Repositioning the offer and rebuilding the website with AI
We repositioned NaveLAB’s offer toward product and AI, and I rebuilt its website, migrating from Squarespace to a custom site built with Lovable, Supabase and Vercel. The site’s annual cost dropped 78%, and the site itself showcases the service we sell.
Visit navelab.pe
Results
−78%
Annual website cost, from $276 to about $60: the website’s share of Lovable, with no Squarespace
~15 min
To publish a change, versus at least an hour before
May 2026
Website live with the new offer
Own code
In our own repository, with no dependency on a closed platform
Desktop
- Performance
- Accessibility
- Best Practices
- SEO
Mobile
- Performance
- Accessibility
- Best Practices
- SEO
PageSpeed Insights scores, November 2026. The site hasn’t been promoted yet, so there’s no lead data so far.
Context and problem
Context
NaveLAB is the consultancy I co-founded with my partner in 2017. Its website is meant to generate leads and sales for our innovation and digital product development services.
The site had two problems:
- It no longer reflected what we do. It offered innovation and agility consulting, visual thinking workshops and graphic recording. Since 2025, NaveLAB has focused on designing and building products with AI.
- It relied on a closed platform. Squarespace Business cost $276 a year, and publishing a change took me at least an hour.
My role and team
Role
I led the migration and built the site end to end: scope, content, design, AI-assisted development and deployment.
My partner Christian and I defined the content and the new positioning together.
Discovery and insights
Discovery
I didn’t interview clients. I drew on what we’ve learned about our clients since 2017 and on a review of the previous site. Three points shaped the decisions:
- The published offer didn’t match our strategy. The site listed NaveLAB’s early-stage services, not our current focus on product and AI.
- Graphic recording requests came through referrals, not through the website.
- The cost didn’t pay off. We were paying an annual subscription for the website alone.
Key decisions
Decisions
Repositioning the offer
We went from three services to two sprints, each for one segment:
- MVP Visual Sprint, for startups that need to ship a first working version fast.
- AI Design Sprint, 3-day workshops for corporate teams.
Removing graphic recording from the site
It still brought in requests, but they came through referrals and it didn’t fit the new focus. I prioritized consistency with the strategy.
Letting the site prove the service
I built it with AI, the same way we build our clients’ products.
Minimal scope
A single page with the essentials: value proposition, services, team and call to action. Supabase only handles the contact form, with no more backend than needed. Visual thinking workshops, which we still run as occasional open events, have their own events page, outside the core offer.
Shared cost
I use Lovable across five projects, so its subscription is split instead of paying for a platform dedicated to the website.
Before · three services
- Innovation & agility consulting
- Visual thinking workshops
- Graphic recording
After · two sprints
- MVP Visual Sprint For startups that need to ship a first working version fast.
- AI Design Sprint 3-day workshops for corporate teams.
The solution
Solution
A single-page website for the core offer, live at navelab.pe. The site is in Spanish.
- Hero with the value proposition: “We build the right product, at the speed of AI.” (Construimos el producto correcto, a la velocidad de la IA.)
- Services: MVP Visual Sprint and AI Design Sprint.
- Team: the two partners and our mix of UX, agility and data.
- Contact form connected to Supabase.
Stack: Lovable (React and Tailwind) for the interface, Supabase for form data, and the code on GitHub, deployed on Vercel.
Learnings
Learnings
Validate the new offer before launching it
We repositioned the site without demand signals for the new services. Today I’d talk to a few prospects, or run a light test, before changing the message.
Launching the site isn’t launching the offer
Without a plan to drive traffic, a website doesn’t generate leads on its own. Today I’d plan acquisition alongside the build.
How I’d measure it today
- North Star: qualified leads per month from the contact form.
- Leading indicators: site visits and share of visitors who click “Tell us your challenge” (Cuéntanos tu desafío) or submit the form.
- Lagging indicators: proposals sent and projects closed for the two sprints.
- Counter metric: annual website cost, so growth doesn’t drive it up.
Next project
KOOKOO