JDGMNT Research
Access Review Worksheet v0.1
Released 21 Oct 2026

PURPOSE
Compare two ways of completing the same AI-assisted task. Record people,
barriers, cost inputs, evidence, usefulness criteria, and a bounded decision.

FILES
access-review-worksheet-v0.1.html - Editable offline worksheet; fill fields and save a copy.
access-review-worksheet-v0.1.pdf - Four-page printable blank worksheet.
access-review-completed-example-v0.1.pdf - Four-page illustrative completed example.

HOW TO USE
1. Define the same task and review period for both options.
2. Record access barriers and dated sources. Keep unknowns visible.
3. Enter one month's cost assumptions in a single currency.
4. Calculate costs; assess usefulness separately with a matched trial.
5. Record the decision, owner, reversal conditions, and next review trigger.
6. Save a new copy when inputs or evidence change. Keep previous reviews.

For the editable HTML file, open it in a browser, fill the fields, and select
Save completed copy. The saved HTML file keeps your responses and remains
editable. Changes are not automatically saved; save before closing the page.
Use Print worksheet to print your responses. No data is uploaded.

FORMULAS
Data cost = tasks x MB per task / 1000 x cost per GB.
Pay-per-task total = tasks x price per task + data cost + monthly device allocation.
Subscription total = monthly subscription + data cost + monthly device allocation.
Budget remaining = monthly budget - option total.
Use decimal GB (1000 MB). Add any other known monthly costs separately.

EXAMPLE
Illustrative inputs only: USD10 budget, 100 tasks, USD0.02 per task or an USD8
subscription, 0.25 MB per task, USD2 per GB, and USD0 monthly device allocation.
Data cost is USD0.05. Totals are USD2.05 and USD8.05. These are arithmetic
estimates, not measured provider prices or performance. Subscription quota
must cover the task count; otherwise include overages or a different plan.

LIMITATIONS
The worksheet is not a benchmark, price comparison, or accessibility certification.
It does not establish usefulness, accuracy, latency, reliability, or privacy.
Record tests for those questions separately. Taxes, retries, setup, human review,
app overhead, and equipment purchases need separate treatment where relevant.
The completed example uses no real provider, named study, or observed trial.

TERMS
You may use, copy, and adapt this worksheet, including commercially. Keep JDGMNT attribution and identify changes. Provided as a planning aid without warranty.