From Contract Documents to Structured Data: What Digitisation Changes—and What It Does Not

A contract may begin as a signed paper document, a PDF, an email chain or a folder containing several versions and supporting records. Digitisation can make that material easier to search, compare and review by turning selected information into structured data.

But structure is not the same as legal certainty. Extracting a party name, payment amount or assignment clause does not prove that the document is authentic, complete, enforceable or transferable.

Key point: Digitisation changes how information can be organised and used. It does not automatically change the underlying legal rights, obligations or evidential status.

Digitisation can describe several different steps

The term digitisation is often used for several separate activities:

1. Creating a digital copy — scanning a paper document or uploading an existing electronic file.
2. Making content searchable — using text recognition to identify words in an image or PDF.
3. Classifying documents — identifying possible document types, versions and relationships.
4. Extracting information — recording names, dates, amounts, clauses and other fields.
5. Normalising data — presenting dates, currencies, identifiers and labels in a consistent format.
6. Linking evidence — connecting a recorded fact to its source document and location.
7. Reviewing and confirming — allowing an authorised user to correct or accept the extracted information.
8. Using structured information — searching, comparing, reporting or moving the reviewed data into another controlled workflow.

Each step adds usefulness, but each also has limits. A scanned file may be readable without being complete. Searchable text may contain recognition errors. A structured field may reflect an AI suggestion rather than a confirmed fact.

What structured contract data can contain

Depending on the document and the purpose of the review, a structured record may include:

• document title and type;
• parties and their stated roles;
• signature and effective dates;
• term, expiry and renewal dates;
• governing law and dispute provisions;
• payment amounts, currencies and schedules;
• interest, fee or penalty provisions;
• deliverables and performance obligations;
• notice and termination provisions;
• assignment, novation and change-of-control clauses;
• security, guarantee or collateral references;
• confidentiality and data-use restrictions;
• amendment and version relationships;
• references to schedules, annexes and supporting evidence; and
• source-document and review-status information.

The useful output is not merely a list of extracted words. It is a structured description showing where the information came from, which document version it belongs to and whether it has been reviewed.

What digitisation changes

1. Information becomes easier to find

A reviewer can search for relevant parties, dates, amounts or clauses without opening every file manually. This can reduce the time needed to locate information, especially in a large or inconsistent document set.

Search results remain leads for review. They should not be treated as complete simply because a system did not find another clause or document.

2. Documents can be connected

A contract rarely stands alone. Amendments, notices, invoices, payment records, correspondence and court documents may change or explain the position.

Structured relationships can show that:

• one agreement replaced an earlier version;
• an amendment changed a payment date;
• a notice relates to a particular termination clause;
• an invoice arose under a specified contract; or
• a payment should be reconciled against a stated obligation.

These links make the package easier to navigate. They do not establish that the legal interpretation of the relationship is correct.

3. Information can be compared consistently

Normalised fields can help reviewers compare several contracts or claims. For example, dates can use a consistent format and amounts can be separated into principal, interest, fees and payments.

Consistency must not erase context. Currency, as-of date, calculation method, source and qualifications should remain visible. Two fields with the same label may still have different legal meanings.

4. Gaps and inconsistencies become more visible

Structured review can help identify possible problems such as:

• missing signatures or schedules;
• inconsistent party names;
• conflicting dates or amounts;
• absent amendments;
• unexplained payment differences;
• expired notice periods;
• unclear assignment restrictions; or
• information extracted with low confidence.

A gap indicator is not a legal conclusion. It is a prompt for a person to investigate the source material.

5. Provenance can be preserved

A useful structured fact should remain linked to its source: the document, version, page or text from which it was recorded.

Provenance allows a reviewer to ask:

• Was this information supplied by the user or extracted automatically?
• Which document supports it?
• Has a person reviewed it?
• Was it later corrected?
• Which version was used?

This review trail improves transparency. It does not prove that the source document itself is authentic or legally effective.

6. Controlled workflows become possible

Reviewed information can support tasks such as creating an Asset Passport, preparing a document index, requesting missing records or sharing selected information with an authorised reviewer.

Automation should remain conditional. A structured field should not trigger registration, transfer, payment, publication or another material action without the required authority, validation and human control.

What digitisation does not change

It does not rewrite the contract

Extracting and organising terms does not amend the agreement or alter the parties' rights and obligations. If structured data conflicts with the source document, the source and applicable legal rules require review.

It does not prove authenticity

A clear scan can still be incomplete, altered, misattributed or unauthorised. File metadata, a hash or a timestamp may help establish the state or integrity of a particular file at a particular time, but they do not prove that every statement inside the file is true or that each signature is genuine.

It does not create a valid signature

Typing a name into a field, uploading a signed-paper scan or clicking an internal confirmation button is not automatically equivalent to executing a document with an appropriate electronic signature.

Electronic-signature requirements depend on the document, transaction and applicable law. In the EU, eIDAS distinguishes electronic signatures, electronic seals and timestamps and gives particular legal effects to qualified trust services. The fact that a record is electronic does not mean every formal requirement has been satisfied.

It does not establish ownership or authority

A document may name a party without proving that the uploader owns the relevant right, still holds it or has authority to disclose or transfer it. Ownership and authority may require a complete chain of title, corporate approvals, current registry information and other evidence.

It does not make a right transferable

A clearly extracted assignment clause does not resolve every question about consent, notice, debtor defences, priority, competing interests or applicable law. See Transferability Is Not Automatic.

It does not determine value

A contract amount, outstanding balance, royalty rate or revenue figure is not automatically the value of the associated asset or right. See Claim Amount Is Not the Same as Asset Value.

It does not complete official registration

Recording information in a private technology platform does not by itself create an entry in a government, court, intellectual-property, land, company, security or securities register. It also does not turn an ordinary contract into a security or regulated financial instrument.

It does not guarantee performance or payment

Structured data can describe an obligation. It cannot guarantee that a party will perform, that a debtor will pay, that a claim can be enforced or that a buyer will complete a transaction.

A digital copy is not always a digital original

It is important to distinguish among:

• a scan of a paper document;
• a copied PDF;
• an electronically created document;
• an electronically signed document;
• a document bearing an electronic seal or timestamp;
• an authoritative record maintained under a defined legal or operational system; and
• an electronic transferable record recognised under applicable law.

These records may have different functions and evidential effects.

The UNCITRAL Model Law on Electronic Transferable Records illustrates the distinction. Its framework concerns particular electronic records that are functionally equivalent to transferable paper documents or instruments, using reliable methods to identify the authoritative record, maintain control and preserve integrity. Uploading an ordinary contract PDF does not automatically create such a record, and the Model Law applies only where adopted and within its defined scope.

AI-assisted extraction still requires review

AI can help identify document types, clauses, parties, amounts and possible gaps. It can also make mistakes because of:

• poor image quality;
• handwriting or unusual layouts;
• missing pages;
• ambiguous language;
• inconsistent defined terms;
• tables, annexes or scanned signatures;
• conflicting document versions;
• jurisdiction-specific terminology; or
• information that is implied rather than stated.

AI-generated fields should therefore be treated as reviewable findings, not certified conclusions. Important facts should remain linked to their sources and be confirmed by an authorised person before they are relied on or shared.

A practical structured-data checklist

Before relying on a structured contract record, check:

1. Document identity — Is the file type, title and version correctly identified?
2. Completeness — Are all pages, schedules, amendments and signature sections present?
3. Source — Is the origin of the file recorded?
4. Parties — Are legal names, roles and identifiers taken from the correct source?
5. Dates and currency — Are formats, time zones, currencies and as-of dates explicit?
6. Evidence links — Can each important field be traced to the supporting document location?
7. Review status — Is it clear what was extracted, user-supplied, corrected or independently reviewed?
8. Conflicts — Are inconsistent versions or values visible rather than silently overwritten?
9. Access — Is personal and confidential information shared only with authorised recipients?
10. Purpose — Is the structured information suitable for the specific task, without being reused beyond its context?

What DaDepo does—and does not do

DaDepo can help organise uploaded documents, extract available information, preserve source references and present reviewed information through an Asset Passport. Users review the findings, correct errors and control how information may be shared.

Digitising or structuring information through DaDepo does not mean that DaDepo has:

• created or amended the underlying contract;
• authenticated every document or signature;
• confirmed legal validity, ownership or enforceability;
• provided a qualified electronic signature, seal, timestamp or archiving service unless expressly identified as such;
• completed an official registration;
• determined transferability, priority or regulatory classification;
• audited financial or commercial information;
• valued the asset or recommended a transaction; or
• guaranteed performance, payment, sale, settlement or liquidity.

Important: DaDepo provides technology and information tools. It does not provide legal, financial, investment, tax, accounting, regulatory, technical-certification or valuation advice. Users and reviewers should examine the source documents and obtain appropriate professional advice.

Structure supports review—not automatic certainty

The value of digitisation is practical. It can turn a difficult folder into a navigable evidence package, make facts easier to compare and keep findings connected to their sources.

Its limit is equally important: better-organised information is still information that must be interpreted, verified and used within the correct legal and commercial context.

Further reading

• UNCITRAL Model Law on Electronic Transferable Records
• European Commission: eSignature FAQ
• European Commission: Questions and answers on trust services
• Law Commission: Electronic execution of documents