Skip to content

PDQ.com / Personal

My Projects

Shipping AI tools, not just talking about them.

The Challenge

AI ideas are everywhere. The problem is building AI tools that actually change how work gets done — tools reliable enough to trust, specific enough to be useful, and simple enough that people actually use them.

The Solution

I've been building AI tools at PDQ and on my own time: automations that process real data at scale, GPTs trained on competitive intelligence, and a growing system for using AI agents to accelerate product development itself.

What I'm Building

My Projects

Voice of the Customer

PDQ collects customer feedback across call transcripts, help desk tickets, NPS surveys, and more — but that data was scattered, inconsistent, and too noisy to act on at scale. I built an AI agent that automatically ingests all of it, filters out noise, extracts valid feedback, and categorizes it by theme and product area. The next step is surfacing that as a living knowledge base any employee can tap into for research or roadmap planning, without reading a single raw ticket.

6 weeks

of manual work saved every quarter

Voice of the Customer

A GPT that never stops watching the competition

Describe the problem you're solving and the Jobs to be Done you're designing for, and this GPT does the rest. It scrapes competitor documentation, social media, and other relevant sources, then surfaces screenshots and pertinent information you can pull directly into your research. No more manually hunting through five different products to figure out what the market is doing.

1 week

of research compressed into minutes

A GPT that never stops watching the competition

Vibe coding a proof of concept at PDQ

I wanted to know if vibe coding could produce something technical, something to give our engineers a head start. The result: a managed software center that installs software on a device, scans to confirm it's there, and logs each deployment stage in a working mockup of PDQ Connect. It ran in a local environment, but the loop was complete: deploy, verify, report.

Months

of traditional dev time saved on a working POC

Vibe coding a proof of concept at PDQ

Up next: MatMaster

I've spent a year designing a tournament management platform for wrestling. Bracket management, registration, live scoring, athlete profiles. I have the design. Vibe coding is the mechanism I plan to use to bring the full product to life.

Up next: MatMaster

Results

Impact & Outcomes

7+

weeks saved every quarter

6 weeks from automating feedback processing with the VoC agent, plus 1 week from on-demand competitive research.