Real data is never tidy. Your import tool shouldn't need it to be.

In Sloose one sheet feeds the Accounts, Contacts and Deals modules in one import. Sloose works out that the company has to exist first. Each contact and deal is then linked to the company on its own row.

You choose how a company is matched against the companies you already hold, for example by its name or its website. A row whose company is already in your CRM links to that record, and a company on several rows becomes one account: a later row finds the account an earlier row made. Capitals don't count, but "Saltbush Solar Co." and "saltbush solar co" are a full stop apart, so tidy the name with a formula or they stay two.

One file, two ways

In Sloose each module has its own rule for finding a record you already hold. The value it looks for can be a formula, so the email can be put in lower case first and PRIYA.R@… finds the record stored aspriya.r@….

You decide whether a record that is found is updated, and you can add a condition, such as "only when the file has a phone number and it differs". When a row finds several records, Sloose does not pick one for you. You choose to be asked, to skip the row, to create a new record, or to use the first.

Matching existing records, in the docs

The AI proposes a formula where a field needs a split, a join or a change of case, built from functions such as split andproper. You see what the formula makes of every row before you accept it.

There is no built-in address parser. Each part of the address gets its own formula, which you write or ask the AI to write, and the results on your rows show you any row it gets wrong, so you can correct it before the import.

AI mapping

Dates

A formula can read each row's date in the order its country writes it, using the row's Country column. You see the date every row will become before anything is written. A value the formula cannot read, such as4 Mar 2026, is shown as it is, never guessed.

Picklists

Translate values matches each value to an option your CRM's field really has. A free matcher handles differences in case, spacing, punctuation and accents. For the rest, such as "Tassie" for Tasmania, you can ask the AI, and every option it names is checked against the field's real options.

Fixing values with AI mapping

In Sloose a lookup can search for the related record by a field you choose, where your CRM can search that field. The value it searches for can be a formula, so ABN 73 000 100 403 and62-000-100-302 become 73000100403 and 62000100302 first.

You decide what happens when nothing matches, or when several records do: leave the link empty, skip the record, or be asked.

Lookups by search, in the docs

Save September's import as a template. In October, start from it. Sloose checks the new file against the template for free and shows what still fits, what is missing and what is new. If you want help with the gaps, the AI is asked about those parts only.

Templates