Focused View FAQs : Opus 2 Platform

Minimum Software Version	8.17.X Winter25
Solution(s)	Hearings X Cases International ✓ Cases US ✓ Institutions ⨉ Counsel ✓

This article answers the most common questions about Focused View — the new, AI-first interface available in Opus 2 Cases. It covers how to enable the feature, how documents are processed, how the AI assists work, and how permissions and token usage are managed.

About Focused View

Following Opus 2's acquisition of Uncover in October 2025, Opus 2 Cases and some of Uncover's AI technology have been integrated into the Opus 2 platform. There are now two views available:

Detailed View — the existing Opus 2 Cases interface. All functionality you depend on remains available and unchanged.
Focused View — a new, streamlined, AI-first, Uncover-inspired interface where you can access the first capabilities arising from this integration.

You can seamlessly switch between the two views. Documents surfaced by Matter Assist or Documents Assist in Focused View can be added to collections or worksheets in Detailed View.

FAQs

Getting started

1. Can I switch on Focused View myself?

No. Focused View must be enabled by Opus 2 Platform Support.

2. Can Focused View be enabled or disabled for specific matters?

Yes. Focused View can be enabled or disabled on a per-matter or per-template basis. If you need certain matters to have this functionality turned off, let us know which ones during onboarding. The implementation team will apply the appropriate feature flags and ensure those matters are configured correctly.

Document processing and vectorisation

3. Do I need to upload my case documents again before using the new features?

No. After your firm has been introduced to Focused View and trained, OCR'd documents in your Opus 2 Cases will be ready for use in Focused View automatically.

Note: Native files (non-PDFs or transcripts) uploaded to the platform may not be compatible with Focused View, as the platform is unable to extract and display their text via OCR. Files may also be considered native if they do not display or convert after ingestion. For more information, see File upload and processing.

4. What does vectorisation mean, and why does Opus 2 use it?

Vectorising text means converting words into numbers — called vectors — so that a computer can process them. These numbers represent the meaning of the text, not just the exact words, which means the AI Assists can find relevant information even when the same idea is described using different words. Combining keyword search with vector search makes results more accurate and useful.

5. Do users need to do anything for their documents to be vectorised?

No action is required. OCR'd documents in your Cases matters are automatically available for use in Focused View. Documents can be OCR'd on upload, or afterwards by selecting the document, going to Actions, and selecting Run OCR.

After your firm has been introduced to Focused View, new documents are vectorised automatically in the background on a rolling basis. The status column in the Documents Library and the Select documents panel in the Matter Assist dialogue both show the current vectorisation status.

6. Is there a document size limit?

Focused View does not have a hard document size limit. How documents are analysed depends on the amount of content being reviewed.

For queries involving up to approximately 180,000 tokens (typically around 300–350 pages of legal text), Focused View can read and analyse the full text directly.

Where the documents being reviewed exceed this context size, Focused View uses retrieval techniques to identify and analyse the content most relevant to your query. This means you can still work with larger documents or sets of documents; however, rather than reading all of the content at once, Focused View retrieves the most relevant sections to generate its response.

Context limits will continue to be evaluated as AI technology evolves. Any significant changes will be communicated through release notes and product updates.

7. Is there an upper document limit when working with tools in Focused View?

Focused View can analyse documents across your entire matter corpus and does not have a hard document limit.

For optimal performance, we currently recommend matters of up to 20,000 documents. Focused View can still be used with matters containing more than 20,000 documents, but you may experience reduced performance as the size and complexity of the matter increases.

You can also choose to focus your query on a specific selection of documents. When selecting documents directly, you can select up to 2,000 documents at a time. This selection limit does not restrict Focused View's ability to analyse the wider matter corpus.

AI assists

8. Why are Matter Assist and Document Assist being merged?

Customers reported confusion about when to use each assist.

Originally, Matter Assist analysed information across the entire matter, while Document Assist focused on selected documents. Matter Assist was later updated to analyse selected documents as well, which made the distinction between the two less clear. Users had to decide which assist was best suited to each query, creating unnecessary cognitive load.

The unified experience removes that decision-making burden and allows the system to determine the best way to process each request.

9. Will users still need to choose the right assist for their query?

No. The goal of the consolidation is to eliminate the need for users to determine which assist is most appropriate. Backend routing logic directs the query to the most suitable processing method, so users can simply ask their question and receive the best available answer.

10. How should users think about using the new unified assist?

Simply ask your question and trust the system to determine the best processing approach. There is no need to understand the underlying technology or maintain a list of use cases for different assists.

11. What are the key benefits of this change?

Reduced user confusion.
Simplified experience with fewer assist options.
Less cognitive effort in selecting the correct tool.
Backend routing automatically determines the best processing method.
Faster adoption and improved usability.
Direct response to customer feedback.

12. How does Matter Assist work now that Document Assist and Matter Assist have been combined?

Matter Assist automatically selects the most effective analysis method based on the content being reviewed. For smaller document sets, it may analyse the full text directly. For larger matters, it may use advanced retrieval techniques to identify and review the most relevant information. This decision is made automatically behind the scenes.

13. Does Matter Assist still use RAG?

Yes. Matter Assist uses a combination of direct content analysis, retrieval-based techniques, and agentic retrieval workflows. The system automatically determines the most appropriate approach based on the volume and characteristics of the selected content, allowing Matter Assist to work effectively across a wide range of content volumes without requiring users to select a specific analysis mode.

14. Can Matter Assist generate timelines?

Matter Assist is not designed for creating timelines. Timelines should be composed using the Timeline functionality in Documents Assist or through AI Workbench.

15. Why don't I see citations in General Assist?

General Assist works differently from the other assists. It does not use the documents uploaded to the case as input; instead, it acts as a secure gateway to the broader knowledge base that the LLM draws from. Because it does not reference specific case documents, no citations are generated.

16. What do the numbers next to cited source documents mean?

The numbers next to cited source documents indicate the level of certainty that the piece of text from that document — used to formulate the output — was relevant to the query.

AI models, feedback, and usage

17. What AI models are used by Focused View?

Focused View's AI-driven functionality was purpose-built to use multiple large language models (LLMs) from various providers, including OpenAI and Anthropic. Opus 2 continuously tests new LLM versions as they are released and may update to newer versions when they outperform the current ones. Focused View currently uses Anthropic's Claude Sonnet 4.6 and Claude Sonnet 5.0.

18. What is the purpose of the thumbs up and thumbs down icons?

These icons allow users to indicate whether they are satisfied with the AI-generated output. Feedback submitted via these buttons helps Opus 2 monitor feature performance and act accordingly.

19. How can I track AI token usage?

System administrators can view their organisation's AI token usage directly within Opus 2 through the AI usage dashboard in System Admin. The dashboard provides an up-to-date view of your account-level annual token allowance, including total allowance, tokens used to date, and remaining balance.

Prompts and conversations

20. Are saved conversations user-specific or workspace-specific?

Saved conversations are user-specific and are scoped to a single workspace. The system does not delete saved conversations, but users can delete their own conversations at any time. Saved conversations can also be shared with other members of the case.

21. Are the prompt categories fixed or configurable?

The prompt categories are fixed. Users can add their own prompts through the Prompt Builder; user-created prompts are saved in My Prompts.

Permissions

22. How do permissions for viewing documents work in Focused View?

Permissions work as follows:

If no users are permitted to view a document, the document does not enter the AI workflow and is not displayed anywhere.
If a specific user is not permitted to view a document, it is unavailable for selection in the Source panel but is still listed. Citations are greyed out and unavailable for that user.
AI responses may still include insights derived from restricted documents.