Work

Autonomous prospect research that drafts the outreach and then stops at a gate you own.

The problem

Finding the right people and writing the first email is the slow part of outreach. The research is real work, the draft is real work, and by the time both are done the list has gone stale.

What was built

Scout researches the accounts and the people behind them, cites the evidence it found, drafts personalized outreach in the sender's voice, and then stops. Every draft waits at an approval gate you own; nothing sends until you say send. It was built for a real podcast-advertising campaign and then packaged so it can run again for any brief.

What happened

A real campaign ran through it end to end: the ranked shows, the people behind each one, the evidence Scout cites, the drafted outreach, and the log of what went out, all in one place. The walkthrough below is that campaign. Early access and the inquiry form live on the product page.

The trade-show platform: plan the show, work the floor with no signal, prove the ROI.

Lanyard runs an entire trade-show program: the plan and budget before the show, lead capture on the floor that keeps working when the venue has no signal, and the follow-up and ROI report after. It is a product of its own at thewiseoperator.com/lanyard.

Visitor check-in and routing for a front desk. Rolled out to twelve colleagues; in use every day.

A visitor signs in at the door, the right host is told in seconds, and the front desk sees the whole day on one board. Built for a real manufacturer's lobby, shipped with role-by-role manuals and five-minute onboarding, and now a product at lobbyrouter.com.

A messy shared drive turned into something a small team can actually find things in.

The problem

Years of files named by whoever saved them last: 403 loose at the root of one downloads folder, 3,627 in all, with duplicates, cryptic names, and no place for anything. Every search ended in a message to a colleague.

What was built

A small AI-assisted process, run once on the real folder: an inventory of every file, a folder taxonomy drawn from how the files were actually used, a human-readable name for every cryptic file, duplicates moved to an archive instead of deleted, and a short guide file in each project folder so the structure explains itself. I set the rules. The AI did the sorting. A person approved every move.

What happened

Every file preserved, nothing deleted, and a structure that new projects slot into without creating a new pile.

• 3,627files organized
• 403root files cleared
• 60+duplicates resolved
• 105files renamed
• 7guide files created
• 0files lost

A working revenue-operations demo: pipeline lanes, scenarios, drafted follow-ups. Click around.

The problem

A sales team's follow-up leaks. Quotes go quiet, the reminder lives in someone's head, and nobody chases in time. The revenue is not lost to a competitor; it is lost to silence.

What was built

Pipeline lanes that show where each deal sits, a scenario explorer for the mornings that actually happen, follow-up drafts written from the deal's own context, and pricing tiers for how the service would be packaged. Built as a demo anyone can click, on the same components a real install would use.

What happened

This one is a demo, and it says so. It exists to show a prospective client the shape of a follow-up system before they commit to one on their own CRM and inbox. The full page, with the pricing tiers and the questions people ask, is at /revops/.

A media brand teaching non-technical professionals to build with AI.

A daily AI brief, a plain-English AI dictionary, and an AI-workstation starter kit, written through the lens of applied wisdom. It is my other door: thewiseoperator.com.