Hello Seven is a US company teaching business to women, with a focus on Black and minority entrepreneurs on the way to a million-dollar business.
Over time they launched courses, a coaching certification, a membership club, books, and a podcast, and by 2024 the gap between all of them was showing.
Each product had grown up with its own way of teaching, selling, and looking.
Members got lost between offers, the coach certification had no public face, and years of curriculum structure had never been mapped end to end. On top of that, the company was betting its next chapter on AI with no designed product layer to deliver it, so they needed shared infrastructure across the whole portfolio rather than a simple redesign.
Hello Seven's growth levels, H1 to H7, are defined by revenue, and the old curriculum assumed revenue meant knowledge.
Take a member at level H5 earning $500k a year: she may have built that business solo, so her hiring knowledge is still at H2... and the H5 hiring lessons land two levels above her head.
That flaw came out of the study I ran before the company migrated its courses to a new platform (a migration is the WORST moment to carry old problems forward). I went through client surveys and interviews, competitive analysis, past feedback, CS complaints, and stakeholder interviews across the team. To fix it, I built a taxonomy that separates revenue stage from subject skill, tagging every lecture by topic, subtopic, and the range of levels it serves so coaches can prescribe H2 and H3 hiring lessons to an H5 founder.
Certified coaches had no public proof of their certification.
I designed the listing platform at hellosevencoach.co from scratch, with profiles and discovery for more than 60 certified coaches... built around how a visitor finds the right coach and how a coach turns certification into visibility.
I designed Hello Seven's member AI platform to deliver founder Rachel Rodgers' coaching frameworks on demand.
It brings together an AI coach for one-on-one business conversations, a Business Auditor that drafts and sharpens 90-day plans, and a Growth Scale Assessment built on the same H1-to-H7 scale that organizes the teaching catalog.
The main design problem was trust, because AI coaching had to feel like Hello Seven rather than a generic chatbot, from conversation design down to how the product admits its own limits. The human coaching program it plugs into holds an NPS of 92 across 304 surveys, and two days after launch a member described the AI coach in words we didn't script:
"This AI is not just a generative AI. Claude will talk my ears off and next thing you know I've written an entire dissertation... not a money making move. But [the AI coach] answers my question directly. My absolute favorite thing is the action oriented answers: 'Ok, stop. That is a Broke Ass Decision. Change this and come back. When will your VA have this done?' I have picked a launch date, written my launch email, have follow up messages for my VA to send, and have clarity on my funnel."
Hello Seven member, two days after launch
Behind the scenes, I built the agents that run the company's content flow end to end, from topic selection to a ready-to-publish social post in under 20 minutes.
The setup includes a copy reviewer that holds brand voice and a competitive analyst that benchmarks Hello Seven's socials against any competitor and returns a metrics-based report. Each agent has a defined scope and a human handoff point, and the team uses them daily.
Each subbrand and product line got its own documented design system.
With tokens, typography rules, and handoff bundles in place, non-designers can ship on-brand pages, and the sales and marketing pages built on top of those systems are where the $10M+ comes from.
Three product calls shaped the rest of this work.
Tagging 300+ lessons instead of re-recording them. I tagged every lesson by topic and by the range of levels it serves, which cost months of classification work and left a maintenance duty that never ends. Still, the other two options on the table were worse: leave lessons gated by revenue where members missed basics, or re-record 300+ lessons per level.
Keeping the AI coach narrow. A broad chatbot looks more capable in a demo, of course. I chose a narrow coach that pushes actions and admits its limits because members needed to trust it, and they kept using it after launch week.
Design systems built for non-designers. Once marketing had the tokens, typography rules, and handoff bundles, they could publish on-brand without me, which cost me pixel-level control of every page and removed design as a bottleneck on the pages behind the $10M+.
Pages I designed for Hello Seven have processed $10M+ in tracked conversions.
Hello Seven sells exclusively through its sales and marketing pages, and I've designed every one of them since the partnership began in 2023. The figure is checkout revenue processed through those pages, measured in ThriveCart, net of refunds. (The offers, the marketing, and the selling behind them are the team's work. The pages are mine.)
Two page types have their own before-and-after numbers in PostHog from 2025 to 2026: across lead magnet flows, the median conversion rate went from 6% to more than 35%, and program offer landing pages went from 6% to as much as 17%.
PS: Yeah, I also do all the event/graphic work for the brand. But this is something for a coffee and snacks conversation.





















