Mesita

Context & Problem Space

While traveling and exploring concierge services, I noticed a recurring frustration: generalist concierges (Amex, Chase, hotel desks) often disappointed when it came to the one task that matters most abroad: booking great restaurants. The suggestions felt like “Google-able lists,” and even harder, the actual reservations were left to me. From a product perspective, this was a clear opportunity: a highly specific, high-frequency pain point (finding and securing the right table) underserved by broad concierge services.

Discovery & Hypothesis

As in any product discovery, I started lean. I framed my hypothesis: If travellers value curated restaurant reservations and are willing to pay for someone else to handle the friction, then a specialised concierge service could succeed

To test this, I ran small validation experiments:

• Posted on Reddit to gauge whether people would pay for this service.

• Received positive signals, travelers highlighted pain with language barriers, limited online booking systems, and wasted time on research.

• Converted interest into an actual first paying client before building any product.

This sequence: discovery, assumption testing, and early monetisation, gave me confidence to proceed.

Constraints & Prioritization

As a solo builder, I prioritized like a PM

• Validate before building → don’t write code until someone pays.

• Narrow scope → focus only on restaurants, not “everything concierge.”

• Build for speed → use no-code, AI, and automation to reach live MVP quickly.

These trade-offs mirror real-world product management: reducing scope to validate the riskiest assumptions fast.

From Validation to Build

After securing my first client, I shifted from discovery to delivery:

• Started with Lovable.dev to bootstrap a prototype, but quickly hit limitations.

• Switched to VS Code with Claude Code, leveraging AI pair programming for faster iteration.

• Used Vercel for hosting, Supabase for database needs, and Stripe for payments.

• Automated concierge workflows with Make + Airtable (request intake, tracking, notifications).

This stack balanced speed (no-code/AI), robustness (Supabase + Stripe), and automation (Make).

Iterations and Learnings

The first version shipped quickly, but iteration came from real clients:

• Some users valued convenience (time saved, handled calls in Spanish)

• Other users valued the surprise factors (curated lists).

• Limiting the scope to restaurants clarified the value proposition: depth over breadth.

Mesita.club delivered peace of mind as the core JTBD.

Outcome and Impact

From a product lens, Mesita.club achieved:

• Validation before code → early paying client confirmed demand.

• Time-to-market → MVP shipped rapidly using AI + no-code.

• Clear positioning → unlike generalist concierges, Mesita delivers one job extremely well.

• Scalability → Tech architecture (Supabase + Make automations) supports growth to more cities.

Reflections and Product Learnings

Mesita.club reinforced key product management principles:

• Start with problem discovery, not solutioning — the idea came from repeated signals in real communities.

• Validate with real willingness to pay before investing in a full build.

• Use AI as leverage — Claude Code accelerated engineering and freed me to focus on product decisions.

• Scope ruthlessly — by cutting down to just restaurants, I created clarity of purpose and differentiated value.

• Iterate on real feedback — client use shaped positioning and helped refine workflows.

Closing

Mesita.club remains a live experiment, but it’s also a proof point: with a strong product mindset, AI leverage, and lean processes, a solo PM can take an idea from assumption to paying client to functional product in weeks.