Choose one of three approaches. The steps for that option appear below.
Choose an option above. On a phone, swipe the tabs left or right to see all options →
Julius AI groups the estimates and drafts the calculations with inspectable code; people re-derive the arithmetic and own every assumption.
Check, upload, describe file
Group rows, compute measures
Recompute a sample
Quant analyst output check
Use: You
Have ready: share estimates with denominators, periods and coverage notes, the bounded question and permitted-use limits. Upload to Julius AI and describe each column in plain language, including which fields are estimates and which are missing.
Compare counts and key fields against the original file; flag missing or unreadable inputs. Financial amounts and formulas require a finance reviewer; no recommendation authorises spending or a financial commitment. Check your organisation's rules before uploading; this record does not establish that upload is permitted. The vendor page did not list accepted file formats as of the check date; confirm your file loads.
Use: Julius AI
Group estimates by denominator, period and coverage; compute concentration measures within comparable groups only, showing residual-share ranges, assumptions and code.
What to paste into the tool
Task: Share estimates with denominators, periods and coverage notes; Bounded research question and permitted-use limits; Confirmed input spreadsheet with scope notes; Loaded dataset with a written column description; Comparability grouping with reasons. Group the rows in this dataset by denominator, period and coverage note. List which groups are internally comparable and which rows rest on different bases. Do not compute any concentration measure yet. Keep the received originals and earlier results alongside this step output; do not treat carried items as newly generated or checked. Task: Share estimates with denominators, periods and coverage notes; Bounded research question and permitted-use limits; Confirmed input spreadsheet with scope notes; Loaded dataset with a written column description; Comparability grouping with reasons; Draft concentration measures with residual-share ranges, assumptions and the code that produced them. Within each comparable group, compute candidate concentration measures with residual-share ranges. State every residual-share assumption next to the measure. Show the code you ran. Keep the received originals and earlier results alongside this step output; do not treat carried items as newly generated or checked.
Use: You
Compare values, denominators and periods against the original spreadsheet; read the code, recompute a sample by hand, assemble the assessment with residual-share ranges and caveats.
Check scope, source access and permission limits; separate source assertions, inference and what is not known yet. A row grouped as comparable without a matching denominator and period is a mismatch to resolve by hand.
Use: The quantitative analyst
Check each calculation retains its covered market, period and assumptions about the share not accounted for; arithmetic was independently checked; no unsupported market-power conclusion.
The named reviewer confirms the evidence trail and any unresolved items before the brief is treated as complete.
You write every comparability rule and formula; Gemini builds a calculator to cross-check your cells on the same inputs.
Import file into Gemini
Write comparability columns
Write concentration formulas
Build cross-check calculator
Resolve, assemble assessment
Quant analyst output check
Use: You
Have ready: share estimates with denominators, periods and coverage notes, and permitted-use limits. Import into Gemini; check row counts and totals match.
Flag missing or unreadable inputs. Confirm current access.
Use: You
Write same_denominator, same_period, same_coverage and comparable_group yourself, reading the coverage notes row by row.
A row is comparable only when denominator, period and coverage all match; borderline rows stay marked for the quantitative analyst.
Use: You
Write the concentration formulas (for example cumulative top-share sums with residual ranges) over comparable rows, with an assumptions column recording how missing shares are treated.
Financial amounts and formulas require a finance reviewer; keep units and periods from the source rows.
Use: Gemini
Ask Gemini to build a small calculator from your stated formulas and compare its outputs against your formula cells on the same inputs.
What to paste into the tool
Task: Share estimates with denominators, periods and coverage notes; Bounded research question and permitted-use limits; Imported spreadsheet checked against the original; Comparability columns written by hand; Hand-written concentration formulas with assumption notes; Cross-check calculator and a list of any input rows where outputs differ. Build a calculator that applies exactly these formulas to these input columns and shows the result per row. Do not change the formulas. Keep the received originals and earlier results alongside this step output; do not treat carried items as newly generated or checked.
Use: You
Resolve every difference between the calculator and your cells, then assemble the draft assessment with residual-share ranges, caveats and source-linked uncertainty notes.
The calculator is a cross-check, not a result. Separate source assertions, analyst inference and what is not known yet; keep source IDs and cell locations.
Use: The quantitative analyst
Check each measure keeps its covered market, period and assumptions about the unaccounted share; no unsupported market-power conclusion is drawn.
The named reviewer confirms the evidence trail and any unresolved items before the brief is treated as complete.
You upload one prepared file and steer with plain-language questions; Vizly profiles the data and shows the code it generates.
Check file, upload, check parse
Profile and compute measures
Recompute, assemble assessment
Quant analyst output check
Use: You
Have ready: the share-estimate spreadsheet with denominators, periods and coverage notes. Decide whether you are permitted to upload the file; upload to Vizly, which states support for CSV and Excel, and check the parsed columns match.
The vendor states uploaded data is encrypted, isolated and deleted after one hour of inactivity on the free tier; decide with the data owner whether that matches your permission. This record does not establish that upload is permitted. Check row counts and totals against the original.
Use: Vizly
Profile denominators, periods and coverage; compute concentration measures within comparable groups with residual-share ranges, showing the Python or R code.
What to paste into the tool
Task: Share estimates with denominators, periods and coverage notes; Bounded research question and permitted-use limits; Confirmed file and upload-permission decision; Parsed dataset checked against the original; Comparability profile with reasons. Profile this dataset: list the distinct denominators, periods and coverage notes, and identify which rows are not comparable with each other. Do not compute concentration measures yet. Keep the received originals and earlier results alongside this step output; do not treat carried items as newly generated or checked. Task: Share estimates with denominators, periods and coverage notes; Bounded research question and permitted-use limits; Confirmed file and upload-permission decision; Parsed dataset checked against the original; Comparability profile with reasons; Draft measures with the generated code. Within each comparable group, compute candidate concentration measures with residual-share ranges and state each residual-share assumption. Show the code you ran. Keep the received originals and earlier results alongside this step output; do not treat carried items as newly generated or checked.
Use: You
Check the profile against the coverage notes, recompute a sample of measures by hand, then assemble the assessment with residual-share ranges and caveats.
Check row counts against the original spreadsheet; keep source IDs and cell locations; separate source assertions, analyst inference and what is not known yet.
Use: The quantitative analyst
Check each measure keeps its covered market, period and assumptions about the unaccounted share; the finance reviewer checks amounts and formulas; no unsupported market-power conclusion.
The named reviewer confirms the evidence trail and any unresolved items before the brief is treated as complete.
Best when: You want the tool to draft grouping and calculations from shown code
What it does: Julius AI groups and computes; people re-derive and decide
Strong at: The code behind the calculations can be requested and read
Limit: Vendor page did not list accepted file formats at check date
Free plan: Yes, limited (75 one-time + 25 daily credits)
Best when: Every formula written by you, with AI as a cross-check
What it does: You write all formulas; Gemini cross-checks them
Strong at: Comparability and formulas are written by a person
Limit: The free plan has usage limits
Free plan: Yes
Best when: One prepared file and conversational profiling with inspectable code
What it does: Vizly profiles and computes; you re-check by hand
Strong at: Generated Python or R code is visible for inspection
Limit: Free tier allows up to 10 messages a month
Free plan: Yes, limited (10 messages a month)
Before you paste into any AI tool, remove any confidential information. Put it back at the end.
Every calculation retains its covered market, period and residual-share assumptions.
The finance reviewer checks financial amounts and formulas.
Unresolved findings stay visible for the named reviewer.
No unsupported market-power conclusion is drawn.
Check the result against your original inputs. Approve it before sharing or acting.