Last date modified: 2026-Oct-06
Creating an aiR Review project
The instructions you give aiR Review are called prompt criteria. They often mimic a traditional review protocol or case brief in that they describe the matter, entities involved, and what is relevant to the legal issues at hand. To help improve results, we recommend analyzing a small set of documents, tweaking the prompt criteria as needed, then finally analyzing a larger set of documents. Starting with a small set allows you to evaluate how aiR Review's coding compares to a human reviewer's coding and adjust the prompts accordingly.
Refer to Best practices for more information on prompt criteria tips, sample workflow guidance, and iteration workflow.
See these related resources for more information:
- aiR Review Validation Concepts and Definitions (White paper)
- AI Advantage: Aiming for Prompt Perfection? Level up with Relativity aiR Review
- A Focus on Security and Privacy in Relativity’s Approach to Generative AI
- Workflows for Applying aiR Review
- aiR Review example project
- aiR Review Prompt Writing Best Practices
- Evaluating aiR Review Prompt Criteria Performance
- Selecting a Prompt Criteria Iteration Sample for aiR Review
Project workflow
The aiR Review project workflow has three basic parts:
- Setting up the project—defining the project name, description, data source (saved search), and analysis type.
- Developing the prompt criteria—specifying the criteria that the AI should apply when reviewing documents according to the chosen analysis type.
- Running the first analysis—running AI analysis on documents according to prompt criteria.
At any point in this process, you can save your progress and come back later.
Job capacity, size limitations, and speed
Based on the limits of the underlying large language model (LLM), aiR Review has size limits for the documents and prompts you submit, as well as instance-level volume limits for the overall jobs.
Size limits
The documents and prompt criteria have the following size limits:
| Category | Applies to | Maximum | Description |
|---|---|---|---|
| Prompt kickstarter | Documents |
10 documents |
Maximum of 10 documents with a maximum combined total of 300,000 characters. See Using prompt kickstarter (Draft with AI) for more information on using the prompt kickstarter feature for some analysis types. |
| Combined document content | 300,000 combined total | ||
| Prompt criteria | Combined content across all tabs | 15,000 characters | The combined content of all prompt criteria tabs cannot exceed 15,000 characters to avoid errors. See Best Practices for tips on writing prompt criteria and removing excess information. |
| Document extracted text | Text per document | 0.05 - 300KB |
Documents must fall within this range to be eligible for aiR Review analysis. Although aiR Review can handle extracted text up to 300KB, use caution above 150KB as processing errors may prevent results. This range also serves as the limits of an aiR Review "aiR Unit" that is available for aiR Unit Analysis. |
| Custom Analyses limits | Prompt criteria |
1,000 characters |
Custom insight/prompt cannot exceed 1,000 characters. |
| Insight title |
30 characters |
Custom insight title cannot exceed 30 characters. |
|
| Vision analysis native image |
20MB |
|
- The size of a job encompasses both the document size and the prompt criteria size. If the combined total exceeds the per document limit, the job will be too large to process.
- If a document receives an error, your organization will not be charged for it. For more information, see Handling document errors.
Volume limits per instance
The per-instance volume limits for aiR Review jobs are as follows:
| Limit per instance | Applies to | Maximum |
|---|---|---|
| Job size |
Relevance, Key Document, Confidential Business Information (CBI), and Custom analyses |
300,000 documents |
|
|
250,000 documents | |
| Total documents | All queued and running jobs in an instance | 600,000 documents |
| Concurrent large jobs | Jobs with more than 200 documents |
|
| Concurrent small jobs | Jobs with 200 or fewer documents | No limit |
Speed
After a job is submitted, aiR Review can analyze up to 3 million documents per day.
For individual jobs, aiR Review averages between 10,000 and 13,000 documents per hour.
Job speeds vary widely depending on the region, the number and size of documents, the overall load on the LLM, and other similar factors.