Canopy

Inspiration

Our inboxes are full of information, but very little of it tells us what actually deserves our attention.

An important deadline can be buried between newsletters. An event invitation can arrive as an image. A message that requires action can look almost identical to one that doesn't. As emails pile up, it becomes harder to distinguish the signal from the noise.

We wanted to build something that could rise above that noise.

That idea became Canopy: an AI assistant inspired by the view from above a forest canopy. Instead of getting lost in the tangled undergrowth of emails, Canopy helps you rise above it, see what matters, and understand what's on the horizon.

What it does

Canopy uses Azure agents to make sense of incoming emails and sort them based on what they actually mean for the user.

With one click, users can analyze new emails. Canopy looks at both the text and images in an email to figure out what’s important and what can be ignored for now.

The dashboard breaks emails into three simple categories:

🔴 Needs Your Attention: Important emails that require some kind of action

🟡 Worth Knowing: Useful information, but nothing you need to act on right away

🟢 No Action Needed: Emails that don’t require anything from you

For emails that need attention, Canopy pulls out the important details, including a summary, key information, how relevant and urgent the email is, whether you need to do anything, and what the next step should be.

It can also pick out things like deadlines, amounts, eligibility requirements, and events. If there’s something you want to remember, you can add a reminder directly from the email. Reminders and events then show up in the Agenda, organized by day.

The idea is pretty simple: instead of making you dig through your inbox to figure out what matters, Canopy surfaces the things that actually need your attention.

How we built it

We used the deep.space SDK to build a work ready app that runs on React and typescript. We utilized Microsoft Azure's capabilities to analyze emails and assign every email an importance and urgency score, sort emails into the categories.

Challenges we ran into

Our biggest challenge was working with Azure Foundry and agents for the first time. We had to learn how to structure agents and use the Agent API to reliably extract, summarize, and classify information from emails.

We also had to decide how much control the AI should have. We considered automatically adding extracted actions to the Agenda, but decided users should choose which reminders get added. This kept the AI helpful without making decisions for the user.

Accomplishments that we're proud of

We're proud that we built a complete, connected experience rather than just an email classifier. Emails are analyzed, summarized, categorized, and turned into actionable information that users can review and add to their Agenda.

We're also proud of getting Azure Foundry and agents working in a project we'd never built with them before.

What we learned

Our reminder system was a good example that automating everything isn't always better. Giving users control over what gets added to their Agenda makes the experience more useful.

What's next for GirlHacks Submission

Canopy currently uses mock data, but we designed the application with real-world integrations in mind.

Our next steps would be to connect Canopy to real email providers and use the Azure AI to analyze live inboxes. We'd also like to build deeper calendar integrations, improve deadline and event extraction, personalize prioritization based on a user's behavior, and introduce notifications for things that genuinely need immediate attention.

In the long term, we see Canopy becoming more than an email organizer.

We want it to become a personal information assistant that can look across the messages, events, deadlines, and opportunities in someone's digital life and help them understand what's worth looking toward on the horizon.