Skip to main content
Cassidy’s Chat interface is where you interact with AI Agents, access your Knowledge Base, and run Workflows — all in one place.
Cassidy Chat interface showing an Agent conversation with Knowledge Base citations
To start a new chat, click + New Chat in the left sidebar or on the All Chats page.
New Chat button in the Cassidy sidebar
You can also chat through the Chrome extension, Outlook extension, Word and Excel extension, and external deployments like Slack, Microsoft Teams, Google Chat, embed, and API.

​
Select an Agent

Every chat uses an Agent. You can switch between custom Agents, foundational models, and Cassidy’s pre-built Agents at any time.
1

Open the Agent dropdown

Click the Agent dropdown at the top of the chat or in the text box to see your options.
Agent dropdown showing Sales Support Agent selected with a list of custom Agents and favorites
2

Browse and choose an Agent

You’ll see three categories:
  • Custom Agents — built by your team, trained on your Knowledge Base with tailored instructions
  • Default Agents — foundational models like GPT, Claude, and Gemini without your business context
  • Pre-built Agents — Cassidy’s ready-made options like the Data Analyst and Cassidy Help Agent
Default Agents don’t have access to your company’s Knowledge Base or custom instructions.
Claude 4.6 Sonnet selected as a default Agent with a banner noting it has no company context
3

Favorite an Agent

Hover over an Agent in the dropdown and click the star icon to add it to your favorites for quick access.
Agent dropdown showing the Add to favorites star icon on Claude 4.6 Sonnet
4

Switch mid-conversation

Need a different Agent? Click the dropdown again and select a new one. The chat continues with the new Agent — no need to start over.
When chatting with a Q&A Agent, responses include source citations that link back to your Knowledge Base so you can verify the information.
Chat response with Knowledge Base citations highlighted

​
Personalize your chat

Personalization lets you tell the AI about yourself so it provides more relevant, tailored responses in your chat conversations. These settings apply across all your chats regardless of which Agent you use.
1

Open a new chat

Click + New Chat in the left sidebar or on the Home page.
2

Open the personalization modal

Click the Personalize button in the top-right corner of the new chat screen.
Personalize button in the top-right corner of the new chat screen
3

Set your details

Fill in the following fields:
  • Name / Nickname — Pre-filled from your profile name. Change it to whatever you’d like the AI to call you.
  • Role — Enter your job title or function (e.g., “VP of Sales” or “Software Engineer”). This helps the AI understand your context.
  • Custom Instructions — Tell the AI about your preferences, work focus, or formatting expectations. For example: “I manage the enterprise sales team focused on AmLaw 200 firms. Keep responses concise and action-oriented. Use bullet points when listing key takeaways.” Supports up to 2,000 characters.
Personalization modal showing Name, Role, and Custom Instructions fields
4

Save your settings

Click Save to apply your changes, or Cancel to discard them.
Personalization modal with Save and Cancel buttons
Personalization settings only apply to your personal chat conversations within the Cassidy app. They do not apply to Workflow runs or external deployments like Slack, Microsoft Teams, Google Chat, embed, or API. Changing your name in personalization does not change your profile name elsewhere on the platform.
For organization-wide AI behavior rules that apply to every Agent and Workflow, see Global instructions.

​
Reference your Knowledge Base

You can pull specific items from your Knowledge Base directly into a chat to ask targeted questions.
1

Type the # key

In the chat text box, type # to open a search dropdown for your Knowledge Base.
Knowledge Base reference dropdown showing folders like Carrier Data, Call Intelligence, and Meeting Notes
2

Search and select an item

Start typing keywords to filter results. Click the item you want to reference — it appears as a tag in your message.
New Employee Onboarding Guide added as a Knowledge Base reference tag in the chat text box
3

Ask your question

Send your message with the referenced item. The Agent uses that specific document as context for its response. For example:
  • “Can you summarize the main points of this employee handbook?”
  • “How many vacation days am I entitled to according to this policy?”
  • “What is the process for requesting time off?”
The # reference is useful when you want to point the Agent at a specific document for a single question. To have an Agent always search certain Knowledge Base items, add them to the Agent’s configuration — see Build and configure an Agent for details.

​
Upload files and media

Share files, images, videos, and audio directly in chat to ask questions about their content. These uploads are temporary — they are not saved to your Knowledge Base.
1

Click the + button

In the bottom-left corner of the text box, click + to open attachment options.
Attachment menu showing file and image upload options
2

Choose your upload type

Select Attach a File or Attach an Image from the menu. You can also drag and drop files directly into the chat window.
3

Ask questions about the content

Once uploaded, the Agent can analyze the content. Try prompts like:
  • “Summarize the key findings in this report”
  • “What text is visible in this screenshot?”
  • “Extract the data from this chart”
Agent response analyzing an uploaded file
Uploaded files are temporary and tied to the chat session. If you need to store a file permanently, upload it to your Knowledge Base first, then reference it with #.

​
Use the tools picker

The tools picker lets you control which capabilities and connectors are active for a given chat. Click the Tools button at the bottom of the text box to toggle tools like Data Analysis, Web Search, Image Generation, Artifacts, and Browser Use on or off.
Tools picker panel showing toggles for Web Search, Data Analysis, Image Creation, Artifacts, and Deep Research

​
Use voice mode

Voice mode lets you have a spoken conversation with your Agent instead of typing. Click the microphone icon in the text box to start speaking. The Agent transcribes your speech, processes it, and responds — both in text and, optionally, with audio playback.
Dictate button and microphone icon in the chat text box

​
Use the writing editor

The writing editor is a side-by-side panel where you can draft and refine content while chatting with an Agent.
1

Open the editor

Click Open editor in the top-right corner of the chat interface.
Open editor button in the chat interface
2

Write alongside your chat

Type directly in the rich text editor while chatting with the Agent. This is useful for researching ideas, drafting content, or getting real-time writing assistance.
Writing editor open alongside a chat conversation
3

Add Agent responses to the editor

When the Agent generates something you want to keep, click Add to Editor on the message. The content is automatically added to your document.
Add to Editor button on an Agent response
4

Save to your Knowledge Base

Click File then Save to Knowledge Base to store your document permanently. Choose a name and folder location, then click Save.
Save to Knowledge Base option in the writing editor file menu
Save dialog with document name and Knowledge Base location fields
You can also export to Microsoft Word or PDF format.
Once saved, changes sync automatically between the writing editor and the Knowledge Base document.
The writing editor is tied to each individual chat. Save your work to the Knowledge Base before closing if you want to keep it.

​
Run a Workflow in chat

You can execute Workflows directly within a chat conversation to automate tasks without leaving the interface.
1

Open the Workflow menu

In the text box, click the Workflows button to see available Workflows.
Workflows menu showing a list of Workflows including Sales RFP, Marketing Blog Post, and Generate In-Depth reports
2

Select a Workflow

Choose the Workflow you want to run from the list.
3

Fill out the form

A form replaces the text box with input fields for the Workflow. Fill in the required information.
Generate In-Depth Company Report Workflow form with Company Name and Company Domain fields
4

Generate and review results

Click Generate to run the Workflow. Results appear directly in the chat.
Workflow results displayed within a chat conversation
5

Continue the conversation

After the Workflow runs, you can ask follow-up questions, refine inputs, or explore the results further in the same chat.

​
Use the prompt library

The prompt library lets you create reusable prompt templates so you can quickly apply consistent instructions across chats.

​
Manage your prompts

1

Open the prompt library

In the chat text box, type / and click Manage prompt library from the menu.
Slash command menu showing the Manage prompt library option
2

Create a new prompt

Click + New prompt. Describe what you want, and Cassidy generates a refined prompt for you. Or click Create prompt manually (advanced) to write it yourself.
New prompt creation interface in the prompt library
Prompt description field where Cassidy auto-generates a prompt
When writing prompts manually, you can:
  • Use [brackets] for placeholders that change each time (e.g., [professional] or [witty] for tone)
  • Type # to reference Knowledge Base items
  • Add example outputs so the Agent matches your preferred style
Prompt editor showing placeholders, Knowledge Base references, and example outputs
3

Save and use

Click Save to store it. You can then click Use prompt to apply it immediately.
Saved prompt with Use prompt and Edit buttons
Prompts created in the library are visible to everyone in your organization.

​
Apply a prompt in chat

1

Type / to open the prompt menu

The menu shows Cassidy’s default prompts (like Improve writing and Summarize) and your custom prompts.
Prompt menu showing default and custom prompts
2

Insert and customize

Click a custom prompt to insert it. Replace any placeholders with your actual values — these edits do not change the saved template.
Custom prompt inserted into the chat with placeholders highlighted
3

Send and get results

Send the message. The Agent generates a response using your prompt template.
Agent response generated from a custom prompt template
For default prompts like Improve writing or Summarize, a quick form appears where you paste your text and click Generate.

​
Analyze data

Upload spreadsheets or data files and let the Agent create charts, run calculations, and extract insights.
1

Select a data-capable Agent

Choose the Data Analyst default Agent, or use a custom Agent that has the Data Analysis capability enabled.
Data Analyst Agent selected in the Agent dropdown
2

Upload your data

Click + and attach a CSV or Excel file, or drag and drop it into the chat.
CSV file being uploaded to a chat for data analysis
3

Ask questions and request visualizations

Describe what you need:
  • “Create a bar chart showing sales by region”
  • “What are the key trends in customer behavior?”
  • “What was the total revenue for Q1?”
Data analysis chart generated from an uploaded spreadsheet
4

View the code behind the analysis

Click Analyzed sources in the results, then click the code icon (</>) next to any step to see the Python code the Agent used to process your data.
Analyzed sources dropdown showing processing steps
Python code modal showing the data analysis logic
  • Use CSV or Excel format — these are the only supported file types for data analysis
  • Organize data in rows and columns — each row is a record, each column is a variable
  • Include descriptive column headers in the first row — avoid blank rows above the headers
  • Keep data consistent — no empty rows between entries, no multiple tables in one file, avoid multiple worksheets
Well-formatted spreadsheet with clear headers and consistent data
The Data Analysis capability can also:
  • Manipulate branded templates — fill out forms, convert reports to presentations, edit Word documents and PDFs
  • Create synthetic data — generate sample datasets for testing or demonstrations
For more on generating files, see Create files in Cassidy.

​
Edit and regenerate messages

You can modify a sent message and have the Agent generate a new response.
1

Hover over your message

Move your cursor over the message you want to change. An edit icon appears on the right.
Edit pencil icon appearing on hover over a chat message
2

Edit and re-send

Click the edit icon, modify your text, and press Enter. The Agent generates a new response based on your updated message.
Chat message in edit mode with modified text
3

Browse response versions

Click the left and right arrows at the top of the Agent’s response to cycle through all generated versions.
Response version navigation arrows on an Agent message
You cannot edit Agent messages directly, but you can click Regenerate to produce a new response with the same prompt.

​
Branch a conversation

Branching creates a copy of a conversation up to a specific message, so you can explore a different direction without losing the original thread.
1

Find the branching point

Open a chat and scroll to the message you want to branch from.
2

Branch in a new chat

Click the … menu below the message and select Branch in new chat.
Branch in new chat option in the message menu
3

Continue in the new chat

A new chat opens in a separate tab with the full conversation history up to that message. The original chat remains unchanged. The new chat is labeled “Branched from [Chat Title].”
Branched chat showing the conversation history and branch label
Continuing a branched conversation in a new direction

​
Share a chat

Share chat conversations with team members so they can view or continue the discussion.
1

Copy the chat URL

Open the chat you want to share and copy the URL from your browser’s address bar.
Browser address bar with a Cassidy chat URL
2

Send the link

Share the URL with your teammate through Slack, email, or any other channel.
Keep these things in mind when sharing chats:
  • The recipient must be a member of the same Cassidy workspace
  • Workspace permissions apply — they may need admin access depending on your settings
  • New messages added to a shared chat appear in real time for everyone with access

​
Use Cassidy in multiple languages

While the Cassidy interface is in English, the AI models can understand and respond in virtually any language. Type your message in the language you prefer, and the Agent responds in the same language.
Chat conversation in Spanish with the Agent responding in Spanish
If the Agent doesn’t consistently use your preferred language, add a specific instruction to the Agent’s settings — for example, “Always respond in Spanish.”
Agent instruction field with a language preference set to Spanish
For more on customizing Agent behavior, see Build and configure an Agent.

​
Next steps

Build and configure an Agent

Create a custom Agent with tailored instructions and Knowledge Base access.

Choose the right AI model

Compare GPT, Claude, and Gemini to pick the best model for each task.

Prompt engineering guide

Write effective prompts for better, more consistent AI outputs.

Create files in Cassidy

Generate documents, spreadsheets, and presentations through chat.