ChatGPT Weekly Report Rollup: Counts, Rates and Blockers
Start with the counting basis
A ChatGPT weekly report rollup becomes misleading when a cumulative count is treated as new weekly work, percentages are averaged across different batch sizes, or a shared blocker is counted twice. Ask ChatGPT to interpret the inputs first, then wait for your explicit confirmation before calculating.
Define the reporting cutoff, unit, completion rule, denominator and record overlap. Confirm whether scope changed or records were removed, reopened or reclassified. A later report does not automatically resolve a source conflict. Use only approved, nonconfidential data; omit credentials and client identifiers.
A fictional set of weekly reports
Paste the following text into ChatGPT as a practice example. These are editorial fixtures, not actual team reports or a recorded ChatGPT run. R4 supplies the comparison basis; real reports may require clarification first.
R1 = 'Batch A, 2026-10-05 baseline: 40 of 100 records reviewed. At 2026-10-12 cutoff: 55 of the same 100 reviewed. The 55 includes the baseline 40. Open blocker BL-7: waiting for a reference file; no owner or due date supplied.'
R2 = 'Batch B, 2026-10-05 baseline: 10 of 20 records reviewed. At 2026-10-12 cutoff: 15 of the same 20 reviewed. The 15 includes the baseline 10. Open blocker BL-7: waiting for the same reference file; no owner or due date supplied.'
R3 = 'Batch C, 2026-10-12 cutoff: 8 records reviewed. Baseline and assigned total are not supplied. No blocker status supplied.'
R4 = 'User-confirmed basis: all records are counted once with the same reviewed definition. Batches A, B and C contain different records. A and B have unchanged scope and denominator between baseline and cutoff; reviewed totals have no removals, reopened records or reclassification. Dates are Asia/Seoul reporting dates, not event timestamps. BL-7 is one shared blocker affecting A and B.'Calculate the comparable subtotal
A’s current 55 already includes its baseline 40, so its net increase is 55 − 40 = 15, not a cumulative total of 95. B adds 15 − 10 = 5. The comparable A+B subtotal is baseline 50, current 70 and net increase 20.
Compute the combined rate from counts: (55 + 15) / (100 + 20) × 100 = 58.3%, rounded to one decimal place. The simple average of 55% and 75% is 65%, which is wrong for this combined count. This weighting reflects assigned record counts, not effort, difficulty, productivity or schedule health.
On a small screen, swipe the table to see all columns.
| Batch / coverage | Baseline reviewed | Current reviewed / assigned | Net increase | Count-based completion |
|---|---|---|---|---|
| A [R1, R4] | 40 | 55 / 100 | 15 | 55.0% |
| B [R2, R4] | 10 | 15 / 20 | 5 | 75.0% |
| C [R3] | Unknown | 8 / Unknown | Unknown | Unknown |
| Comparable A+B subtotal [R1, R2, R4] | 50 | 70 / 120 | 20 | 58.3% |
Keep missing coverage and repeated blockers visible
C has a known current count of eight, but no baseline or assigned total. Exclude C from the comparable A+B subtotal and state that this is a partial subtotal. Never report 78/120: that mixes C’s numerator with a denominator covering only A and B. Missing values stay Unknown.
BL-7 is one confirmed shared blocker affecting A and B. C’s blocker status is unknown, so this does not establish that the entire portfolio has only one blocker. The blocker owner and due date remain Unknown. Questions about those fields are proposed follow-ups, not confirmed assignments.
In other datasets, do not clip a negative difference to zero or call it gross work completed. Reopened records, removals, reclassification and scope changes can make a difference noncomparable or change its meaning.
Use a two-stage prompt and review the draft
The reusable prompt below first extracts source IDs, dates, units, denominators and comparison assumptions, then stops for confirmation. After confirmation, it can calculate confirmed rows while keeping affected unknown rows separate. You can paste a text table; if using a file, choose readable text or structured data. File availability varies.
A concise draft for this confirmed example could read:
As of the 2026-10-12 Asia/Seoul reporting cutoff, the comparable A+B subtotal is 70 of 120 records reviewed (58.3%), up 20 from the 2026-10-05 baseline of 50 [R1, R2, R4]. C separately reports 8 reviewed; its period increase and completion rate are Unknown [R3]. This subtotal does not cover the full portfolio. One known shared blocker, BL-7, affects A and B; owner and due date are Unknown [R1, R2, R4]. C’s blocker status is Unknown [R3].Check the report against its sources
Before sharing, check that cumulative and period values are distinguished, every rate has a matching denominator, and subtotal exclusions are explicit. Recalculate the differences and aggregate rate; check shared blocker identity and every Unknown against the original reports. Keep proposed questions separate from confirmed actions.
OpenAI’s data-analysis guidance recommends clear columns and one record per row, and reviewing assumptions and outputs. Its accuracy guidance also explains that ChatGPT can make factual errors. A polished rollup still needs source checks.
Open the weekly report rollup prompt
Related guides
Official OpenAI references checked on October 15, 2026. R1–R4 are editorial fictional inputs. The example amounts, ratios and blocker counts were independently checked; this is not a ChatGPT run or an analysis of real team data. The sources support data preparation and verification, not a guarantee of a correct rollup.
OpenAI Help Center: Data analysis with ChatGPT · OpenAI Help Center: Does ChatGPT tell the truth? · OpenAI Academy: Analyzing data with ChatGPT
