Podcast Summary: AI and I — "How We Built Our AI Email Assistant: A Behind-the-Scenes Look at Cora"

Host: Dan Shipper
Guests: Kieran, Brandon, Natasha
Date: June 26, 2025

Episode Overview

In this special episode, Dan Shipper brings together the core team behind Cora, an AI-powered email assistant, to celebrate its public launch. The discussion provides an in-depth, behind-the-scenes look at the process of building Cora: the product philosophy, technical challenges, team dynamics, and reflections on how Cora transforms the email experience for its users. The team shares memorable moments, lessons learned, and their vision for the future of AI-powered productivity tools.

Key Discussion Points and Insights

1. What Is Cora? (05:07–09:10)

• Cora’s Purpose:
  • Functions as a “chief of staff” for your inbox—summarizing, categorizing, and drafting responses to emails.
  • Offers productivity and peace of mind for $15 a month, versus hiring a human assistant.
• Core Features:
  • Screens Emails: Delivers only urgent or human-generated emails to the main inbox.
  • Drafts Replies: Prepares response drafts based on the user's writing style and previous emails.
  • Daily Briefs: Twice daily summaries of emails requiring reading but not replying, easily skimmable in under a minute.

Quote (05:07):

“It’s sort of like having a chief of staff in your inbox for 15 bucks a month instead of 150k a year... I just have that kind of ‘it’s handled’ feeling.”
— Dan Shipper

2. Product Walkthrough & User Experience (05:56–10:06)

• Visual and Usability Distinction:
  • Briefs come with a visually appealing interface (e.g., dynamic oil painting backgrounds for different times of day).
  • Sections like "important info," newsletters, payments, etc., all summarizing and categorizing emails for easy digestion.

Quote (07:13):

“The most important stuff is on the top. It’s called ‘important info.’ It means you should know about it, but it doesn’t mean you have to take action.”
— Kieran

• Language Adaptation: Summaries respect the user’s language settings.

3. Product Philosophy: Opinionated Design (10:06–12:43)

• Starting Point:
  • The team’s agreed dislike of email shaped Cora’s “opinionated” approach—automating away the inbox as a perpetual, externally-written to-do list.
• User Response:
  • Polarizing but effective—"80% of people absolutely love it... 20% absolutely hate it."
  • The experience can be initially uncomfortable ("anxious attachment" to inbox), but most users find relief and new productivity.

Quote (10:46):

“Email is a to-do list written by everyone but yourself.”
— Brandon

Quote (12:11):

“We still had that urge to check and to refresh... But there was nothing. It was just... quiet.”
— Kieran

4. Technical Journey: From MVP to Customization (16:19–24:17)

• Early Iterations and Insights:
  • Realization that only 10% of emails require a response; the remaining 90% needed efficient sorting and summarization.
• Personalization via Assistant:
  • Introduction of the in-inbox Assistant: Users can train Cora on personal preferences (e.g., flagging kids’ daycare emails as important).
  • Supports natural language interaction (“vibe-based rules”), adapting the brief and inbox dynamically over time.

Quote (19:42):

“Your email is the story of your life... Cora can use AI to pick out things, create memories and learn about you.”
— Kieran

• Technological Complexity:
  • Balancing traditional rule-based algorithms and advanced language models (LLMs) for classification and summarization.
  • Huge challenges around cost, LLM limitations, and scaling from a few users to thousands.

Quote (22:57):

“It is really hard. There are many moving parts... What is the right thing? The right thing is also different for everyone.”
— Kieran

5. AI Tools in Development Workflow (27:34–32:58)

• Game-Changing Tools:
  • Monologue: A smart dictation app, improved brainstorming and code-writing through voice, delivering richer context to LLMs.
  • Claude Code: Revolutionizes how software is built and maintained.
• “Compounding Engineering” Mindset:
  • Focus on recording workflows and feedback, distilling operational knowledge into reusable prompts and procedures.
  • Creating a library of “muscle memory” for both engineering and editorial processes.

Quote (31:27):

“Instead of doing the work, you do the thing that does the work going forward... So instead of always doing a code review, you’re recording that so that it compounds.”
— Dan Shipper

6. User Feedback and Product-Market Fit (36:07–40:21)

• Community & Engagement:
  • Cora reached 10,000+ waitlist signups and 1,000+ DAUs even before public launch.
  • Users are reporting lifestyle transformation; many become “super fans.”
• Pricing Lessons:
  • Initial transition from free to paid was difficult; learning the importance of not “taking things away,” but found a price point and model that works.

Quote (38:02):

“It’s very special... to work on something that is not even publicly available, but has a 10,000 person waitlist... People love this.”
— Dan Shipper

7. Creative Uses and The API Mindset (40:21–42:30)

• Users have creatively integrated Cora with workflows (e.g., forwarding emails to trigger AI processing, using Zapier for categorization).
• Cora’s assistant acts like an API—users can forward emails and have the assistant take actions (unsubscribe, recategorize, summarize).

8. Vision for the Future (42:30–45:05)

• Short-Term Roadmap:
  • iOS app in development
  • Assistant to become more capable and possibly entirely replace legacy systems
  • Aspirations to build a fully immersive alternative to Gmail—where all email activity happens within Cora.
  • Unified brief across multiple inboxes (personal and work).
• Long-Term Aspirations:
  • “What would it be like to work with the same assistant for 10 years, who knows everything about you and can help you with anything?”

Quote (43:30):

“I think I’m just done with Gmail. I want to be in Cora at all times... I want it to be the whole thing.”
— Brandon

Quote (44:29):

“I want to just make whatever we do with categorization and summarization... a hundred times better.”
— Kieran

Standout Quotes and Notable Moments

• On the product’s emotional impact:
  “It felt emotional. It... we felt chill, like we could feel it, which is really cool if software makes you feel a certain way.”
  — Kieran (15:30)
• Competing with simple expectations:
  “I love making products that are so complicated in the backend that they can be described as simply as ‘just an email summarizer.’”
  — Brandon (26:20)
• Summing up the product’s philosophy:
  “Email is a to-do list written by everyone but yourself.”
  — Brandon (10:46)
• On building for fans:
  “I just like the fact that it's so opinionated and because of that people have come back and been like this has changed my life.”
  — Brandon (38:59)

Timestamps for Major Segments

• Product introduction and philosophy: 05:07–12:43
• First experiences with Cora: 12:43–15:30
• Deep dive: personalization and assistant features: 16:19–22:57
• Technical discussion: categorization and AI challenges: 22:57–27:34
• AI workflow tools and compounding engineering: 27:34–32:58
• User feedback and community stories: 36:07–40:21
• Creative API-like use cases: 40:21–42:30
• Roadmap and team’s future vision: 42:30–45:05

Summary

This episode provides a lively, detailed, and candid insider’s tour of how Cora was conceived, prototyped, stress-tested, and brought to market. The team’s mix of technical rigor and honest introspection reveals both the challenges and joys of building truly useful, opinionated AI products for real people. Cora isn’t just a new email tool—it represents a philosophy shift, aiming to give busy professionals their time, focus, and sanity back. The team’s dedication to continual improvement—both through technology and by listening to users—sets the stage for a product that will keep evolving as AI advances.

Try Cora: cora.computer