Illustrative demo dataThis is not a customer result. Rumbo has no customers yet. Every figure here comes from a fictional pilot at a fictional company, generated by the real product so that what you see is exactly what a team would receive.

Pilot report · A mid-market food importer · 6 weeks

What a team receives after six weeks

This is the real report, generated by the real product, from a fictional pilot. Nothing has been tidied: two of four measurable criteria were met, three could not be measured at all, and one risk is still open. That is what a six-week pilot actually looks like.

PAGE 1

Executive result

Needs attention

20 people produced enough work to measure, and the team's average AI Score is up 14 points since they started.

  • 20 of 22 invited people have activated (91%).
  • 154 tasks completed, 7.7 per active learner.
  • 86% of members were active in the last seven days.
  • Against the criteria agreed before the pilot: 2 of 4 measured criteria met, 3 not measurable from this pilot.
  • 8 of 22 learners reported using a workflow on their own work — 14 workflows in total, 5 confirmed by a manager. Self-reported.
  • Estimated 0.1 hours per learner per week on those workflows. Confidence: MEDIUM. Not measured by the platform.

Written from the figures below by arithmetic. No part of a pilot report is generated by a language model.

20/22
Activated
86%
Active weekly
154
Tasks completed
5
Verified workflows
PAGE 2

Capability: baseline versus now

34→48+14

Across 20 learners with a recorded starting point. 2 learners are excluded from this comparison for having no baseline — they are counted in the cohort but not in the change.

EstimatedEvery task in this sample was graded by the offline rubric engine rather than a language model, so its scores are estimates.

SYSTEM MEASUREDObserved by the platform. No assumption involved.

Assessment: baseline and final

Not reported. 20 baselines were taken and 0 finals. Below three learners with both, a paired result is withheld — an average of two people identifies both of them. This is the strongest claim the product can make, and this pilot did not earn it.

PAGE 3

Which skills moved

Skill averages appear once at least three learners have completed work in a skill. Below that an average would identify individuals.

PAGE 4

Did it reach real work

8/22

learners reported using a workflow at work

14

workflows reported · self-reported

5

confirmed by a manager

reconciliation · 4financial commentary · 3invoice review · 3anomaly review · 3management reporting · 1
SELF-REPORTED / MANAGER VERIFIEDReported by learners. Only the verified count was confirmed by a manager.
PAGE 5

Productivity evidence

0.1hmedium confidence

per learner per week, on the workflows reported

14 reported workflows from 8 learners over 6 weeks, 5 of them confirmed by a manager. Each report is a learner's estimate of the same task before and after; reports a manager confirmed count fully, and unconfirmed ones count for less. Nothing here is measured by the platform.

No euro figure appears here because this pilot did not supply an hourly rate. When a buyer enters their own fully-loaded rate the report shows a projection with the arithmetic in full, labelled estimated potential annual value — never ROI generated, never savings delivered.

PAGE 6

Against the criteria, and what is still wrong

Criteria agreed in week zero

2 of 4 measured criteria met, 3 not measurable from this pilot.

  • Activation91%target 75%
  • Learner usefulnessself-reportednot measured · target 4/5
  • Workplace adoptionself-reported36%target 50%
  • AI Score improvementnot measured · target 8points
  • Final assessment completion0%target 65%
  • Weekly participation86%target 60%
  • AI Verification improvementnot measured · target 10points

2 of 4 measured criteria met. Criteria the pilot could not measure are shown as unmeasured rather than dropped — omitting them would turn 2 of 4 into 2 of 2.

Open risks at the end of the pilot

  • The final assessment is still shutThe last week of the pilot, and the final has not been opened.

What this report cannot say

  • 2 learners are excluded from the comparison for having no recorded starting point.
  • No learner has answered a time survey yet, so there is nothing self-reported to show.
  • Capability is measured on platform tasks. Workplace impact is reported by learners and, where marked, confirmed by their manager — it is not measured by the platform and never will be.

What is missing from this sample

A customer’s report also carries a recommended next phase and, where the pilot earned it, a paired baseline-to-final assessment result. This one does not show a paired result, because not enough of the fictional cohort completed both assessments — which is exactly what would happen in a real pilot where the final was opened too late.

Illustrative demo dataThis is not a customer result. Rumbo has no customers yet. Every figure here comes from a fictional pilot at a fictional company, generated by the real product so that what you see is exactly what a team would receive.

See what your own team would look like

Six weeks, twenty people, and you keep the report whether you continue or not.