The AI Split and Index action examines a multi-page stack of any document type, works out where one document ends and the next begins, and splits the stack into a new Record per document. It can also index each document as it goes, capturing its header fields and its line items, so the split Records come out indexed without a separate extraction per Record. It is the general-purpose sibling of AI Invoice Split: instead of being locked to invoices, it takes its sense of what a document looks like from the Record Type you point it at.


Use it for a scanned batch that contains many separate documents of the same kind — a stack of invoices, statements, applications, remittances, tax forms, or delivery notes fed through the scanner in one pass.


AZURE ONLY: This is an AI vision action and runs only on Azure-hosted / AI-licensed sites.

How the split is found

The action looks at the pages and watches the identity fields you name. When a page asserts different values for those fields, that page starts a new document; pages that agree, or that say nothing about them, continue the current one. This is the one thing the AI cannot work out on its own, and it is the only setting the action requires.


Identity fields are field names, not the values to look for. This is the opposite of the Auto-Split family, where you list the values printed on a separator sheet. Nothing has to be printed on the page for the split to work except the identity values themselves — no separator sheets, no barcodes, no page counts.


Use more than one identity field when a single one is not unique on its own. For a stack of statements, Account Number alone repeats across months, so Account Number plus Statement Date is the honest identity.


A Record Type is not required for the split. When one resolves, its field descriptions sharpen the prompt; when it does not, the identity field names alone are enough. Indexing works best with one, because the fields it captures come from it: without a Record Type, only the identity fields and any fields named in Header Field Instructions are captured, and line items are skipped.


Options

The action's settings are grouped into five tabs.

Document(s) to Split

  • New Document Cat/Sub/Name — The Category, SubCategory and Name applied to each split Record's document. Leave blank to use the defaults.

The Record's active document is the one that gets split, or the Record's first document if none is active.

Splitting

  • Identity Fields — Record Type field names, one per line, whose values identify a single document. This is the only required setting.
  • Commit Split — On by default: the action creates a new Record per detected document and routes them straight away. Turn it off to stage the detected splits instead, for a later Auto-Split Commit action to commit.

Indexing

The AI has to read each document's header to know where the boundaries are, so the values are already in front of it. This tab keeps them instead of throwing them away, and can read each document's line items as well.

  • Enable Indexing — Off by default, and with it off the action splits only. Turn it on and the same AI call also captures the current Record Type's header fields for each document, writing them to each split Record as Index_<name>. The rest of the indexing settings are on the two tabs that appear below the switch, Header and Line Items.

Header

  • Write to Record Fields — Off by default. Turn it on and each captured value is also written onto the Record Type field of the same name, so no Generic Indexing step is needed. With Capture Line Items on, this one switch covers the line items too. See What each split Record starts with below for the fields it leaves alone.
  • Header Field Instructions — Optional. Tells the AI how to read the header fields it captures. Use it to constrain a field the Record Type already defines (allowed values, format, where it sits on the page), or to add a field beyond the Record Type by naming and describing it. Line item rules go in Line Item Instructions on the Line Items tab. Note this is a different box from Extra Instructions on the Advanced tab, which is about finding boundaries rather than capturing values.
  • Exclude Fields — Optional. Record Type header field names, one per line, to leave out of the capture. See below.

Line Items

  • Capture Line Items — Off by default. Turn it on and the action also reads each document's line items, using the current Record Type's line item fields, and writes them to its split Record as IndexLine_<name>, one row per line. The Record Type must have line item fields; if it has none, the workflow log notes that the line items were skipped and the split goes ahead as usual.
  • Line Item Instructions — Optional, and shown only when Capture Line Items is on. Tells the AI how to read the line items: which rows count, what a column means, or a line field to add beyond the Record Type by naming and describing it.
  • Exclude Fields — Optional, and shown only when Capture Line Items is on. Record Type line item field names, one per line, to leave out of the line items. This is a separate list from the one on the Header tab.


The line items are read in a second AI call that reads the same pages again. The AI keeps those pages for a short time after the split, so the second read costs far less than the first; the line items themselves are counted like any other AI output. Both calls appear as one row on the AI Workflow report.


Line items are read the way AI Generic Extraction reads them. Each row of the document's item table becomes one line, in the order printed, and a table that carries on to the next page stays one list. Subtotal, total and footer rows are left out. Values are read where they are printed on the page and never corrected: if the document itself is wrong, the values come back as printed. Amounts come back without currency symbols or thousands separators.


If a document's line items cannot be read, its split still goes ahead and its Record gets an Index_AIWarning value you can search or filter on. Index_AIWarning also carries the AI's notes about a document's line items, such as a field it could not find or a value it could only partly read.


Indexing is a token-saving option rather than a convenience. Splitting an eighty-document stack and then extracting each Record separately is eighty-one AI calls; indexing on the way through is one split call plus a line call for each batch of pages. In our tests on stacks of invoices, capturing header fields and line items this way used roughly half the workflow tokens of splitting first and then running AI Generic Extraction on each Record. Writing the values onto the Record fields, whether here or with Generic Indexing afterwards, makes no AI call.


AI Level

  • Quality — Balanced suits most business documents; Maximum costs more and is best for poor scans, dense pages, or subtle boundaries. Vision detection needs a capable model, so there is no Economy option for this action.

Advanced

  • Extra Instructions — Boundary rules specific to your documents, for example "A continuation page repeats the account number in the top-right" or "Terms and Conditions pages belong with the document before them". When in doubt the AI keeps pages together rather than over-splitting.
  • Date Format — How ambiguous dates such as 05/06/2025 are read when Indexing captures header fields and line items. Use System Default follows the Default Date Format in System Options; Month/Day/Year (US) or Day/Month/Year (European) forces the order for documents this action processes, and the dates then come back as yyyy-MM-dd. Not used when Indexing is off, since nothing is captured. This is the same setting AI Generic Extraction has.

Exclude Fields

There are two Exclude Fields lists, one on the Header tab and one on the Line Items tab. Each takes fields back out of what the AI is asked for: enter Record Type field names, one per line, and those fields are left out entirely — the AI is never asked for them, so they cost no tokens and cannot come back wrong.


Names must match the Record Type field name exactly, although capitalization does not matter, and wildcards are not supported. The Header list matches header fields only, and the Line Items list matches line item fields only.


Identity fields are never excluded from the Header list, whatever you put in it. The split finds its boundaries by watching the identity fields change from page to page, so leaving one out would break the detection rather than simply skip a value. An identity field that appears in the list is still captured.


What each split Record starts with

Each new Record inherits every data value from the original Record, which is how it keeps the right Record Type and any filter values that Record Type depends on. Captured values are always saved under the Index_ and IndexLine_ prefixes, a separate set of names that cannot collide with a field the Record Type is relying on.


With Write to Record Fields on, each captured value is also written onto the Record Type field of the same name, and each line item value onto the Record Type line item field of the same name, at its line number. Two kinds of field are never written: the RecordType field, and any field with a Filter Value set, since those decide which Record Type the new Record belongs to. Fields listed in either Exclude Fields list are skipped too, except identity fields. Values are copied exactly as captured, with no further currency or date cleaning, and each split Record is stamped with who indexed it and when, as Generic Indexing does.


A header field the AI did not find on a document keeps the value that document's Record inherited from the original Record. Line items work differently: for each line item field the split writes, the lines it writes replace the lines of that field the Record inherited, so a document with three lines does not keep the original's lines four onward. Fields added through Header Field Instructions or Line Item Instructions that are not on the Record Type stay under Index_ or IndexLine_ only.


Leave Write to Record Fields off and follow the action with Generic Indexing when you need a value moved to a field with a different name, or cleaned as currency or as a date. The Index_ and IndexLine_ values are there either way, so the two can also be combined.


TIPS

  1. Choose identity fields that are printed on the page. A calculated field, or an internal status the document never shows, cannot identify anything the AI can see.
  2. Too many Records usually means an identity field is too narrow — something that changes within a single document, like a page number. Too few means it is too broad.
  3. Turn on Enable Indexing whenever the split Records need values, and Capture Line Items when the documents have line items. With Write to Record Fields on as well, the split Records come out indexed, header and lines, with no further action.
  4. Set the field Header text on the Record Type, for header fields and line item fields alike. It is sent as each field's description, so a clear Header improves the values you get back without any instructions at all.
  5. Use the Exclude Fields lists to keep the capture to what the documents actually show. Fields the documents never contain cost tokens on every document and give the AI something to guess at.
  6. If the split is close but not right, describe the exception in Extra Instructions before reaching for Maximum quality — a single sentence about continuation pages usually fixes more than a bigger model does.


NOTE ON (AI): Reading the document with AI consumes tokens and may incur extra fees. Workflow tokens are billed separately from End-User or Admin tokens — check with your sales rep for more information.