Context
Top Bliss, the parent company of Easebrew, wanted a slimming-product brand and had nothing to launch it with: no name, no category entry, no campaigns, on a timeline measured in weeks rather than quarters.
My partner and I ran it. Two people, the whole build: the name, the product direction, the research, the website and funnel, the questionnaires and forms, the launch plans, the ad creative, and the media buying. Top Bliss's project manager and CEO reviewed and approved as we went, in weekly lockstep. That rhythm, multiple touchpoints a week, is what kept a fast launch from becoming a sloppy one.
Research & development
The launch didn't start with ads. It started with weeks of R&D:
- Naming the brand. TrueSip came out of the positioning work, not from a brief handed to us. We invented it.
- Finding the target market. Category and competitor analysis for a slimming product entering a space with strong incumbents. Who actually buys, at what price, moved by which promise.
- Niching down. The wide category was unwinnable on a startup budget. The work was narrowing until we found a segment the incumbents were underserving.
- Filling a market gap. Positioning TrueSip where the incumbents' messaging wasn't, instead of shouting against them where it was.
- Validating before scaling. Testing the positioning against live response, not assumptions: small spends, real signals, then commit. Research that survives contact with the market is the only kind that counts.
The launch system
Research only pays off if there's somewhere to send the traffic. We built that too:
- Website and funnel. The full path from ad click to conversion, built and wired by the two of us instead of farmed out.
- Questionnaires and forms. How prospects told us who they were, feeding both qualification and the next round of positioning.
- Launch plans. Sequencing, budgets, and what happens in which order, written down and taken to the project manager and CEO for sign-off.
The creative pipeline
Graphic ads and video ads, produced with AI workflows I built and learned in the doing. The pipeline itself became a deliverable:
- Graphics with Canva Magic Studio. Product shots, promo banners, and ad imagery at iteration speed, in one cohesive visual system.
- Video with CapCut AI. Short-form video ad variants at the pace paid social actually demands: multiple variants per week, not per month.
- Copy with ChatGPT and Claude. Hooks, captions, and A/B variants, generated in volume so paid creative could be iterated against live performance instead of hunch.
Media buying
The last leg: learning and running proper media buying, full-cycle.
- Setup. Campaign structure, audiences, and ad formats built out properly instead of boosted-post shortcuts.
- Optimization. Ongoing creative A/B against performance data, leaning on Meta's own campaign optimization, because the platform's ML knows things outside tools can't.
- Scaling. Winners got budget; losers got killed; spend was reallocated across campaigns and ad sets as the numbers moved.
Outcome
- The brand went live on Facebook, with our creative library and campaigns feeding the launch.
- A registered trademark. The name we invented is registered with the Intellectual Property Office of the Philippines, certificate issued, with Top Bliss holding the filing. The brand outlived the campaigns as an owned asset.
- A validated niche. Target market found, position narrowed, market gap confirmed against real demand.
- A reusable creative pipeline. The AI ad workflows, graphic and video, remained in use after launch: production capacity, not one-off assets.
- Media buying as a capability. Setup through scaling, learned by running it live with real budget on the line.
What I'd repeat
The pattern that made this work was picking AI tools by job, not by hype. Canva Magic Studio for iteration because the speed-to-finished-asset beat every standalone generative tool I tested. Meta's own optimization for ad placement because it has data nobody outside the platform has. The mistake most "AI-first launches" make is using one general-purpose tool for everything. The right pattern is a small stack of specialists, coordinated by an operator who knows what each one is actually good at.