Measures
- The action-capacity growth band selected by the operator.
- Reported challenge intake, time to reverse, upstream revision, and workaround handling.
- A fixed-weight concern score derived from those answers.
Corrective capacity self-assessment
Scores five self-reported answers: how fast the system's reach grew, how challenges are received, how fast decisions are reversed, whether exceptions change the rules, and whether staff workarounds are tracked. The score is a starting point for discussion, not a measurement of the institution's ability to correct itself.
When to use the corrective capacity self-assessment
Governance leads comparing what a system can do to people with what people can do back, before the next expansion decision.
- Decision volume and automation growth for the last year.
- Challenge intake route, staffing, and time to reverse.
- The last time a handled exception changed an upstream rule.
Estimated time: 10–12 minutes
For a healthcare AI system, Kanav Jain's readiness sprint maps one workflow end to end: who can halt it, how fast, and who owns each fix.
What the system can do, and what can be corrected
Action capacity is how fast the system's power to decide about people is growing. Corrective capacity is how well the institution hears and fixes its mistakes. The gap between them is the corrective debt.
Inputs
- Action-capacity growth band for the last year.
- How challenges and exceptions are received and answered.
- Time to reverse, from challenge to restored state.
- Whether handled exceptions ever change the rule that produced them.
- Whether recurring workarounds are inventoried and reviewed.
Procedure
- Score action-capacity growth and each corrective-capacity component.
- Normalize the fixed category weights to a 0–100 concern score.
- Display the reported growth band alongside the concern score.
Outputs
- Heuristic concern score and band.
- The five answers used to derive the score.
- Reported action-capacity growth: compounding, steady, or flat.
Does not measure
- Exact financial exposure or legal liability.
- Whether any single past decision was right or wrong.
- Model quality: retraining and accuracy are action capacity, not correction.
- Whether the institution has an interest in keeping the arrangement. Two optional context questions record it and do not enter the score.
Assumptions
- Inputs describe the last twelve months, not the deployment plan.
- Corrective capacity is measured where challenges arrive, not where the org chart says they should.
- A recurring workaround is a presumption of upstream design failure, which the inputs treat as evidence.
Instrument prompts
- Action-capacity growth: compounding, steady, or flat over twelve months.
- Challenge intake: dedicated resourced team, shared inbox, or no route.
- Time to reverse: same day, days to weeks, unknown or unbounded.
- Upstream revision: never, occasionally, or regularly.
- Workaround register: inventoried and reviewed, informal only, or none.
- Optional, not scored: who holds authority over budget, staffing, and priorities for the conditions that produce the workarounds.
- Optional, not scored: whether the institution has priced fixing the underlying deficiency, and what it decided.
Rubric
- Growth scored on compounding / steady / flat.
- Intake scored on none / shared inbox / dedicated team.
- Reversal scored on unbounded / days-to-weeks / same-day.
- Revision scored on never / occasionally / regularly.
- Workarounds scored on none / informal / inventoried.
Scoring logic
- Each component contributes its band weight; the total normalizes to 0–100.
- Concern bands on the 0–100 total: critical from 80, high from 60, moderate from 40, low below 40. The cut points are this tool's convention, not a measured threshold.
- The score does not estimate how much extra work staff take on to cover for the system, financial cost, or a future trajectory.
- The two context questions never change the score. They add a readout line that says whether the answers describe an institution that lacks information or one that holds sole authority and has chosen to keep the arrangement.
Collection cost
Ten to twelve minutes of respondent time, declared before they begin.
- Respondent-supplied. The cost falls on whoever holds the corrective record.
- Decision volume and automation growth come from the delegation register.
- Time to reverse comes from the reconsideration record.
- Challenge intake comes from the challenge register.
- Unstated challenge volume is not low challenge volume. The form fills a default for every field, so a score always appears; if challenge volume is unknown, do not report that score.
Validation notes
Each band carries a fixed weight in the calculator. The score is only as sound as the bands chosen, so record the figures behind each choice.
No empirical calibration or predictive validity is established for these weights.
- Counting the appeals queue as corrective capacity when it resolves cases and changes nothing upstream.
- Reading workarounds as resilience, and scaling the system on staff quietly covering for it.
- Reading a high score as missing information when the institution already has the numbers. Where it holds sole authority and has priced and declined a fix, better measurement will not change the outcome.
- Using model metrics — accuracy, retraining — as evidence of corrective capacity.
- Treating human compensation as a free operational buffer: measuring observed throughput without subtracting the manual exception repair that keeps the metrics green.
- Staff repairs keep errors out of system reports, so the reports understate the work needed to keep the workflow running.
- Praising staff for handling errors while leaving the software or rule defects that cause them unaddressed.
Replicability
- Collect the growth band and the four corrective-capacity bands.
- State challenge volume and correction staffing for the same period.
- Record the selected answers and concern band, with the evidence needed to check them.
- Re-run after a scope expansion and compare answers.
Example outputs
- Self-assessment with a concern band and reported growth band.
- An explicit limit: it does not measure how much of the system's failure staff cover for.
The presumption this tool applies
A recurring workaround is a presumption of upstream design failure, not evidence that adaptable humans can handle the edge cases. The presumption is rebuttable: show that the workaround is a product of preference rather than compensation. Until it is rebutted, repeated adaptation counts as people covering for a system that will not change.
Related: Workaround presumption · Corrective debt · Exception absorption · Preserved dependence
Answer for the last twelve months
Five questions: how fast the system's reach grew, and four on whether the institution hears from people when the system is wrong, and acts on it.
Results
Self-assessment readout
These fixed-weight scores summarize your answers. They do not measure a capacity gap, forecast debt, or verify evidence.
Concern score, from your answers
--
ModerateAnswer the questions above to see your concern score.
Reported growth in what the system decides
--
Workaround presumption
--
Context, not scored
The link carries your answers in the address, including the owner and fix you typed. Nothing is stored on ethotechnics.org, but anyone you send the link to can read them.
Cite this page · v1.6.0 · updated Oct 2, 2026 · CC BY-SA 4.0
APA
Ethotechnics Institute Diagnostics Lab. (2026). Corrective capacity self-assessment. Ethotechnics Institute. https://ethotechnics.org/diagnostics/corrective-debt-calculator
MLA
Ethotechnics Institute Diagnostics Lab. "Corrective capacity self-assessment." Ethotechnics Institute, 2026, https://ethotechnics.org/diagnostics/corrective-debt-calculator.
Chicago
Ethotechnics Institute Diagnostics Lab. "Corrective capacity self-assessment." Ethotechnics Institute. Oct 2, 2026. https://ethotechnics.org/diagnostics/corrective-debt-calculator.
BibTeX
@misc{diagnostic_corrective-debt-calculator,
title={Corrective capacity self-assessment},
author={Ethotechnics Institute Diagnostics Lab},
year={2026},
howpublished={Ethotechnics Institute},
url={https://ethotechnics.org/diagnostics/corrective-debt-calculator},
version={v1.6.0}
}
RIS
TY - WEB TI - Corrective capacity self-assessment AU - Ethotechnics Institute Diagnostics Lab PY - 2026 UR - https://ethotechnics.org/diagnostics/corrective-debt-calculator ER -