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 →
You paste both lists. Gemini structures, normalises, compares and drafts; you approve each change and confirm every duplicate.
Paste both lists
Structure, normalise
Approve changes
Compare, list duplicates
Confirm duplicates
Draft gap list
Read against records
Use: You
Paste or import both lists into a spreadsheet as two raw lists.
Check both lists are yours; pasted values match your records; condition and ownership notes stay exactly as recorded.
Use: Gemini
Ask Gemini to structure the lists into table A (identifier, name, notes) and table B (scope), then propose corrected identifier formats.
What to paste into the tool
Structure my two pasted lists into table A (identifier, name, notes) and table B (scope entries). Keep every value exactly as pasted. Do not change or invent values. List identifiers that do not follow one consistent format and propose a corrected format for each. Keep my original identifier in its own column. Do not change my values.
Use: You
Check both tables against your pasted records and approve or reject each proposed normalisation; originals stay in their own column.
Check each normalised value is one you approved.
Use: Gemini
Ask Gemini to compare the tables on the approved identifiers into matched, only-here and uncertain lists, then list duplicate candidates.
What to paste into the tool
Compare table A with table B on the approved identifiers. List matched rows, rows only in A, and rows only in B. Put anything unclear in an uncertain list. Do not delete anything. List rows in table A whose approved identifiers differ by one character or spacing, as duplicate candidates. Do not merge or delete anything.
Use: You
Read the uncertain and matched lists; confirm or dismiss each duplicate candidate. Doubtful matches move to the uncertain list, not the bin.
Spot-check five matches against your records; each pair is marked confirmed-duplicate or not-a-duplicate by you.
Use: Gemini
Ask Gemini to draft a gap list from the only-in-B rows with a 'research question' column.
What to paste into the tool
Draft a gap list from the only-in-B rows with a "research question" column, for example "confirm whether variant X belongs to this set". No prices, valuations or purchase suggestions.
Use: You
Read the register, the uncertain list and the gap list against your records. The gap list holds no prices, valuations, authentication claims or purchase suggestions.
Check every scope member is marked held, missing or uncertain with evidence; counts reconcile to the scope. An incomplete scope is not called a complete collection audit.
You paste both lists and keep the checks. Hex structures, compares and drafts; you confirm matches, duplicates and questions.
Paste both lists
Structure, normalise
Check formats
Compare, list duplicates
Reconcile, confirm
Draft gap list
Read against records
Use: You
Paste or upload both lists into a Hex notebook as two raw lists.
Check both lists are yours; the pasted values match your records and the originals are kept. Plan terms beyond the trial offer are not stated.
Use: Hex
Ask Hex to structure the lists into tables A and B as pasted, show identifier formats, then add the normalised column you choose.
What to paste into the tool
Structure my two pasted lists into table A and table B with one identifier column each. Keep every value exactly as pasted. Do not change or invent values. Show me the distinct identifier formats in table A. Then add a normalised column using the format I choose, keeping the original column.
Use: You
Check the format listing and the normalised column against your records; original identifiers stay preserved.
Check each normalised value is one you approved.
Use: Hex
Ask Hex to compare A and B on the normalised identifier into matched, only-in-A and only-in-B lists, then list duplicate candidates.
What to paste into the tool
Compare A and B on the normalised identifier. Show matched, only-in-A and only-in-B as three lists with counts. List rows in A whose normalised identifiers differ by one character or spacing, as duplicate candidates. Do not merge or delete anything.
Use: You
Reconcile the counts (matched plus only-in-A should equal your catalogue size) and confirm or dismiss each duplicate candidate.
Check each pair is marked confirmed-duplicate or not-a-duplicate by you; anything unclear stays in the uncertain list.
Use: Hex
Ask Hex to draft the gap list from the only-in-B rows and your confirmed decisions, with a research-question column.
What to paste into the tool
Draft a gap list from the only-in-B rows and my confirmed decisions, with a research-question column, for example "confirm whether variant X belongs to this set". No prices, valuations or purchase suggestions.
Use: You
Read the register, uncertain list and gap list against your records. Gap list: no prices, valuations, authentication claims or purchase suggestions.
Check every scope member is marked held, missing or uncertain with evidence, and counts reconcile to the scope.
You paste both lists. Coda sets up tables and proposes normalisations, matches, duplicates and gap questions; you confirm each one.
Paste both lists
Set up, propose matches
Set final statuses
Draft gaps table
Read against records
Use: You
Paste both lists into a Coda doc as two raw lists.
Check both lists are yours; the pasted values match your records and the originals are preserved.
Use: Coda
Ask Coda to set up the Catalogue and Scope tables as pasted, propose a normalised identifier, a matching scope row and duplicate flags per row.
What to paste into the tool
Turn my two pasted lists into a Catalogue table (identifier, name, notes) and a Scope table. Keep every value exactly as pasted. Do not change or invent values. Propose a normalised identifier for this row following one consistent format. Keep the original identifier untouched. Propose the scope row that matches this catalogue row, or write NO MATCH. Suggestion only. Flag rows in this table whose normalised identifiers differ by one character or spacing, as duplicate candidates. Do not merge or delete anything.
Use: You
Accept or correct each normalised cell, set the final match status (matched, not in scope, uncertain) for every row, and confirm or dismiss each duplicate.
Check each suspected pair is marked confirmed-duplicate or not-a-duplicate by you; uncertain rows stay in an uncertain view.
Use: Coda
Add a 'Gaps' table linked to Scope rows with no confirmed match, with one drafted research question per row.
What to paste into the tool
Draft one research question per gap row, for example "confirm whether variant X belongs to this set". No prices, valuations or purchase suggestions.
Use: You
Read the tables and the gap list against your records. Gap list: no prices, valuations, authentication claims or purchase suggestions.
Check you edited each question; every scope member is marked held, missing or uncertain with evidence; counts reconcile to the scope.
Best when: You want the tool to normalise and compare while you approve
What it does: Gemini structures and compares; you approve and confirm
Strong at: Plain-language requests in one spreadsheet
Limit: Match correctness unproven; spot-check against your records
Free plan: Yes
Best when: You want to question each pass and reconcile counts
What it does: Hex runs the passes; you reconcile and confirm
Strong at: Built for analytics notebooks with inspectable passes
Limit: 'Trusted answer' is vendor wording; counts need reconciliation
Free plan: Yes, limited (agent on the free plan is a trial)
Best when: You want cell-by-cell proposals you confirm row by row
What it does: Coda proposes per cell; you set every final status
Strong at: AI columns draft inside the tables you control
Limit: Needs internet to open a doc
Free plan: Yes, limited (Docs AI assistant on the free plan is a trial)
Before you paste into any AI tool, remove any confidential information. Put it back at the end.
Never paste credentials or full payment details.
Original identifiers are preserved beside any normalised form.
Anything doubtful goes to an uncertain list, not deleted.
Gaps carry research questions only: no authentication, valuation or purchase recommendation.
Counts reconcile to the declared scope; an incomplete scope is not called a complete collection audit.
Check the result against your original inputs. Approve it before sharing or acting.