Ask Your Email Archive Questions (and Get Smart Answers)

Imagine asking your inbox a question and getting a smart answer. Not searching through folders. Not trying to remember exact keywords. Just asking a natural question and receiving a synthesized response drawn from everything you've saved.

This is the shift from searching to asking—and it fundamentally changes how you interact with your personal knowledge.

How Email Q&A Works

The Q&A Pipeline

1. Send question to your Lolodex address
2. AI identifies intent as "ask"
3. Semantic search finds relevant notes
4. AI generates answer from your content
5. Response delivered to your inbox with citations

The entire process typically takes 10-30 seconds, depending on how much content needs to be analyzed.

Crafting Effective Questions

Good Question Examples

Question Formats That Work

• Specific > Vague: "What did Sarah say about pricing?" beats "What about pricing?"
• Include context clues: Names, dates, projects help narrow results
• Natural language is fine: Ask like you'd ask a colleague

Questions That Don't Work Well

• Too broad: "Tell me everything about marketing"
• No context: "What was that thing?"
• Outside your knowledge: Questions about topics you haven't captured

Understanding Answers

Anatomy of a Response

When Lolodex answers your question, the response includes:

• The answer: A synthesized response based on your notes
• Source citations: Links to the specific notes that informed the answer
• Confidence context: Indication of how much relevant content was found

When Notes Are Insufficient

If Lolodex can't find relevant information, you'll receive a response indicating this, along with:

• Suggestions for refining your query
• Option to trigger research mode for external information
• Related notes that might be tangentially relevant

Following Up

You can reply to continue the conversation:

• "Tell me more about [specific aspect]"
• "What else do I have about [related topic]?"
• "Can you summarize that in bullet points?"

Real-World Examples

Example 1: Finding Meeting Context

Question: "What were the action items from last week's product meeting?"

Answer: Based on your note "Product Sync - Jan 8," the action items were: (1) Sarah to complete design mockups by Friday, (2) Mike to investigate API performance issues, (3) Lisa to schedule customer interviews for next week.

Example 2: Newsletter Recall

Question: "What did Stratechery say about Apple's AI strategy?"

Answer: Based on your saved Stratechery newsletters, Ben Thompson argued that Apple's AI approach focuses on on-device processing for privacy, though this limits model capability compared to cloud-based competitors. Key insight: Apple is betting that privacy will be a differentiator as AI becomes more personal.

Example 3: Decision Context

Question: "Why did we decide against the dark mode feature?"

Answer: From your note "Dark Mode Feature - Decision," the team deprioritized dark mode for Q1 because: (1) Only 12% of surveyed customers requested it, (2) Engineering estimated 3 sprints of work, (3) It would delay higher-priority features. Decision was to revisit in Q2 based on updated customer feedback.

Tips for Better Results

Build a Rich Knowledge Base First

Q&A is only as good as the content you've captured. More notes = better answers. Forward valuable content consistently.

Use Consistent Terminology

If you always call it "the PM meeting," ask about "the PM meeting." Consistency helps the AI match your questions to your content.

Combine with Search

• Use Ask when you want synthesis or specific answers
• Use Search when you want to browse or explore

The shift from searching to asking changes how you interact with your own information. Instead of navigating folders and trying keywords, you simply ask what you want to know. Your knowledge base becomes a conversation partner that remembers everything you've ever captured.