Match two Excel lists by meaning, not just spelling.

First, exact matches are found inside Excel: names that differ only in capitals, accents, punctuation or a legal form such as Inc. or Corporation are matched for free and never leave your workbook. Then each remaining row is sent with only the columns you ticked, together with its candidates from List B: the whole list when it has up to 200 entries, otherwise the 20 entries closest in spelling. A matching model from TypeSafe AI picks one of those entries or answers No match, with a confidence. It can only choose from your list, so it cannot invent a name. Answers below your threshold (80% unless you change it) are marked Review, and all results are written into new columns next to List A.

Often, yes. Fuzzy merge is free, runs offline, and handles typos and small spelling differences well. It scores how alike two texts look, so it cannot know that AMZN MKTP is Amazon unless you add that pair to a transformation table yourself, and two different businesses with similar names can pass your threshold. If your differences are spelling, use it. If they are abbreviations, renames, bank descriptions or look-alikes where a wrong pair costs money, compare both on your own list.

For a one-off list of a hundred rows, ChatGPT is fine. Microsoft's own Copilot FAQ says it "can sometimes make mistakes" and advises against using it "for decisions in sensitive areas such as finance". AI Fuzzy Match does not use ChatGPT or any chatbot. It works inside Excel, gives every row its own decision, an entry from your list or No match, with a confidence, and writes the results in columns next to your list.

One credit is one AI answer for one row. A normal run uses 1 credit for each List A row the AI looks at, however long List B is; the optional thorough search for long lists costs more per row. Exact matches, repeated rows and skipped rows cost nothing and are never sent. Before every run the add-in shows how many credits it will use, and nothing happens until you click Run.

No. Credits from a pack never expire, and there is no subscription. If you buy another pack, the add-in keeps both keys and uses up the older pack first.

Credit packs are non-refundable once they are delivered. That is why every install comes with 100 free credits: they are the trial, so you can see the results on your own data before you buy. This does not affect your statutory rights as a consumer. The details are in the terms.

Only the columns you select are sent: the values of the columns you tick for each row, the List B entries they are compared with, and your extra rule if you add one. For bank lines that can be the payee text alone. They go to Cloudflare and TypeSafe AI only, not to OpenAI or any chat service. The rest of your workbook is never uploaded, exact matches never leave Excel, your data is not used to train AI models, and we do not store your cell data. If your policy allows no outside service at all, use Power Query's fuzzy merge, which runs offline. The privacy policy has the details.

Excel on Windows, Excel on Mac and Excel on the web. You install it once from Microsoft's add-in store inside Excel.

No. It only adds new result columns: Match, Confidence and Status next to List A, plus any columns you chose to bring back from List B. Rows to review get a light fill. Nothing outside those columns is changed, and Clear results removes only them.

No tool is right on every list. Every row gets an entry from your list or No match, a confidence and a status, and the doubtful rows are marked Review so you check them first. We know of two weak spots. On a reference list of more than 200 names, a complete rename that shares no letters, such as FACEBK for Meta Platforms, can come back No match; Search thoroughly can check those rows against the whole list. And a parent company and its subsidiary with near-identical names can be matched as the same company. The fairest test is your own list, and the first 100 credits are free for exactly that. More in How accurate is AI Fuzzy Match.

Up to 50,000 rows in List A per run, and up to 100,000 rows in List B. For a longer List A, run it in parts: rows that already have a result are skipped unless you choose to redo them. Exact matches are free and instant. The rest takes longer on big lists, so the add-in shows its progress, and if you cancel, everything finished so far stays in the sheet.

No. Install the add-in and your 100 free credits are there, with no signup and no card. When you buy a pack, the license key arrives by email; paste it into the add-in and its credits are added to what you have.