AI Product Manager

Dan Shipper Portfolio

Co-founder and CEO of Every, which has shipped named AI products including Cora (email), Sparkle (file organization), Spiral (writing), Monologue (dictation), Proof (collaborative editing), and Plus One (a Slack AI agent), while running a media business alongside them.

AI-Native ProductsFounder-Led PMAI Writing ToolsProduct Portfolio

Location

Remote-first, in the most literal sense — no idea where.

Skills

AI-Native ProductsFounder-Led PMAI Writing ToolsProduct Portfolio

What makes it work

A breakdown of the choices that make this portfolio stand out.

1

Six named, shipped products instead of one flagship case study

Listing six distinct, named AI products under one roof shows a repeatable product-development process rather than a single lucky hit.

2

Media and product built as one deliberate system

Running a newsletter and podcast alongside the product suite means every launch has a built-in distribution and feedback channel — a structural choice worth naming.

3

A public, named argument for why the company can ship this fast

Publicly describing Every as operating with a very small engineering footprint relative to its output ties the product portfolio to a specific operating thesis, not just a list of outputs.

What AI Product Managers can take from this

Specific, actionable tips to apply to your own portfolio — no generic advice.

  • If you've shipped multiple small AI products rather than one big one, list them by name — range and repeatability are their own kind of evidence.

  • If you pair product work with an owned audience, explain how the two reinforce each other rather than describing them separately.

  • State the operating thesis behind how you built what you built — the "how" is often more differentiating than the feature list itself.

Case studies

Dan Shipper builds in public twice over: Every's software products, and the weekly writing and podcast episodes that explain how those products get made. The cards pair the products with the media that carries them.

Case StudyintermediateWeekly column; archive back to 2022

Chain of Thought

Every.toNewsletterEssaysHands-on model reviews

Shipper's weekly column on Every about AI and knowledge work. It mixes essays ('The Knowledge Economy Is Over. Welcome to the Allocation Economy.') with hands-on reports from building products, e.g. 'Agent-native Architectures' and what happens when a vibe-coded app goes viral.

It is a public record of product thinking written while he runs a company. Readers can check his calls against what actually shipped.

  • 'GPT-4 Is a Reasoning Engine' (Mar 2023) and 'The Allocation Economy' (Jan 2024) are listed as selected writing on his site

  • 2025–26 pieces cover agent-native apps and 'The Two-slice Team'

  • Earliest piece in his selected writing: 'AI and the Age of the Individual', Nov 2022

Case StudybeginnerOngoing; 100+ episodes

AI & I

PodcastInterviewsPublished transcripts

An interview podcast about how people actually use AI to think, create and work. Guests range from Every's own staff walking through internal workflows to outside founders and executives, and each episode is published with a full transcript.

100+ structured interviews about how people really use AI. That is user research, done in public.

  • Episode on how Every's team used AI to ship its biggest launch

  • Guests from Granola, Surge, GitHub and Figma

  • Transcripts published alongside every episode

Projects

SaaS ProductadvancedLive product; $20/month

Cora

Gmail / Google WorkspaceGoogle, Anthropic and OpenAI modelsEmail automation

An AI email assistant that sits alongside Gmail. It learns from your history which emails matter, pre-drafts replies in your voice into your drafts folder, and rolls everything non-urgent into a 'Brief' delivered twice a day. It deliberately cannot send or delete email.

The design choices are the lesson: drafts instead of sends, archive-to-Brief instead of delete. Scoping an agent's permissions to keep user trust is core AI product judgment.

  • Twice-daily Brief: non-urgent mail condensed into a 30-second read

  • No send or delete permissions, by design; Google Verified and CASA Tier 2

  • Every's compound-engineering method grew out of building Cora, per Every's own write-up

SaaS ProductintermediateAcquired 2014

Firefly

Co-browsingCustomer support softwareB2B SaaS

A customer-support tool that let two people browse the same web page together without installing anything. He co-founded it with Justin Meltzer while a student at Penn.

It is a full founder cycle with honest numbers. Slow early revenue, a focus on the customer-support market, then an exit.

  • $11,000 total revenue in the first 10 months, per First Round Review

  • Narrowed to customer-support teams as the core buyer

  • Acquired by Pegasystems in 2014

Ready to build your portfolio?

Follow the AI Product Manager roadmap — skills, projects, and timeline to get hired.

AI Product Manager Roadmap