Manuel Zamora · Founder · Downshift portfolio
I run Downshift, a portfolio of 10 production SaaS products. Fair, FairStack, Rivalize, JamWise, RepairSnap, FleetSnap, TIH, TodayIn, Travlist, Wonderbell. Some you have heard of, most you have not. They share infrastructure, share an AI pipeline, share a small team. They have all paid the bills for years.
In 2025 I tried to grow Fair through paid social. I hired an agency, brief in hand, and watched five figures evaporate over three months on ads that all looked like the same generic template with different logos pasted on top. The work was not bad in the abstract. The work was not Fair. Same composition, same captions, same visual register as every other DTC ad in the feed. Performance reflected that.
I went back to the brief and realized the bottleneck was not creative volume. The agency could ship more ads. The bottleneck was brand voice: there was no portable, structured representation of how Fair sounds, what Fair stands for, who Fair is for, that any tool or human could read and use. Every new ad started from a blank page.
So I started building what I wished existed. Pull a brand profile from a URL. Ground every ad, every social post, every UGC reel, every email in that profile. Edit anything in plain English. Run it as a daily queue, not a weekly miracle. Use the same AI pipeline that already powered the portfolio so it was battle-tested before any outside brand touched it. That is mani.
I begin building small SaaS tools as a portfolio. The first three are scrappy, profitable enough to keep building, and not famous enough to attract VCs. Good place to learn.
After ChatGPT lands, every product in the portfolio needs creative AI fast. We build FairStack as the shared engine that powers all 10. The same engine eventually grows into mani.
Trying to grow Fair on paid social, I burn through five figures on agency-produced ads that all look like templates. Every dollar feels wasted because every ad feels generic. The bottleneck is brand voice, not creative volume.
After 12 months of internal use across the portfolio, mani opens to outside brands. Today: 53 marketing pages live, free Brand DNA tool, daily creative queue, $19.99/mo entry tier. Everything I wish I had two years ago.
Six principles that shape every product call we make.
Your brand voice, your audience profile, your visual identity. We never train shared models on your data. Your Brand DNA never leaks into prompts you do not control.
The right metric for marketing software is the time it gives back to founders, not the impressions it claims to optimize. We measure ourselves on the first.
We will tell you when mani is the wrong tool. We will tell you when a competitor fits better. The free tier is genuinely free, and the comparisons are genuinely fair.
Marketing teams need 30 brand-consistent assets per month, not one polished hero ad per quarter. We optimize for the calendar, not the case study.
No prompt engineering. If something is wrong with an ad, you say what is wrong with it in normal language and the model fixes it. You should not have to learn a new vocabulary to use a creative tool.
mani is paid for by Downshift, not by a VC clock. That gives us the freedom to ship the right product over five years instead of the loudest product over five quarters.
Want the portfolio-credibility version of this story?
Read the Built-on-Fair version →Interviews about AI ads, Brand DNA, honest pricing, and building a multi-product SaaS portfolio.
Browse podcast appearances →Drop your URL. Get a free Brand DNA report in 90 seconds. No signup, no email gate.