← All work

2018 - Present / eCommerce / AI Systems / Growth

BlackboxMyCar

Director of Marketing & New Business

Instead of using AI as a writing shortcut, I used it as an operating layer.

Pulling data, QA'ing attribution, surfacing revenue leaks, drafting executive-ready recommendations, monitoring customer demand, and turning fragmented systems into repeatable business workflows across 10+ platforms.

Visit website
Case study

eCommerce / AI Systems / Growth

10+connected business systems running on AI-assisted workflows
1,600+reporting and QA artifacts produced
250K+support-ticket records analyzed
12.14xaverage monthly paid media ROAS
25%revenue growth after the Shopify rebuild
Connected platform stack
Connected platform stack: Shopify, Google Ads, Meta Ads, GA4, Klaviyo, Gorgias, Merchant Center and more feeding into one AI operations layer.

Before

Disconnected systems, manual everything, and data nobody trusted

  • The business had deep category expertise and strong customer trust, but growth created friction. The Shopify storefront, paid ads, email marketing, social media, product launches, and customer support all ran as separate silos with no shared data or process connecting them.
  • Manual workflows dominated. Weekly price changes, promotions, and product feed updates were done by hand across two separate stores (US and Canada). The team of 7 had no standardized operating rhythm. Reporting lived in spreadsheets that took days to compile. Nobody trusted the numbers enough to make fast decisions.
  • AI was available but was being treated as a writing shortcut. The real opportunity was deeper: building AI-assisted systems that could pull data, check it for quality, surface what mattered, and produce decisions rather than more spreadsheets.

After

An AI-powered operating layer driving measurable business decisions

01

AI-Powered eCommerce Command Centre

Built a connected operating system that turned fragmented data into faster decisions across the whole team. Instead of waiting days for spreadsheets, leadership got daily readouts on what mattered: revenue, channel performance, product winners, and tracking gaps. 1,600+ decision-ready artifacts replaced manual reports, giving everyone alignment on real data that actually moved the business forward.

02

Executive Reporting Automation

Automated reporting that turned raw channel data into clear business decisions. Built recurring daily sales reports, weekly paid media reviews, monthly marketing reports, and strategic CMO-style memos. Generated 100+ daily sales reports and 50+ weekly reporting deliverables, giving leadership faster readouts on revenue, AOV, order volume, channel performance, product winners, tracking gaps, and next actions.

03

Support Intelligence and Revenue Attribution

Turned customer support from a reactive inbox into a measurable revenue channel. Analyzed Gorgias ticket history, pre-sale tags, response coverage, AI handoffs, CSAT, macros, and Shopify order matching across 250K+ support-ticket metadata records. Discovered that replied pre-sale conversations converted dramatically better than unreplied ones, reframing support coverage as a revenue lever.

04

Paid Media Waste Reduction

Built AI-assisted workflows to audit Google Ads, Meta Ads, Merchant Center, GA4, and Shopify together. Identified products getting impressions or spend without conversions, then built exclusion and negative-keyword workflows to reduce wasted Shopping/PMax exposure. Cleaned conversion-goal noise so purchase actions stayed central to bidding. Created 590+ GA4 backfill artifacts, 120+ Shopify-to-GA4 reconciliation reports, and 40+ tracking coverage reports.

05

Feed Health Automation

Built feed-health monitoring and cleanup across US and Canada stores covering product approvals, disapprovals, availability mismatches, local feed issues, refurbished-product leakage, and ad/free-listing destination status. Ran 100+ feed-health and cleanup workflows, protecting thousands of product/variant records from bad feed states and wasted ad eligibility.

06

Competitive Pricing Intelligence

Stayed on top of competitors across US and Canada storefronts. We matched or dropped pricing whenever we caught a gap, enabling faster MAP changes, quicker promo launches, and cleaner execution across both markets. 600+ pricing intelligence runs kept us ahead of the competition.

07

Campaign Execution and Lifestyle Branding

Supported campaign operations from setup through post-campaign reporting: promo-code infrastructure, Shopify banners, landing pages, collection merchandising, redirects, email copy, and final analysis. Every campaign was designed to drive desire and build lifestyle branding, helping customers feel like the products were something they genuinely wanted. Urgency and excitement drove the creative, not discounts.

08

AI-Assisted Content, SEO and Brand Reputation

Built workflows converting search data, customer questions, and social signals into SEO blogs, content briefs, support articles, social scripts, and campaign emails. Produced 50+ dash cam trend reports and published Shopify content across US and Canada. Also built public brand monitoring for Reddit and social mentions with tracked response workflows connecting complaints to operational root causes.

The outcome

Instead of using AI as a writing shortcut, I used it as an operating layer: pulling data, QA'ing attribution, surfacing revenue leaks, drafting executive-ready recommendations, monitoring customer demand, and turning fragmented systems into repeatable business workflows.

1,600+artifacts produced
600+pricing reports
100+feed health runs
$19K+campaign value tracked
Discuss a similar challenge

Project gallery

In the news

July 15, 2026
There's nothing impressive about weaving in and out of traffic.

Joseph Choi on how social media has driven demand for dash cams, featured in The Wall Street Journal's coverage of young drivers, dash cams and TikTok content culture.

Read the full article on WSJ.com