Building in public // 2026

A lot of work
is still way harder
than it needs
to be.

I'm trying to fix that.

See what I'm building
workflow_v2.ts
L
System Status
Running smoothly
// automating the boring parts
const friction = analyze(work);
if (friction > threshold) {
buildTool({
target: 'real_work',
goal: 'practical',
public: true
});
} else {
keepIterating();
}
Section 02

Help me help you.

Tell me what's
slowing
you down.

AI isn't magic — but it is a very good lever. Tell me what's tedious, broken, or boring in your week. Let's see if we can build something practical for it.

"If you find yourself copying and pasting the same thing more than twice a week, we should talk."

Common frictions

Read by a human. No automated jargon.
Section 03 — Experiments

Featured Builds

Less theoretical roadmaps, more practical tools turning chaos into structure.

Active Experiment #CRE_AUTOMATION

Meeting Notes → Action Items Pipeline

A three-step workflow that takes raw CRE meeting notes, extracts structured action items with owners and deadlines, and drafts a professional follow-up email — in under 90 seconds.

01
02
03
A
In Testing

Lease Renewal Analysis Automator

A two-step Excel + PowerPoint workflow that turns raw comp data into an executive-ready lease renewal memo — showing exactly how far automation can go without manual templating.

Personal Tool

Tone Matcher Prompt Library

A dynamic prompt pipeline that analyzes writing style on the fly and helps generate first drafts that actually sound like me — not like a generic AI assistant trying to be helpful.

Section 04 — Resources

Library

Prompts worth keeping, workflows worth copying, honest experiments worth learning from.

All AI Workflows Prompts Worth Keeping What I Tried Building in Public Honest Takes
AI Workflows
01
Prompts Worth Keeping
02
What I Tried
03

// Browse the full library at lilyp.ai/library

Section 05 — Live

Now /
Currently
Testing

What's live, what's messy, what's being figured out right now. This updates when things change.

Live
Building lilyp.ai

Designing and building this site in public using Claude Code. The whole thing is an experiment in itself.

In Progress
Meeting Notes Pipeline Demo

Recording a full walkthrough — meeting transcript → Claude → structured action items → Airtable, live.

Up Next
September Content System

Two posts per week for 8 weeks, building an audience before telling anyone I'm building an audience.

Still Messy
The Library Architecture

Category structure, URL design, and how to make it actually useful instead of just another link dump.