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 →
GitHub Copilot extracts formulas and drafts the glossary; you check queries, rule on discrepancies and approve or hold.
Collect queries and official KPI
Draft formulas, glossary and discrepancy
Check every KPI and rule
Use: You
Have ready: all dashboard SQL queries and official KPI definitions in one folder; check the set matches the live dashboard.
Proprietary queries go into an external tool only if your organisation's rules allow it; this record does not establish that upload is permitted.
Use: GitHub Copilot
Ask GitHub Copilot for CSV: KPI, formula, tables, filters, query_ref; check five queries before asking for glossary entries and discrepancies.
What to paste into the tool
Task: Draft formulas (CSV: KPI, formula, tables, filters, query_ref). For each SQL file, explain the query and return a CSV row: KPI name, formula in words, source tables, filters, and the file name as query_ref. Describe only what the code computes. Task: Draft glossary and discrepancy list. Using this CSV, the official definitions and the query files, draft a glossary entry per KPI: formula in words, source tables and filters as given in the CSV, and caveats visible in the query referenced by query_ref. Then list every difference between the CSV formulas and the official definitions. Do not decide which is right.
Use: You
Check glossary wording against queries and official definitions; add missing differences, record which definition is right and why, then approve or hold.
Check five queries against extracted formulas, filters and tables; correct mismatches before glossary drafting. Each KPI's wording matches its query and the owner has ruled on each discrepancy. The glossary does not redefine KPIs Nothing real is opened with the assistant.
Amazon Q Developer drafts formulas and the glossary; you check quoted SQL, rule on differences and approve or hold.
Collect queries and official KPI
Draft a formulas table
Check formulas against the query
Draft glossary entries and definition
Check the glossary
Use: You
Have ready: dashboard SQL queries and official KPI definitions in the IDE; check the set is complete and matches the live dashboard.
Compare the query set with the live dashboard: it must be complete and match. Proprietary queries go into an external tool only if your organisation's rules allow it; this record does not establish that use is permitted.
Use: Amazon Q Developer
Ask Amazon Q Developer for KPI, formula in words, source tables, filters and query_ref, with supporting quoted SQL per row.
What to paste into the tool
Task: Build a formulas table for the KPIs in these SQL query files. One row per KPI with columns KPI, formula in words, source tables, filters, query_ref, and a short quoted snippet of the SQL that supports the formula. Describe only what is in the code. If the code does not make a formula clear, write "not clear from code" in that cell.
Use: You
Compare every formula, table, filter and quoted snippet with its source query; correct or mark rows that do not match.
The formulas come from code only. Fix or mark any row that does not match.
Use: Amazon Q Developer
Ask Amazon Q Developer for glossary entries with caveats and a discrepancy list against official definitions.
What to paste into the tool
Task: Using the checked formulas table and the official KPI definitions below, draft two things. First, a glossary with KPI, formula in words, source tables, filters and known caveats taken from the query code. Second, a discrepancy list: every difference per KPI with a quoted snippet from each side. Do not decide which side is right and do not redefine any KPI. [paste the checked formulas table and the official definitions]
Use: You
Compare entries with queries and official definitions; add missing differences, record your decision on each and approve or hold.
Rule on each discrepancy. The glossary does not redefine KPIs The owner approves only when each entry matches its query and every discrepancy is ruled.
You read SQL and write the formulas; Notion AI drafts glossary wording, then you check entries and rule on differences.
Write the formulas and collect
Draft glossary wording
Check every entry
Use: You
Have ready: SQL, for example 20 KPI queries, and official definitions. Write KPI, formula, tables, filters, query_ref; paste into Notion AI.
Read five queries a second time and check formulas, filters and tables. The formulas come from code only Proprietary definitions go into an external tool only if your organisation's rules allow it; this record does not establish that upload is permitted.
Use: Notion AI
Ask Notion AI for one entry per checked CSV row: formula in words, tables, filters, caveats and differences from official definitions.
What to paste into the tool
Task: Draft glossary and discrepancy list on the page. For each KPI row on this page, draft a glossary entry: formula in words, source tables and filters as given in the row, and caveats taken from the pasted query named in query_ref. Then list differences from the official definitions quoted on this page. Use only page content. Do not decide which is right.
Use: You
Compare entries with CSV rows and queries; reject added meaning, rule on every discrepancy and finish the glossary.
Each KPI's wording matches its query, discrepancies with official definitions are listed, and the owner has ruled on each.
Best when: You want SQL-derived formulas and a first glossary together.
What it does: GitHub Copilot extracts and drafts; you check and rule
Strong at: Extracts formulas, source tables and filters from query files.
Limit: You must check five queries and decide every definition discrepancy.
Free plan: Yes
Best when: You want supporting SQL quotations beside each formula.
What it does: Amazon Q Developer drafts; you check and rule
Strong at: Drafts a formulas table with a quoted SQL snippet per row.
Limit: You must check each quotation and decide every definition discrepancy.
Free plan: Yes, 50 agentic requests per month
Best when: You want to read SQL yourself and delegate glossary wording only.
What it does: Notion AI drafts wording; you read SQL and rule
Strong at: Turns checked formula rows into glossary wording on the same page.
Limit: Paid plan needed; you still decide every discrepancy.
Free plan: Paid plan needed
Before you paste into any AI tool, remove any confidential information. Put it back at the end.
Each KPI's wording matches its query.
Discrepancies with the official definitions are listed.
The KPI owner rules on each discrepancy.
Formulas come from the code only.
Check data-handling rules before placing proprietary queries into any tool.
Check the result against your original inputs. Approve it before sharing or acting.