Files in Athena are called Assets. You can easily upload the supported file types and interact with them using Athena Chat. Athena supports the following asset types:
Athena supports page-level citations in PDFs!
Athena enhances document analysis with its capability to support page-level citations in PDFs. This allows you to pinpoint specific pages within a document that are referenced during analysis, ensuring transparency and accuracy in the information retrieval process.Whether you’re summarizing, extracting data, or generating reports, Athena’s page-level citations provide granular details that make your outputs more reliable and trustworthy. You can click on the citation to scroll to the highlighted source behind Athena’s response.
1
Upload assets
There are multiple ways to upload assets in Athena. The simplest way is to drag and drop assets from your computer onto the file tree view on the left.
2
Interact
Enter the following prompt in the Chat window on the right. Press Enter.
Prompt
What role does the Operations Room play in supporting Red Bull Racing’s trackside performance during races?
This section provides an overview of all the different capabilities offered by Athena Chat.
Choose from a number of toolkits designed to optimize and automate workflows. Toolkits are specialized suites designed to handle discrete tasks efficiently within larger workflows. They are the building blocks of complex processes, each performing a specific function such as document analysis, email management or coding tasks.
Toolkit Suggestions - Athena will automatically suggest relevant toolkits based on the assets you have open and the content of your chat messages. For example, if you have a document open, Athena may suggest enabling the Document Toolkit. Similarly, if your message mentions creating a spreadsheet, Athena will suggest the Spreadsheet Toolkit. Simply click Enable to activate the suggested toolkit.
Toolkits
Description
Canvas
Collaborate and visualize ideas interactively on a creative platform designed to bring strategies, designs, and brainstorming sessions to life.
Collection
Create and manage collections of assets for organizing and grouping related content
Visualizations
Create and manage interactive visualizations, charts, dashboards and figures
User Interface
Create and deploy interactive Athena user interfaces with tool registration and Athena SDK integration.
AOP
Create, update, and execute Agent Operating Procedures (AOPs) with template variables and structured output schemas.
Browser
Have Athena control a Chrome browser for web navigation, interaction, downloads, screenshots, and automated web tasks.
Presentation
Create professional presentations with support for slides, layouts, and visual content.
Notebook
Execute code-based workflows, analyze data, and run simulations in an AI-enabled environment for technical and analytical tasks.
Spaces
Manage your files and folders effortlessly with a centralized toolkit that supports organization, sharing permissions, and collaboration.
Python
Run Python scripts and automate complex computational tasks, making it a go-to tool for developers and data scientists.
Email
Manage email workflows by drafting, responding, and organizing communications with AI-powered insights and automation.
Spreadsheet
Optimize data manipulation and analysis with tools for tabular data processing, including formulas, summaries, and visualizations.
Document
Seamlessly create, edit, and analyze text-based files with Athena’s Document toolkit, which supports workflows like summarization, information extraction, and report generation.
Agent Personas allow you to optimize Athena’s behavior for different types of tasks. Each persona is designed for specific use cases to help you get the best results.
When to use each persona:
Chat - Best for brainstorming, Q&A, summaries, and ideation. Read-only mode for fast responses without modifying your workspace.
Example prompts: “Summarize this PDF,” “Give me 5 approaches to solving this problem,” “Explain this concept in simple terms”
Use when: You want quick answers or insights without creating or modifying files
Assist (Default) - Best for day-to-day work that requires reading and writing. Uses tools to draft emails, create documents, modify spreadsheets, and execute AOPs. Balanced speed and depth.
Example prompts: “Draft a customer reply and save as a Doc,” “Pull these rows into a Sheet and create a chart,” “Create an AOP to analyze quarterly reports”
Use when: You need Athena to take action and create or modify assets in your workspace
Deep Think - Best for multi-step, complex reasoning across multiple assets. Slower by design and may use multiple agents working together for comprehensive analysis.
Example prompts: “Compare these 5 contracts and produce a redline memo with citations,” “Analyze this dataset and identify trends, anomalies, and recommendations”
Use when: You have complex problems requiring thorough analysis across multiple documents or data sources
Tip: Pair with the Effort Dial for even more rigorous analysis
Not sure which to pick? Start with Assist for most tasks. You can switch personas as your needs evolve during the conversation.
The Agent Context menu lets you configure contextual knowledge that Athena has access to, for example, knowledge base, memories and workbench. Learn more about these in our guides on contextual knowledge.
Learn Preferences - Toggle this on for Athena to learn your preferences from your conversations, such as formatting styles, tone, and workflow patterns.
Use Preferences - Toggle this on for Athena to apply your learned preferences to personalize her responses and outputs.
Effort Dial - Have Athena execute a request for a dedicated period, making her spend a minimum amount of time thinking & processing before responding.
AOP Library - Use this button to view or create Agent Operating Procedures (AOPs) which enable Athena to dynamically execute multi-step workflows.
Transcribe - Use Athena Voice to give speech commands to Athena for her to transcribe and execute your requests.
LLM Selector - Choose from a wide range of LLMs based on the complexity of your task.
