allocation Recommended allocation Health 70% Dental 20% Mental 10% Flexible 0% Total budget: EUR 1,800 / employee / yr Health 55% Dental 28% Mental 20% Flex 7% Same budget. Better coverage.
3a

Review the recommendation

A structured PDF and interactive dashboard show every proposed change, its rationale, and its projected cost delta.

3b

Adjust and re-run

Change the budget ceiling or lock any category and Toduba recalculates the rest in seconds, maintaining feasibility constraints.

3c

Export for leadership

Generate a board-ready summary with supporting data, ready to attach to your annual renewal submission.

Under the hood

Built for European compliance from the start

Toduba was designed to meet GDPR constraints before writing a line of business logic. Data never leaves EU infrastructure, processing is logged, and no individual-level record is retained after the analysis run.

GDPR-ready architecture

Data controller agreements available. All personal identifiers are stripped during ingestion. Processing logs are retained for audit for 90 days.

EU-only data residency

All compute and storage runs on AWS Frankfurt. No data transfer to third-party AI providers. Model inference happens inside the Toduba boundary.

Aggregation threshold

Cohort outputs are suppressed when the underlying segment has fewer than 10 employees, preventing indirect re-identification of individuals.

Integrates with your existing stack

Personio
HiBob
Workday
CSV upload

See the process on your own data

Book a 30-minute pilot session and we will run the ingestion and modeling steps live with an anonymised sample from your organisation.

for questions