> ## Documentation Index
> Fetch the complete documentation index at: https://labs.laer.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Best practices for Libraries

Use these best practices to get the most value from libraries in Epiq AI and ensure consistent, high‑quality results across workflows.

### Plan libraries before you run workflows

* Build libraries **before** running workflows whenever you already know key people, organizations, or prompt content.
* Treat libraries as shared inputs, not workflow‑specific fixes.
* Start with high‑confidence entries to avoid introducing noise early.

### Use the right library type

* **Person libraries.** Use for known individuals who are likely to appear repeatedly in the data.
* **Organization libraries.** Use for known companies, firms, domains, or other organizations.
* **Prompt libraries.** Use for approved, reusable prompt text that should remain consistent across workflows.
* **Custom libraries.** Use for reference data the other three types cannot represent — matter codes, deal names, product identifiers, custodian groupings.

Do not mix content types across libraries. Each library type serves a distinct purpose.

Reach for a custom library only when the built-in types genuinely do not fit. A person recorded in a custom library is invisible to the workflow components that read person libraries, so representing people or organizations in a custom shape costs you the automatic enrichment and tag sync that make libraries worth building.

### Scope libraries at the level the data belongs to

* **Client** for anything true of the client rather than of one matter: its in-house counsel, its subsidiaries and domains, its regular outside firms, its approved prompt language. One library then serves every matter for that client, including matters that do not exist yet.
* **Project** for reference data that is genuinely specific to one matter — an opposing party's counsel, a custodian list for this collection.

Set client organizations deliberately rather than relying on the default. A library created inside a project inherits that project's client automatically, which is usually what you want — but it also means the library cannot later be applied to another client's matter. Leave the client organizations field empty if the library is meant to be shared more widely.

### Design custom library columns before you import

* Choose the **key field** first. It is what Overlay imports match on and what the library is searched by, so a poorly chosen key produces duplicates every time you re-import.
* Prefer a stable identifier over a display name for the key field. Names get re-spelled; codes do not.
* Set each column's **type** rather than leaving everything as text. Dates and numbers filter and sort correctly only when typed correctly.
* Use **Import from file** to seed columns from a spreadsheet's headers, then correct the types. It is faster than defining a dozen columns by hand, but it replaces every existing column, so use it when starting out rather than to add one field.
* Download the library **template** before preparing an import file. Its headers are exactly the columns you defined.

### Keep entries clean and normalized

* Use a single, canonical version of each person or organization name.
* Avoid duplicate entries that represent the same entity.
* Remove outdated or incorrect entries during periodic reviews.
* Prioritize clarity over completeness. Large, unreviewed libraries can reduce accuracy.

### Update libraries during review cycles

* Add new people or organizations discovered during workflow review.
* Feed corrected workflow output back into libraries where appropriate.
* Update prompt libraries when prompt guidance changes or improves.
* Treat library maintenance as part of the review process, not a one‑time setup step.

### Reuse libraries across workflows and matters

* Share libraries across workflows that involve the same participants or objectives.
* Reuse prompt libraries to enforce consistent instructions and outputs.
* Avoid creating duplicate libraries for the same purpose unless required by data separation needs.
* Prefer one client library applied broadly over the same list rebuilt per matter. Client scope exists precisely so that curation done once carries forward.
* Feed verified entities back with **Save to Libraries** at the end of a matter. The work a review team did establishing who an attorney is should not have to be redone on the next matter for the same client.

### Version and govern prompt libraries carefully

* Store only approved or validated prompt content in prompt libraries.
* Make prompt changes deliberately and document the reason for changes outside the library if required.
* Avoid experimental or workflow‑specific prompts in shared prompt libraries.

### Review libraries regularly

* Schedule periodic reviews to:
  * Remove duplicates
  * Correct errors
  * Retire unused entries
* Validate that libraries still reflect current matters and workflows.

### Common mistakes to avoid

* Uploading incomplete or unverified lists
* Treating libraries as disposable, one‑off inputs
* Allowing uncontrolled growth without review
* Mixing experimental data into shared libraries
* Creating a custom library and closing the dialog before adding columns, leaving it unusable
* Re-importing corrections with **Append**, which skips every row whose key already exists and silently applies none of them. Use **Overlay** to update