[REDACTED] Episode 12: Prototype, Survey, Repeat, and Building the Product Loop With AI

On this episode of Redacted, we skip the guests and get back to the heart of the show: what it really looks like to build and run Offline with AI in 2026.

David walks through the Hot List, a prototype feed of restaurants opening around Raleigh. An AI monitors Instagram and local news the way a person would, and David built the front end by having Claude recreate Offline's existing app with the new data. Taylor then shows the other half of the loop: a homegrown survey tool that embeds the live prototype, plays an intro video, and asks AI-generated follow-up questions based on each person's answers.

From there, the conversation turns to strategy. We cover Offline's North Star metric, the inactive members they most want to reach, and the evidence-guided approach that starts with the most basic question of all: are people even going out to eat?

The episode closes with a bonus segment on rebuilding the company website in roughly an hour of human time, and what that says about how small teams can multiply their output with agents.

Timestamps

00:00 Cold open

00:45 Welcome back: just the two of us

01:55 Showing how it came together, not just what we built

02:40 The old prototype → interview → meeting loop

03:50 Introducing the Hot List

05:40 The Updates tab and opening-day reels

07:20 Back end: hiring an AI to watch the city