Validate and dry run
See every record before it's written
Sloose checks the whole import before it writes a single record. Then a dry run goes through every row and every rule, and shows you what each row would create or update in your CRM. Nothing is written until you press Import.
Test an import before importing it
Validate lists everything Sloose can tell you without writing anything, in one place.
- Every file, not only the one on screen. With several files in one import, each problem says which file it is about.
- Grouped by where you fix it. The worst come first, and Go to fix opens the right file at the right part of the import.
- Errors stop the run. Notes do not. A required field nothing fills stops it. A field that will be empty on 40 rows is a note: you decide whether it matters.
- Free. Validate's checks cost no AI credits. Credits are spent only if you ask the AI to explain or fix a problem.

Why does my import fail? Validate finds it first
Many import failures come from something you can see in the file before it runs. These are some of the things Validate checks before a run.
"Invalid enumeration option" and other picklist errors
Values that are not options of the picklist, such as "N.S.W." for "New South Wales". Each one is listed, with a link to translate it.
Required fields
A required field nothing is mapped to stops the run. Rows where a required field is blank are listed by row number.
Columns that are not there
A field mapped from a column the file does not have, and formulas that do not compile.
Links between files
A row that looks up a record no row of the other file provides, and files that look each other up in a circle.
Matching existing records
A module set to "Some may exist" with no field to match on, or a match method your CRM's search does not support.
Percentages that are still fractions
Excel stores 15% as 0.15. When every value in a percentage column is below 1, Validate says so and offers to multiply by 100.
| A | B | C | D | E | F | |
|---|---|---|---|---|---|---|
| 1 | First name | Last name | Company | Country | State / region | Lead source |
| 2 | Marcus | Bell | Saltbush Solar Co. | Australia | N.S.W. | |
| 3 | Grace | Liu | Orchard Lane Catering | AU | new south wales | Adwords |
| 4 | Ravi | Patel | Willowmere Dental Group | Australia | Tassie | |
| 5 | Maria | García | Solera Tapas LLC | US | Calif. | Google Ads |
| 6 | Dana | Whitfield | Lakeside Rowing Club | USA | CA | linkedin.com |
| 7 | Priya | Shah | Maple Row Clinics | Canada | ON | Trade show |
| 8 | Anna-Lena | Schröder | Bergmann Werkzeuge GmbH | Deutschland | Bayern | Messe München |
| 9 | Hannah | Price | Fernhill Bakery Ltd | UK | Gtr Manchester | word of mouth |
The full list is in the docs: What Validate checks.
Explain, and a fix to review
Press Explain on a problem and the AI answers in plain words: why it stops the run, whether it matters, and how to fix it. What it proposes lands as a suggestion on the part it is about. Nothing changes until you use it, and Undo turns them all down.
Ask AI at the top of the list reads every problem together and answers questions likeIs this safe to run? and What should I fix first?
Explain uses AI credits. A request that fails is not charged. Some problems also have a one-click fix that uses no AI at all, such as leaving values that have no match blank, or multiplying a percentage column by 100.

A dry run shows what each row would create or update
A dry run goes through every row and every rule: matching, formulas, value maps and lookups. It searches your CRM for existing records exactly as the real run would. It writes nothing.
- Every row shows, for each module, Would create, Would update, Would skip or Would ask.
- With several files, a row that links to a record another file makes says it would link to the record this run creates.
- A dry run is not counted as a run, and it still works when your month's runs are used up.
- It never spends AI credits.
- Open any row to see what it would send to each module and, for an update, how that differs from the record you already hold.
Run one whenever you want one. When you are ready, Import asks once, saying how many rows the run will send and which modules it creates or updates in.

What a dry run cannot tell you
It cannot prove that your CRM will accept every record. A validation rule, a mandatory field Sloose was not told about, or a workflow in the CRM only shows up on a real write. Sloose also has no undo: records a run wrote stay written. That is why the checks come first.
When a row matches more than one record, you choose
Each module has its own answer for when several existing records match a row: use the first, skip the row, create a new record, or ask me.
With ask me, the run pauses on that row and shows each matching record with its fields. Pick the right one, skip the row, or create a new record. You can apply a skip or a create to every row like it for the rest of the run.
A dry run never stops to ask. It marks those rows Would ask, so you know how many decisions the real run holds before you start it.

From a file to your CRM, checked at each step
Each column goes to a field, with a formula where the value needs changing.
- Validate
Every file and every rule is checked. Fix what stops the run, and read the notes.
- Dry run
See what each row would create, update, skip or ask about. Nothing is written.
Confirm once. Rows are written one at a time, and you can pause or cancel.
Failed rows are grouped by why. Retry them without writing again what worked.
Questions people ask
Can I test an import before importing?
Yes. Validate checks the whole import for free, and a dry run goes through every row and writes nothing. It shows what each row would create or update in each module.
Does a dry run cost anything?
No. A dry run is not counted as a run on your plan and never spends AI credits. It still works when the month's runs are used up.
Can I undo an import?
No. Sloose has no undo, and records a run wrote stay written. What it gives you instead comes before the write: Validate, a dry run, and one confirmation before Import. If some rows fail, a retry runs only those rows.
Why does my import skip records?
Usually because a required field is empty, or a value does not fit the field, such as an email address the CRM cannot read. Validate lists those before the run. After a run, failed rows are grouped by the reason the CRM gave. See What failed in the docs.
Does Validate send my data to the AI?
No. Validate's checks run without the AI. Something is sent only when you press Explain or Ask AI, and that uses credits. Your data and the model says what is sent.
Check your next import before it writes
Start on the free plan and run as many dry runs as you like.