All work on Jylo is contained within a Project. Projects are collaborative environments where individuals or teams can interact with artificial intelligence models via chat features and apply a product to document review.
All users will have a personal project created upon establishing an account that is specific to them. AI chat activities that the user wishes to undertake can be conducted in the user’s personal project.
There are two tiers of user within project:
| Project Administrator |
|
|
| Project Member |
|
All documentation relevant to the project can be securly stored on the platform ready to be analysed using the Chat and Document Review features.
Chat exists in three formats; Conversation, Document Analysis and Text Analysis. Conversation allows users to conversate with the selected model in a secure space, unlike browser versions, Jylo does not allow user information to be accessed by any AI provider. Document Analysis gives users the ability to interact with documentation saved within the project, users commonly ask for document summaries or to compare various documents simultaneously. Text Analysis performs the same function as Document Analysis but with Text which has been typed or pasted into the text field.
Rather than relying on each user to understand how to maximise the potential of AI, Jylo uses knowledge templates, known as Products which have been created by experts (Product Creators) to be used repeatedly for designated task across organisations and industries, see our article on Jylo Roles for more information. Products are a series of prompts which are applied in Document Review. Products are designed to allow users to navigate the dataset quickly and identify the source of information ready for instant human verification, an essential part of combating hallucination, the main complication in the effective use of AI. Products apply additional questions to individual documents to provide more specific and nuanced information found inside the document.
| 1 | Document count | Documents in the selected view. If only one document is selected, this will not show. |
| 2 | Filename | Name and extension of review documents |
| 3 | Filters | Used to narrow the dataset with results populated by AI. This is not present during single document reviews. |
| 4 | Filter buttons | Used to takes user to AI results within the document and present the model’s reasoning justification |
| 5 | Document viewer | Viewing screen for the document. Selecting a new filename switches the document within the viewer |
| 6 | Filter Justification | Source justification for filter responses |
| 7 | Verification buttons | Used to verify the AI’s output by either approving, rejecting, or editing the model’s output |
| 8 | Comments | Used to add comments to an AI’s output, for tracking purposes |
For help or more information regarding Document Review, see our Document Review Guide or any of our product demonstration videos.
Add comment
Please sign in to leave a comment.