Last week, we published a post about using Claude AI to plan a mobile dog grooming business. The response was clear: founders want to see how this works in practice, with real stakes and a real timeline. So we found a founder who did exactly that. This is the story of how a solo operator in Singapore used Claude to go from a vague idea to a launched D2C skincare brand in 90 days.
The Starting Point: An Idea and a Deadline
Sarah (a composite of several founders we spoke to) had a corporate background in marketing and a side interest in clean beauty. She wanted to launch a direct-to-consumer skincare brand for Singapore's humid climate, targeting women aged 25–40. Her constraints were typical: a full-time job, a S$15,000 personal savings runway, and zero experience in cosmetics manufacturing or e-commerce. She gave herself three months to launch. Her first step was not a business plan template. It was a blank chat window with Claude.
She started with the same prompt style from the original post: "I want to launch a D2C skincare brand for humid climates, based in Singapore. I have S$15,000. Help me plan the first 90 days." Claude's response set the tone. It did not list product ideas. It asked: "What specific skin concerns does humidity exacerbate in your target demographic? Do you have a formulation chemist in mind, or will you use a private label manufacturer? Is your S$15,000 inclusive of inventory purchase?"
This forced specificity from day one. Sarah answered: focus on oil control and barrier protection for combination skin, no chemist contact, open to private label or contract manufacturing, and the budget needed to cover everything from legal to first inventory batch.
The 90-Day Plan: From Spreadsheet to First Sale
Claude generated a phased plan. We are sharing the concrete numbers and decisions that came from it.
Weeks 1–4: Validation and Sourcing
The AI advised against designing a full product line immediately. It suggested a single hero product to validate demand. Using Claude's analysis of search trends and local forums, Sarah settled on a "gel-cream moisturiser with niacinamide and green tea." Claude provided a list of Southeast Asian contract manufacturers, their typical minimum order quantities (MOQs), and lead times. Sarah contacted five. The quotes ranged from S$8–S$12 per unit at a 500-unit MOQ. Claude helped draft the RFQ emails to ensure she asked for certificates of analysis and ingredient sourcing details.
Concurrent to this, Claude outlined the legal requirements: registering a business with ACRA, applying for a cosmetic importer license from the HSA, and drafting website Terms & Conditions. The estimated cost and timeline for this phase was S$1,200 and three weeks.
Weeks 5–8: Brand and Build
With a manufacturer selected (cost: S$5,500 for 500 units), Sarah used Claude to workshop brand positioning. She rejected the AI's first ten name suggestions as generic. By refining the prompt ("suggest names that are short, easy for Singaporeans to pronounce, and evoke freshness without being floral"), she landed on "Apex Skin." Claude generated a brand voice guide, tagline options, and a content calendar for Instagram and TikTok focused on educational content about humidity and skin.
For the tech stack, Claude recommended a simple, cost-effective setup: a Shopify Lite store (S$29/month), a Canva Pro subscription for asset creation (S$16/month), and a basic accounting template. It warned against over-investing in custom website design before proving the concept. Total setup cost: under S$300.
Weeks 9–12: Pre-Launch and Launch
This is where the plan met reality. Claude suggested a pre-launch email list built through a simple landing page offering a 15% discount. Sarah set this up using MailerLite. The goal was 100 sign-ups. She hit 87 in two weeks by sharing the page in two targeted Facebook groups for skincare enthusiasts in Singapore, using content ideas Claude drafted.
The launch plan was a soft launch: no ads, just an email to the list and three Instagram posts. Claude drafted the email sequence and post captions. On launch day, Sarah sold 31 units at S$38 each, generating S$1,178 in revenue. Her net profit after subtracting the unit cost was negative, as expected for a first batch, but she had validated that people would buy.
What the AI Got Right (and Where Sarah Overruled It)
The plan's structure was its greatest strength. It broke an overwhelming project into weekly tasks with clear deliverables. The regulatory and cost guidance for Singapore was accurate, saving Sarah dozens of hours of research.
The financial model Claude built was sobering but crucial. It projected that at her price point and margins, she would need to sell approximately 1,200 units to break even on her initial S$15,000 investment. This grounded all decisions in unit economics from the start.
But Sarah deviated from the AI's advice in two key areas. First, Claude heavily recommended investing S$1,500–S$2,000 in Meta ads at launch. Sarah disagreed, citing the low intent of cold traffic for a new brand. She chose to reinvest her first revenue into micro-influencer collaborations (sending free product to 5 nano-influencers in exchange for reviews), which she negotiated herself. This drove more cost-effective early growth.
Second, Claude's marketing content, while grammatically perfect, sounded slightly robotic. Sarah spent time rewriting every caption and email to sound like her—a knowledgeable friend, not a faceless brand. The AI provided the first draft; she provided the authenticity.
The Outcome After 90 Days
At the end of her timeline, Sarah had a live business. Key metrics:
- Total spent: S$14,220 of her S$15,000 budget.
- Units sold: 142 of 500 initial inventory.
- Revenue: S$5,396.
- Email list: 311 subscribers.
- Primary sales channel: Direct via Instagram (65%), followed by website (35%).
She was not profitable yet, but she had a real product in the market, real customers, and a clear roadmap to her next batch. The AI did not build the business for her. It acted as a relentless, instant research assistant and planning partner, compressing months of foundational work into weeks.
If You Want to Try This
The lesson is not that AI will write your business plan. The lesson is that the right AI, used as a thinking partner, can dramatically accelerate the validation phase. It forces you to answer hard questions early. It gives you a structured skeleton to build on. Your job is to supply the real-world judgment, the human touch in marketing, and the courage to press send. If you are staring at a blank page for your own idea, the fastest way forward might be to start with a simple prompt and see where the conversation goes. For founders managing every part of their pipeline, from product to sales, tools that save time on planning create space for execution.
— Mira