# Is OCR Text Part of the Client File? What to Hand Over

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2025-03-11T22:22:06Z
- Updated: 2026-10-11T19:22:20Z
- Category: Confidentiality
- Canonical: https://www.schemalaw.com/blog/is-ocr-text-part-of-the-client-file

> Whether OCR text is part of the client file depends on your jurisdiction's test. How the two approaches treat converted copies, and what to hand over.

Whether OCR text is part of the client file depends on the test your jurisdiction uses for what a client is owed. Under an entire-file approach a converted copy of a client's document is hard to hold back. Under an end-product approach the answer turns on whether the copy was made for the client or for your own use.

This post explains how to sort converted copies when a client asks for the file. It is not legal advice. The duty comes from your own state's rule, case law and ethics opinions, and those control over anything a national opinion says.

## Is the OCR text of a client's document part of the client file?

The Model Rules do not name OCR text, so the answer comes from the general duty to surrender what the client is entitled to. The ABA's opinion on the subject does not mention converted copies. It does say the duty reaches electronic material, and it lists categories you can sort your exports into.

The starting point is [ABA Model Rule 1.16](https://www.americanbar.org/groups/professional_responsibility/publications/model_rules_of_professional_conduct/rule_1_16_declining_or_terminating_representation/), paragraph (d), which reads in part: "Upon termination of representation, a lawyer shall take steps to the extent reasonably practicable to protect a client's interests, such as giving reasonable notice to the client, allowing time for employment of other counsel, surrendering papers and property to which the client is entitled". The rule does not define those papers.

[ABA Formal Opinion 471](https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-471.pdf), issued July 1, 2015, fills in some of the gap. It says that the "use of the term 'paper' in Rule 1.16(d) includes all communications noted above, whether tangible or electronic." So a file that exists only on a server is not outside the rule for that reason.

OCR output is an awkward fit for the opinion's categories, because it is a second version of something else. A searchable PDF of the client's own lease is the client's document plus a text layer you added. A Word file of that lease is a working copy. A spreadsheet built from the client's bank statements took your staff's time to check. The opinion's two approaches treat those three differently.

## Entire-file and end-product approaches to what a client is owed

Opinion 471 describes two approaches. Under the entire-file approach the lawyer surrenders everything related to the representation unless a specific exception applies. Under the end-product approach the lawyer surrenders finished work and what the client supplied, and may keep material generated for the lawyer's own purpose.

The opinion's words on the first: "A majority of jurisdictions follow what is referred to as the 'entire file' approach." In those jurisdictions, it says, a lawyer must surrender papers and property related to the representation "unless the lawyer establishes that a specific exception applies". The exceptions it lists include material that would breach a duty of non-disclosure to another person and "documents reflecting only internal firm communications and assignments."

On the second, the opinion says other jurisdictions "follow variations of an end-product approach", which separate the end product of the lawyer's services from "other material that may have led to the creation of that 'end-product,' which need not be automatically surrendered."

The ABA committee's own position is a floor. On the facts before it, the opinion says a lawyer must at a minimum surrender, among other things, "any materials provided to the lawyer by" the client, "discovery or evidentiary exhibits", and "third party assessments, evaluations, or records paid for by" the client. It says the client is unlikely to be entitled to material "the lawyer generated for internal use primarily for the lawyer's own purpose".

| Converted copy | Entire-file approach | End-product approach |
| --- | --- | --- |
| Searchable PDF of a document the client supplied | Related to the representation; look for an exception before withholding | The source is client-provided material; the text layer rides along with it |
| OCR text of the other side's production | Related to the representation | Discovery material is on the opinion's minimum list; ask whether the text version is part of it |
| Word or text working copy never sent to anyone | Related to the representation; internal-use exceptions vary | Closer to material generated for the lawyer's own purpose |
| Spreadsheet built from scanned statements and used as an exhibit | Related to the representation | An exhibit is on the opinion's minimum list |

Use the table to frame the question, then check it against your jurisdiction's authority. The opinion itself warns that lawyers "are cautioned to review the law in the jurisdiction in which they practice".

Drafts and margin notes you add after conversion may raise the work-product doctrine, and that call belongs to counsel in your jurisdiction. Client confidentiality still covers every copy while it sits in the firm. If an engagement letter already names the formats the client will receive, follow it when you pack the box. Ask how the duty of loyalty bears on a chronology the next lawyer will need.

## Which converted exports go to the client when the matter ends?

Hand over the exports that stand in for the client's own documents or that were filed, served, sent or relied on. Searchable PDFs of client-supplied scans and spreadsheets used as exhibits are the clearest cases. Unsent Word working copies are the least clear, and the answer for those depends on your jurisdiction's test.

A practical sort for a closing matter:

1. **List every export made from each source scan.** Schema Law, for example, exports in eight formats: Word, Excel, CSV, JSON, searchable PDF, plain text, Markdown and HTML. One scan can have several children. The [export formats on the homepage](https://www.schemalaw.com/#formats) describe what each holds.
2. **Mark each export by use.** Filed or served. Sent to the client. Used as an exhibit. Used inside the firm only.
3. **Hand over the first three groups** with the originals.
4. **Decide the fourth group under your jurisdiction's test,** and write down the reason.
5. **Label what you send.** A converted Word file is not the signed instrument. The post on [converting a signed agreement to Word](https://www.schemalaw.com/blog/convert-signed-agreement-to-word) explains how to mark a working copy so nobody mistakes it for the executed document.

There is also the question of format and cost. Opinion 471 says it "does not address whether and in what circumstances a lawyer must convert an electronic document into paper for a client or who will bear the cost of this conversion". It does not discuss conversion in the other direction, from scan to text, at all. It agrees with the reasoning of D.C. Bar Opinion 357, which it quotes: "Lawyers and clients may enter into reasonable agreements addressing how the client's files will be maintained, how copies will be provided to the client if requested, and who will bear what costs associated with providing the files in a particular form". An engagement letter that says the file will be returned as searchable PDFs settles the point before it becomes a dispute.

When you send the copies, protect them on the way out. The post on [turning a converted PDF into a password protected PDF](https://www.schemalaw.com/blog/convert-pdf-to-password-protected-pdf) covers that step.

## Annotated and corrected OCR copies as the lawyer's working papers

A converted copy that carries your highlights, margin notes or issue tags is two things at once: a copy of the document and a record of your thinking. Opinion 471's list of material a lawyer need not provide includes "personal notes" and "internal legal memoranda and research materials". Whether your jurisdiction agrees is the question to check.

Separate two kinds of change before you decide.

- **Corrections.** You fixed a misread name or a wrong digit in the OCR text so that it matches the page. A corrected copy is a more faithful version of the client's document. It carries no opinion of yours, and withholding it only hands the client a worse copy.
- **Annotations.** You flagged clauses, added comments or built an issues column in the spreadsheet. Those show your analysis.

Where both sit in one file, a clean export solves the problem: keep the annotated version and hand over an unannotated one with the corrections intact. If you corrected the text in a separate editor after export, save the clean corrected version before you start marking it up.

One caution from the opinion. If the representation ends before the matter is finished, it says the former client "may be entitled to the release of some materials the lawyer generated for internal law office use". Its example is a draft prepared for an imminent filing deadline. A half-built chronology spreadsheet that successor counsel needs next week may fall on the same side of that line. Ask what withholding it would cost the client.

## Keeping or deleting converted copies after the file is returned

Returning the file does not end the firm's handling of the copies. Three sets remain: what the firm keeps under its retention schedule, what sits in the conversion tool's account, and what the client now holds. Decide each on purpose and record it.

**The firm's copy.** Rule 1.16(d) ends with this sentence: "The lawyer may retain papers relating to the client to the extent permitted by other law." Opinion 471 notes that the Model Rules "do not directly address the length of time a lawyer must preserve client files after the close of the representation" and that jurisdictions give guidance by court rule or ethics opinion. Follow your jurisdiction's period and your malpractice carrier's instructions.

**The vendor's copy.** A converted document may still sit in the account of the tool that converted it. Delete it there once the firm's own copy is filed, unless a hold applies. The post on [whether OCR software stores your documents](https://www.schemalaw.com/blog/does-ocr-software-store-my-documents) explains where copies can remain, and the post on [OCR vendor subpoenas](https://www.schemalaw.com/blog/ocr-vendor-subpoena) explains why a copy left with a vendor can be demanded from the vendor by someone else.

**The closing note.** Record what was handed over, in which formats, on what date, what the firm kept, and when the tool's copies were deleted. Two lines in the closing memo are enough, and they answer the question if the client comes back in three years asking where the spreadsheet went.

## Try it on a real document

Upload a scanned PDF and export it as Word, a searchable PDF or plain text on the free plan. Your documents are never used to train an AI model. [Start free](https://www.schemalaw.com/api/auth/google)