Conversing with documents

Converse allows you to have a conversation with your documents, getting on-demand information from documents of nearly any type or format. A conversation can also act as the starting point for a chatbot.

Creating conversations

Conversations can be created in your personal workspace.

• From the homepage, click Create on the Chats panel. An empty conversation opens in Converse.

• In Workspaces, select your personal workspace, then click Create > Chat.

An empty conversation opens. While the chat panel on the right displays some sample prompts, your first step before you can submit a meaningful query is uploading files to your conversation.

Uploading files

In an empty conversation, you can use the center pane to upload files. You can upload a variety of file types in various supported languages. You can also add documents present in a connected drive, assuming the drive is accessible to your personal workspace.

Upload files to your empty conversation by doing one of the following:

• Drag and drop files into the center pane.

• Click the links present in the Select files or folders, browse drives, or paste a URL text in the center pane to open your local file explorer, the AI Hub file explorer, or the URL upload dialog.

  When using URL upload, you can upload files available at a URL or you can import a webpage. Content uploaded by URL is captured as a PDF, meaning the content doesn’t refresh and can’t reflect future changes. Inline links might not be preserved. URL upload is supported for public websites without paywalls or loading animations.

Uploaded files are digitized and stored in your default drive. After uploading files, two new panels are visible alongside your chat panel:

• The document list on the left displays all documents uploaded to the conversation. From here, you can:

  • Upload more files to the conversation by clicking the Add files icon .

  • Permanently delete selected files by clicking the Delete selection icon .

  • Control your message scope---the set of documents you’re conversing with in each query---by selecting and deselecting files.

• The document view in the center displays the file selected in the document list. The view pane includes a toolbar, auto-hidden by default, with controls for viewing the selected document, including image or text-only views, keyword search, and page selection. When object detection is enabled, you also see a highlight icon at the top of the document view pane for documents containing detected objects.

You can hide the document list and document view by clicking the X at the top of the view pane. To bring both panels back, click the Open file panel icon floating in the top right of the chat panel.

Digitization and object detection

All uploaded files are digitized according to the default digitization processes. If encountering low-quality responses or using documents that might benefit from advanced digitization and object detection settings, you can modify your conversation’s digitization settings.

In your conversation’s digitization settings (Settings > Digitization), you can preview how changes impact machine-readable text with up to three documents from your conversation.

Any time you change digitization settings, all files in your conversation are redigitized.

Choose the digitization settings suitable for your documents and AI Hub subscription. For details about optical character recognition (OCR) support for various languages, see Supported languages.

• Tables — Provides better results when extracting information from tables, and enables table highlighting, which lets you enlarge, copy, or download highlighted tables directly from the document viewer.

• Checkboxes — Provides better results when extracting information from checkboxes.

  Table and checkbox recognition change the OCR processor used, which slows digitization slightly and might impact accuracy, particularly with less common languages. Enable tables and checkboxes only if needed.

• Non-Latin characters Commercial & Enterprise — Enables support for many common languages that use writing systems other than the Latin alphabet (a, b, c…). Support for non-Latin characters is offered in standard and advanced language sets. For details, see Supported languages.

• Process spreadsheets natively — Processes Excel spreadsheets in their native file format instead of converting to PDF. This option offers better results for wide tables, but doesn’t support embedded objects or source highlighting in results.

• Treat files as images Commercial & Enterprise — Digitizes files as they appear, discarding any embedded machine-readable text. This option often provides better results for documents that use non-Latin characters, handwritten text, and visually complex documents.

• Pages Commercial & Enterprise — Limits digitization to specified pages.

Conversing with your documents

Conversing with your documents is straightforward: type a message in the chat box, then press enter or click the send icon . The model provides a response and lets you know which documents it referenced when generating that response by listing them below the response. Selecting one of the listed documents opens it in the document view panel to show the provenance or information source—where Converse found the relevant information.

To get the most out of Converse, it helps to understand what functionality is supported. The following sections provide an overview of supported features, querying techniques, and more, grouped into the following categories:

• Defining message scope — Learn how to narrow your query to a subset of uploaded documents.

• Models and research mode — Understand which models are used to answer queries and the capabilities they offer.

• Asking great questions — Review advice for writing better, more specific queries.

• Specifying result formats — Explore supported response formats beyond plain text.

• Querying detected objects — Learn how to fully leverage AI Hub’s digitization and object detection processes.

• Querying visual elements — Trigger visual reasoning in research mode so you can query visual elements such as images, diagrams, and markup.

And finally, user interface tips collects miscellaneous tips to make sending queries and understanding responses even easier.

Defining message scope

Message scope is the set of documents that you’re conversing with in each query. By default, the message scope for a query is all files uploaded to the conversation, but you can modify this to narrow your query to single or specific documents.

The chat box displays your current message scope at the bottom, such as Message scope: All files or Message scope: 5 files. For more details, open the document list: the files selected in the document list are your current message scope. You can edit your message scope by selecting and deselecting files in the document list. Click Select all or Clear to quickly select or deselect all.

Models and research mode

Converse automatically chooses the best model for your query based on the number of documents included in your message scope. Understanding this automatic selection helps you not only be aware of how many consumption units a query consumes but also understand the selected model’s capabilities.

• If the query is directed to a single document, the advanced model is automatically used.

• If the query is directed to multiple documents, the multistep model is automatically used.

Regardless of your message scope, you can enable research mode for any query by turning on the Research mode toggle in the chat box. Research mode uses a more powerful variant of the multistep model and is suited for complex reasoning queries, but can result in longer execution times.

Asking great questions

While Converse can certainly answer straightforward queries such as What were the total tax deductions from this paystub? or Do all these documents have the same address for Jane Smith?, getting the most out of Converse means understanding how to ask great questions.

Provide clear instructions

Use verbs like extract, identify, calculate, find, explain, and summarize, depending on your documents and question. If you want the answer in a structured format such as a table or in JSON, specify the result format explicitly.

Ask the model to think step-by-step

Especially for complex tasks, ask Converse to think step-by-step by adding “explain step by step” to the end of your query. When the model focuses on each task individually, it improves the accuracy of each response.

Provide additional context

Give Converse more context about your question to help it better understand what information you need.

Specifying result formats

Converse can return responses in plain text or in other formats, including rich graph formats such as tables, lists, charts, and code blocks. These result formats can be copied or converted into other formats and downloaded.

Available formats for Converse responses include:

• Tables, which you can copy in tab-separated values (TSV) format or download as CSV files.

• Charts, including line, bar, column, pie, scatter, and multi-axis. Charts are also downloadable as CSV, PNG, or SVG files.

• Code blocks, with more than 25 formats available, including JSON, Python, bash, and JavaScript.

To get results in a specific format, ask for that format in your query to Converse. For example:

Querying detected objects

You can use Converse to query detected objects such as tables, checkboxes, signatures, and barcodes.

Tables

To extract information from all tables in a document, begin your query with Extract all tables. To get results from multiple tables in a specific format, specify either Markdown or JSON. For example:

• Extract all tables and return in Markdown

• Extract all tables and return in JSON format

To extract information from a single specific table in a document, include the title or header of the table in your query. For example:

• Extract the transactions table for the month of January 2023

• Extract the monthly transaction summary for the month of January 2023

Converse returns single-table extraction results as a table, which you can copy or download as a CSV file.

You can also filter columns or rows, sort columns, and perform other manipulation of table data. For example:

• Extract transactions and filter for amounts greater than $1,000

• Extract transactions and return results for 01 May through 15 May

• Extract transactions table and sort amounts from smallest to largest

• Extract transactions and add a column Flagged with values set to Yes if the debit is greater than $70

Checkboxes

You can extract information from checkboxes in single or multipage documents.

• For a group of checkboxes with a label, such as the Filing Status field on a tax form, use a query asking about which checkboxes are selected, such as What filing status is claimed?

• For a standalone checkbox, use a query that indicates whether the checkbox is ticked. For example, Is the filer claiming capital gains or losses?

Signatures

You can extract information about signatures, including whether a document is signed, who the signer was, and the signature date. Extraction of signature images isn’t supported.

For example, you can ask Converse:

• Extract all signatures

• Is this document signed?

• Who signed this document?

• Are these documents signed by the same people?

Barcodes

You can extract information about barcodes and their embedded values. Both numeric and non-numeric formats and one-dimensional and two-dimensional formats, such as PDF417 or QR codes, are supported. For example:

• Are there any barcodes in this document? If yes, what are their values?

• How many barcodes are in this document?

• What does the QR code in this document link to?

Querying visual elements

With research mode enabled, you can leverage visual reasoning capabilities. The model can analyze visual and stylistic elements, including elements that OCR doesn’t capture, such as images, diagrams, watermarks, layout, colors, text styling, and handwritten markup. Think of it as the model being able to “see” your documents and answer questions about them accordingly. You can ask the model basic questions like What color is the car in this image, or make more complex requests like asking it to describe a diagram.

To leverage visual reasoning, include keywords in your query, such as “image” or “diagram”, or specify a visual element. For example:

• Explain what happened in the car accident based on the accident diagram.

• Is any text in @contract-agreement.pdf crossed out?

• Based on the shaded areas in the pain diagram, where does the patient report experiencing pain?

User interface tips

• To resend a previous query, click the chat box and press the ↑ (upwards arrow) key to populate your last query in the chat box. You can use your arrow keys to move up and down your conversation history. Use this shortcut for editing previous messages or resending the same message with a different scope.

• You can direct Converse to prioritize referencing specific files by including @file-name in your message. Mentioning files is an alternative to narrowing the message scope to only that file. You can also use the @file-name approach to direct Converse to compare one mentioned file against another.

• When using research mode, it’s noted in the response window. You can make use of this to compare how enabling research mode affects responses to the same query.

• You can provide feedback on low-quality responses by hovering over the response and clicking the thumbs-down icon . Your feedback is used to improve model quality and might be viewed by Instabase. If you add a feedback message, don’t include any personally identifiable or sensitive information.

Deleting conversations

Deleting a conversation permanently deletes the conversation itself as well as your conversation history. You can also delete only your conversation history by clicking Clear chat.

• In Workspaces, open your personal workspace and locate the conversation to delete. Click the overflow icon then select Delete conversation. Click Delete to confirm.

• In Converse, hover over the left side menu to display a list of your conversations. On the conversation to delete, click the delete icon . Click Delete to confirm.