Classification

Identifying the document type is one of the first steps to begin document processing. Users can specify the document type by populating the document_type field with one of the document types listed below when uploading or updating a document. Sikoia also automatically classifies the uploaded file/document and populates the sikoia_document_type field too.

📘

Document type comparison check

Many customers compare the document_type and the sikoia_document_type to request an automatic resubmission when the two disagree.

Sikoia leverages advanced ML/AI deep-learning models that combine layout and language features to accurately detect and identify documents.

Document file splitting

During classification, Sikoia also detects whether a single file contains more than one document. When it does, the file is split so that each document can be handled on its own: every individual document identified within the file is assigned its own data_source_id, distinct from the file-level document_id/source_id. This split is what allows each document to be classified, validated, and extracted independently, and it's the same data_source_id referenced throughout the document lifecycle.

📘

Multiple Documents in a Single File

Currently our platform supports input files containing multiple documents of the same document type within a single file.

Each document is detected and processed individually, then mapped into a separate internal structure called a 'data source' to ensure accurate and organised data handling. Data sources are utilized for features such as Income & Employer Verification, Affordability Insights, and Completeness Checks. Learn more about data sources here

Learn more about data sourceshere.

Document Types

The currently supported document types are:

Document typeDescriptionStandard
extraction
Deep
extraction
BankStatementDetailed report of transactions and balances in a bank account.✅✅
BenefitAwardLetterLetter from government agency confirming eligibility for social welfare benefits.✅
ChildBenefitLetterLetter issued from the Child Benefit Office.✅
CompanyRegistrationOfficial record confirming a company's legal existence.✅
ContractAny form of contractual agreement.✅
CreditReportDocument summarising the credit history of an individual or business.✅
EmploymentContractAgreement setting out employment conditions, rights and responsibilities.✅
EmploymentLetterLetter from an employer stating employment details.✅
FactFindDocument summarising essential information about an individual or business.✅
FinancialAccountsDocument showing profit and loss, balance sheet, and cash flow statements for an entity.✅
IdentityGovernment issued card or record to verify an individual's identity.✅
P45HMRC form filled by the employer when an employee stops working.✅
P60End-of-year certificate from HMRC.✅
PayslipDocument issued by an employer showing salary details.✅✅
PensionStatementPension plan statement detailing the amount in a pension pot✅
SA302Detailed income and tax calculation summary, either generated via HMRC or submitted by the individual.✅✅
SavingsStatementBank statement on savings, ISA, and other accounts.✅✅
StakeholdersDisclosureOfficial record about entities holding an interest in a business or project.✅
StudentLoanLetterLetter from a lender specifying details of a student loan.✅
TaxAny other document issued by tax authorities.✅
TaxOverviewLess detailed tax calculation summary from HMRC.✅✅
UtilityBillStatement detailing charges for essential services (e.g. electricity, water, gas).✅
OtherAny other document that doesn’t fit the types above.✅