Coda AI: Everything you need to know

Explore

Coda AI: Everything you need to know

Coda AI: Everything you need to know

Coda AI: context aware

In order for an AI-enabled tool to be considered a work assistant, it needs to fit within your existing workflows, not feel like a separate experience.

For example, if your goal is to write a product brief and then get feedback, you’d want both:

An AI integration that can reference relevant context, so that the first draft is helpful and informed.

A surface that makes it easy for your team to find it, leave comments, and upvote ideas.

⁠

⁠

An AI assistant should be able to understand relevant information about your team, and once it’s answered a question or created new content, enable you to actually use it in an impactful way.

AI that understands your work

Its like having a VERY knowledgeable assistant looking over your shoulder, who understands the content of your document, who can answer specific (and general) questions, editing their answers directly into your document.

- Max Xyzor, CTO, Agile Dynamics

Starting from scratch every time not only limits AI’s potential use cases but also adds unnecessary copying and pasting if you’ve already generated relevant context elsewhere. For example, if you’re writing a blog post about dogs, you usually don’t need to share much information with AI to get a satisfactory draft. If you are trying to create a project brief, or synthesize a weekly team update, then sharing the relevant context matters—otherwise AI won’t be able to generate content you can actually use.

When looking at the current landscape of AI-enabled productivity tools, most fall into the following categories:

Disconnected experience: Tools like ChatGPT can be accessed in a separate browser, and your team and project data live elsewhere.

Integrated experience, with limitations: AI integrations like Atlassian Intelligence (Confluence AI) and Notion AI can reference information you have within the tool, but you can’t specify what data is relevant. For example, if you are using AI to learn the latest customer feedback, you have two options:

Put all the data, notes, and other details on one page, then request a summary.

Ask questions based on all the information across your workspace(s), including data that’s outdated or needs to be cleaned up.

Integrated experience, with flexibility: Coda AI can reference any relevant information across your doc, and you get to choose which pages, tables, or text are relevant. For example, if you’re writing a product brief based on the latest notes and customer feedback, you can specify where these are located.

⁠

⁠

An AI integration that’s available in a surface where you spend all your time can do more than reference existing context—it can meaningfully contribute to your work.

AI that contributes to your work

While it can be helpful to look at any tool’s comprehensive feature set, as you explore the examples throughout this doc, I’d encourage you to think about how each one fits into your team’s broader use cases. Once your team finishes a task, there’s always a series of next steps that get put in motion, and opportunities for an AI work assistant to help.

For example, when your team finishes a writeup, what happens next?

⁠

⁠

⁠

For many, the next step would be to get feedback, either async or in a team meeting. Coda features like voting buttons give everyone a voice, and canvas columns help keep meeting notes organized.

After that, you might want to post a summary in Slack or Teams. Called

Packs⁠

, Coda has a robust integrations ecosystem, which allows you to sync in data from or send updates to hundreds of tools. You can explore

3. Task assistance⁠

for more examples of how you can use Coda AI with Packs.

Overall, an AI assistant should be an intuitive extension of your tools, designed to work within your processes, not outside them.

AI that respects your permissions settings

Many teams have a lot of context in Coda, but not all of it should be accessible to the whole company. Some information might be available at the team level—for example, the HR team hub is only available to its members. Other permissions might be more intricate—a manager’s notes on a team member’s performance might be limited to the two of them.

Coda AI leverages the security and access controls that power our search functionality today. Coda search works across all your Coda documents and folders, respecting the permissions of what you can access, and how the owners have set this up. So when you use Coda AI to reference available context, it can only access the information that’s been made available to you in Coda.

⁠

Ready to explore how teams are using Coda AI?

⁠

1. Writing assistance⁠

⁠

2. Knowledge assistance⁠

⁠

3. Task assistance⁠

Gallery

Want to print your doc?
This is not the way.

Try clicking the ⋯ next to your doc name or using a keyboard shortcut (

CtrlP

) instead.