I read every ChatGPT and AI Overview answer in the category.
Commas earned 8.4% of citations. Peers sat between 21% and 33%. The gap became 221 page opportunities and a 35-topic roadmap.
Growth marketer. I find the customers who pay, prove what brought them in, and run the whole machine on AI agents I built.
Open to growth leadership roles at tech companies
I’ve spent ten years running growth where the numbers are real. $20M+ a year in paid media at Shopify. About $1M a month at Commas. A B2B engine that cut the cost of a qualified opportunity by 74%. Now two coding agents run the reporting, research and ad operations, so my time goes to the decisions that move revenue.
Each one comes with the spreadsheet, the test design and the version that didn’t work the first time.
in paid media across Google, Meta, TikTok, YouTube and Reddit, scaled 4x while holding a 3.2x ROAS.
cost per sales-qualified opportunity, in two quarters.
cost per new patient, with 75% more new patients and no added spend.
in paid acquisition, about $5M over five months, optimized to sellers who actually process payments.
of new-business wins came through AI assistants, revenue the CRM had credited to paid search.
tokens through Codex alone. One autonomous task ran for 15 hours and 37 minutes.
Find the customers who pay, prove what brought them in, then put the money behind it.
Claude Code and Codex handle the reporting, research, ad operations, tracking fixes and builds. I make the calls that need judgment and approve anything that spends money.
Hover a system to trace it. Nothing that spends money ships without my go.
The long version covers the knowledge base, the guardrails and how a marketing team can start.
How I run growth on Claude Code and CodexBuyers ask ChatGPT, Perplexity and Google’s AI Overviews before they ever click. I measure who gets cited, then build the pages that earn it.
Commas earned 8.4% of citations. Peers sat between 21% and 33%. The gap became 221 page opportunities and a 35-topic roadmap.
The CRM had credited that revenue to paid search. Asking buyers what they actually used, and tracking referrals from ChatGPT, Perplexity, Gemini and Claude every month, put it where it belonged.
This site runs the same playbook: one clear entity, structured data, an llms.txt file, open access for AI crawlers and answers written to be quoted.
Optimize to paying merchants, approved sellers and booked patients. Platform conversions are where the analysis starts, not where it ends.
If a channel can’t show lift in a geo or audience test, the budget moves somewhere that can.
Attribution, dashboards and agents first, so every week after gets faster than the one before.
When a buyer asks an AI who to trust, the brand should be in the reply, with a page that earns the citation.
How I actually do the work, written for the people who’ll do it next.
Short bio, for introductions
Farhan Munshi is a growth and performance marketer based in Miami. He leads growth marketing at Commas, a payments platform for high-ticket online sellers, and has run growth at Pipe17, Shopify and Dental Team FL. He runs his work on AI agents he built himself.
I started at McMaster with a hoodie store I grew to about $300K in sales on Facebook and Google ads I learned from YouTube. That became an agency, then Purolator, where I built paid media from scratch and scaled it 5x, then three years running paid at global scale at Shopify. Every job since has been a bigger version of the same question: which dollars bring in customers who pay?
“The engine room of our growth motion… it’s moved the numbers. Our lead-to-SQO conversion and cost-per-influenced-SQL both prove it.”