Import your data

Most businesses arrive with years of records in spreadsheets. The onboarding importer reads those files and compiles a reviewable starting catalog — units, suppliers, customers, items, opening stock, and recipes — for you to check before anything lands. It’s an import compiler, not an autonomous agent: it works only from what’s in your files, and no business records are created in your workspace until you approve them.

4 steps · ~15 minFiles: CSV, Excel, PDF, imagesAccess: Inventory, Sales & Purchasing admin

Where it lives, and who can run it

Import runs from the AI Onboarding flow when a workspace is new (the Import step, before Review and Connect). Because it can create records across your whole business, it’s limited to the workspace owner or a member who is an admin in Inventory, Sales, and Purchasing.

What you can import

A single import builds the whole foundation. Each record type maps to a real object in Ashicore, and the importer keeps them internally consistent — an item points at a unit, opening stock points at an item, a recipe points at its components.

Record	Creates	Key fields
Units	Unit definitions	Display name, a numeric size, and a measurable base unit (e.g. cu ft, lb, kg). Packaging words like bag, tote, or pallet live in the name, never the base unit.
Suppliers	Purchasing suppliers	Name, code, contact name, email, phone.
Customers	Sales customers	Name, email, phone. Master data only — no orders are invented from transactional rows.
Items	Materials & products	Type (material or product), name, SKU, stock unit, optional purchase unit & factor, default purchase/selling prices, sellable, lot tracking.
Opening stock	Starting balances	Item, quantity, unit cost, lot number, and received date — captured as of a single opening date.
BOMs	Recipes	Product, output quantity & unit, and components (item, quantity, per-unit or per-batch).

The import flow

1. Upload your files. Drop in whatever you already keep your operation in — inventory spreadsheets, price lists, supplier and customer lists, recipe sheets. Upload up to 12 files by drag-and-drop or the file picker (50 MB each, 100 MB total). Across a workbook the importer reads more than the active sheet — including hidden sheets and cell comments — so a quantity tucked in a comment can still be picked up. When the files are in, choose Extract my data.

   AI Onboarding · Bring in your data Extract my data

   Drag files here, or browse your computer

   current-inventory-2026.xlsxwholesale-price-list.xlsxsuppliers.csv

2. Ashicore reads and structures your files. The importer extracts your files into the catalog above, maps shorthand to real units, normalizes obvious spelling variants, and prefers full product names from your current-inventory sheets over abbreviations buried elsewhere. Where the source is genuinely ambiguous, it flags it for review rather than guessing.

   Reading your filesRecognizing products & quantitiesMatching suppliers & customersStructuring into your ERP

   It reads signals, not just columns

   A unit price sitting beside a stock quantity on a current-inventory row is read as that lot’s cost. A comment like ”≈ 40 kg available, aged” becomes a reviewable stock row. Archive sheets, historical orders, and recipe quantities are not mistaken for stock on hand.

3. Review every record before it lands. This is where you spend your time. Each row carries a confidence score, a link back to its source (file → sheet → row), and a match against anything already in your workspace — create, update, or skip. Suppliers match by code or name, customers by name, and items by SKU or name, so a second import doesn’t duplicate your catalog. Deselect rows you don’t want, and filter to Needs review to focus on what’s uncertain — anything below 70% confidence, or that you’ve deselected, lands there.

   Review · Here's what we found

   Action	Item	Source	Confidence
   create	House Espresso product	current-inventory · A14	0.94
   create	Huila Green RVS-GRN-HUILA	current-inventory · A22	0.88
   update	Decaf Espresso was “Decaff Espresso”	price-list · 31	0.71

4. Approve, and commit. When you’re ready, Ashicore validates your selections and returns a preview split into blocking issues (must fix) and warnings worth a look. Opening stock, for example, needs an active item, a positive quantity, and a unit cost — a row missing any of those blocks until fixed. Clear the blockers and choose Approve all & continue — your units, partners, items, opening balances, and recipes are written in one pass. A committed import can’t be edited, so review carefully first.

   On a paid plan, checkout comes first

   If your plan needs payment, approval routes through checkout and your records commit once payment and entitlements clear. On a free trial they commit on approval.

How the importer thinks

1

A compiler, not an agent

It transforms the files you give it into reviewable records. It works only from what’s in your files — it doesn’t browse the web, write to your workspace on its own, or fill in values that aren’t there.

2

Provenance & confidence score on every record

Each row knows which file, sheet, and cell it came from, and carries a confidence score. Low-confidence rows are exactly where your review attention belongs.

3

Matched against what you already have

Re-importing or topping up? Rows are matched to existing records and proposed as create, update, or skip — so a second import doesn’t duplicate your catalog.

4

Ambiguity is flagged, not guessed

Where the source is unclear, the importer flags it for review and tags it as a warning or a blocker — it doesn’t browse the web or fill in values that aren’t there. Blockers must be resolved before you commit; warnings let you proceed with eyes open.

Two things that trip people up

Units need a real base measure. “Bag”, “tote”, and “pallet” are packaging, not units — they belong in a unit’s name with an explicit size and a measurable base unit. Opening stock is “now”, not history. Counts come from your current-inventory sheets and post as starting balances on one date; archived and historical rows are deliberately ignored so you don’t reload last year’s stock.

What’s next

• Connect your integrations — after the initial import you can keep pulling data in: contacts from Xero or QuickBooks, paid orders from Shopify.
• The inventory ledger — opening stock posts as starting balances; see how every balance is projected from events.
• Add a product or material — fill any gaps the import left, by hand.