AI & Marketing Operations Projects

The Grad Trip

Planning a trip with zero boring tours

The Problem

Three countries, two weeks, one graduate who didn't want a single bus tour.

The Build

I connected AI to my Gmail and had it pull every confirmation scattered across months of booking: flights, cruise, hotels, excursions. It turned that mess into one clean itinerary, then built a follow-up survey so we could capture what actually worked once we got back.

The Framework

→Centralize the chaos first. Every confirmation lives in one inbox, let AI go find it instead of you digging
→Use it to surface the weird stuff, not the obvious stuff. Algorithms default to the top-rated tourist trap. Push past that
→Build the boring logistics (the itinerary, the timing) so you can spend your actual planning energy on the unusual finds
→Close the loop. A quick follow-up survey turns one good trip into a repeatable process for the next one

What We Actually Did

Custom handbags made by hand in Scotland. ATVs across Iceland's lava fields. Ziplining over rivers. A fish and chip shop with a terrible Google rating that turned out to be the best meal of the trip.

Custom Handbags, ScotlandLava Field ATVs, IcelandRiver ZipliningViking Photoshoot

Pipeline HQ

A real system, not a spreadsheet

The Problem

When I started taking on more fractional engagements, I had inbound leads, referrals, and follow-ups scattered across email threads and sticky notes. Spreadsheets don't scale once it gets real.

The Build

A full Notion command center. Linked databases tracking every prospect, every referral, every follow-up. An automated Gmail sweep that pulls inbound inquiries, scores each lead against a framework I built, and logs anything worth pursuing.

The Framework

→Define your non-negotiables first (scope, timeline, deal size) before you build any automation
→Score leads on fit, not just who reaches out first. Build the rubric before you need it
→Separate “in conversation” tracking from “referral relationship” tracking. They decay at different speeds
→Automate the boring sweep, keep the judgment calls human

What It Actually Tracks

Date of first contact, lead source, referral chain, current status, and a follow-up threshold that auto-flags anything silent for 30 days.

Discovery Call Prep

Prep built like campaigns

The Problem

Generic discovery calls get you generic scope. I needed prep that matched the actual stack and actual pain points of each prospective client, not a recycled pitch deck.

The Build

Custom prep packets per prospect. Pulls in the company's tech stack, the point of contact's background, and the specific operational problems they're likely fighting, then maps my own case studies directly to those problems.

The Framework

→Research the point of contact, not just the company. Their background tells you what they'll want to talk about
→Match your case studies to their stack. A Marketo win lands differently at a HubSpot shop
→Build reusable proof points once, then re-angle them per conversation instead of writing from scratch every time
→Scope and pricing belong in the proposal. Diagnostic depth belongs in the discovery call. Don't mix them

Result

Prep that actually tracks each prospect's real problems instead of guessing my way through a generic pitch.

Lead Scoring

A scoring system for signal over noise

The Problem

Inbound interest shows up everywhere: LinkedIn, referrals, cold outreach. It doesn't tell you what's actually worth your time.

The Build

A scoring system that takes every inbound lead and rates it against criteria that actually matter to me: deal size, scope fit, company stage, timeline. Logged automatically, reviewed weekly.

The Framework

→Write your scoring criteria down before you start seeing leads, not after
→Weight what you can't compromise on heavier than what's nice to have
→Review the log on a cadence. Weekly works. Don't let it pile up
→Kill dead-end leads fast. A clear “no” is more valuable than a maybe sitting in limbo

Result

A discovery system that filters noise before it ever hits my pipeline.

The Resume Review

Scored like a lead, not a formality

The Problem

Every MOps resume I screen gets judged the same six ways in my head. I got tired of only sharing that read one DM at a time.

The Build

An AI-powered resume and LinkedIn review that scores against the same criteria I actually use hiring: systems fluency, quantified impact, ownership versus task language, ATS keyword coverage, and unexplained gaps. Free, no email gate.

The Framework

→Score honestly. A resume with no metrics and no named tools should score low, not get graded on a curve
→Check formatting the way an ATS actually reads it, not the way it looks to a human eye
→Name unexplained gaps directly instead of letting them sit there as risk
→Give people a prompt to fix it themselves with AI, not just a list of problems

Try It

Upload your resume, get scored in under a minute.

Get your resume reviewed