Designing & building AI products — straight answers
Cost, timeline, and process for getting an AI product designed and shipped. No sales fluff — the same answers I'd give you on a call.
How much does it cost to design and build an AI MVP?
Three fixed-price packages. An AI Trust Audit is $1,200 for 5 working days: 10+ ranked fixes and your key flow redesigned in code. Live in 4 Weeks is $7,000–9,000 fixed: a live AI product, design through deploy, with code you own and 30 days of support. Ship Every Week is $3,000/month with at least one shipped improvement a week and a 3-month minimum. The exact scope and fixed price follow within 48 hours of a free 30-minute call, and the online estimator gives a range in 30 seconds. It's faster than an agency and costs less than hiring a separate designer and developer, because one senior person handles both design and production code.
Do I need both a designer and a developer for my AI startup?
Not necessarily. The biggest source of lost time in early AI products is the handoff between design and engineering — specs lost in translation, designs rebuilt from scratch, cycles spent waiting on a second person. A studio where one person both designs and ships removes that handoff entirely. For an MVP or first version, one designer-engineer with AI velocity usually beats hiring two separate roles.
How long does it take to ship an AI product or LLM MVP?
A focused AI MVP takes 4 weeks from scope to live product when one person owns design and code. The Proof-First Build runs Map (days 1–3), Prove (week 1), Build (weeks 2–3), Verify (week 4) and Hand over (day 28), and you get a live build every Friday. Concrete example: EngiBoard — a native desktop SaaS on Tauri 2 + Next.js + Supabase across 4 platforms — went from a blank repo to v0.1.44 in production with solo work and AI tools. Read the full build diary.
What is a solo + AI studio, and how is it different from an agency?
One experienced product person using AI tools (Claude Code, Cursor) as co-workers to deliver what used to require a full team. Versus an agency: no account managers, no design-to-dev handoff, no junior staffing, one point of accountability. Agency-grade output, delivered faster and cheaper, with the founder talking directly to the person doing the work.
Can one person really design and build a production AI product?
Yes. I both design and ship: full-stack SaaS, Tauri desktop apps, Next.js web. Live proof: Vigilo (a global risk monitor with 45 live data sources across 239 countries) and EngiBoard (a 4-platform desktop SaaS built solo with AI tools). AI handles the mechanical volume of coding; 9+ years of product design judgment handles what matters.
Who designs LLM, chat, copilot, and AI-agent interfaces?
That's the core specialisation — UX for AI-driven products: LLM chat, copilots, RAG systems, AI search, agent control rooms. Including the patterns most teams get wrong: streaming responses, showing model uncertainty, source attribution for RAG answers, confirmation before an agent acts, and billing for AI features. I designed AI products at AI Dent (used by 1000+ clinics) and build my own AI tools.
What kind of companies do you work with?
AI-first startups, SaaS companies, and teams in HealthTech and FinTech that need both product design and frontend shipped fast. Typical engagements: an AI Trust Audit of an existing product, a first version live in 4 weeks, or Ship Every Week — an ongoing monthly partnership that acts as an embedded studio without a full-time hire.
Have a different question, or an AI product you want shipped fast?
Book a free 30-min call