JJ MARINE

JJ Marine sells and charters yachts in Dubai, and before the latest rebuild it did so off a Wix version of its site. The catalogue lived as PDF brochures, which meant the specifications buyers search for existed nowhere a crawler or an answer engine could read them, and the inventory could only be updated by hand, page by page. A dealership whose stock turns over cannot run on hand-edited pages. It also sells into a market where the buyer is as likely to read Arabic or Russian as English, and the old site was English only.

Rebuild the site as a database-backed catalogue where the boat row is the source of truth, then treat search visibility as an output of the data model rather than a marketing layer added afterwards. Every boat renders its specifications as on-page HTML and emits Product plus Offer structured data from the same row. The migration off Wix was written down before launch as 15 redirect rules and 4 hard 410s, plus 7 redirects for the old brochure PDF deep-links. The build itself ran as agent work under a coordination protocol, with adversarial review agents as a scheduled phase.

JJ Marine is a yacht and boat dealership in Dubai that sells new and pre-owned boats and runs a charter fleet. Bruno Jaamaa has worked there since 2024 and holds the title of Technical Lead. The role has had two phases: first the website and its search visibility, and now AI systems for the business itself.

The company website has been live for years, and Jaamaa has redesigned it several times since 2024. Before the latest rebuild it ran on Wix, in English only, and its catalogue was delivered as PDF brochures. The specifications a buyer searches for were not readable by a search engine or an answer engine, and stock had to be updated by hand, page by page.

The first phase was the website and SEO. The current version at jjmarine.ae replaced the Wix one. It holds 48 boats (27 power, 12 sail, 5 pre-owned, 4 charter) in a Supabase table, where every row generates its own page, its specifications as on-page text and its structured data. The site is published in four languages, English, Arabic (laid out right to left), Russian and French, and its sitemap lists 384 URLs. Sixteen editorial guides sit beside the catalogue. The move off Wix was written down before launch as 15 redirect rules and 4 pages marked as permanently gone.

The current phase is AI systems for the business. By his account he built a Telegram bot with retrieval over the business's systems that drafts client replies, financials and boat decks; nothing goes out until he or the boss approves it. The stated aim is to leave more of the day for client relationships and face-to-face work. No further detail of the internal systems is published here. The public site is a separate matter: it has no chatbot and nothing customer-facing that talks to a model.

The role is carried out in person as well as in code. By his own account, Jaamaa does "everything from the website to client relations, fueling, taking clients out on the boat, managing SEO and outbound, newsletters, and our Instagram account." He has said that he walked a client through the purchase of a Jeanneau Merry Fisher 1295 Fly.

The rebuild itself was planned as 7 epics and about 50 short sprints. By launch it had run to 111 sprint commits, numbered up to S165, across 22 epics, and it has kept going since. Epics 7 through 21 were all emergent, driven by feedback while the site was already running: a matcher for buyers who do not know what they want, mobile touch behaviour, four languages, a launch runbook, a self-hosted analytics layer, an admin worklist. The gap between the plan and the log is the shape of building in-house, where the people who own the business see the site every day and keep having opinions.

The site was built by AI coding agents working under a written protocol. Sprint S156 wired the repository into an Aethereum coordination room with 13 agent surfaces, and the repository's AGENTS.md makes the loop mandatory: pull team context before touching cross-service code, share intent, declare a contract for every new interface, claim and release before a refactor, check the blast radius before changing a contract others depend on. Briefs from one agent to another are committed artifacts. Adversarial review agents are a phase in the build: S140 fixed three live search defects found in S138's own code, and S162's three headline bugs came out of adversarial review rather than from a user.

It also collided. The project's plan records that sprint numbering is broken for 5 August 2026, because two agents worked the repository concurrently and S156, S157, S158, S159 and S160 each name two different commits. S162's commit body records two renames being swept into the other agent's commit, because git add -A from either working tree picks up whatever the other has staged. Nobody lost work, but the history is no longer readable by sprint number and has to be disambiguated by hash. That cost is recorded in the repository rather than tidied away.

Next.js 16.2.9 App Router on Turbopack, React 19.2.4, TypeScript strict, Tailwind CSS v4 with @theme tokens, Motion 12.42 under LazyMotion, next-intl 4.13, Supabase, Vercel. 28 public page routes. The accent colour is a single token, so a rebrand is a one-line change rather than a search-and-replace across components.

The catalogue is the spine. Boats live in a Supabase table and every row generates its detail page and its Product plus Offer JSON-LD from the same source, alongside filter, sort, facet and full-text search over the collection. Specifications render as on-page HTML instead of living only inside a brochure PDF, which was the single biggest search and answer-engine gain in the build, and the brochures stay available as downloads. dynamicParams is false on /boats/[slug], so a slug that is not in the table returns a true 404 rather than a rendered empty shell; boats that do exist revalidate hourly through ISR.

Internationalisation runs across en, ar-AE, ru and fr with localePrefix 'as-needed' and full RTL for Arabic. Sprints S143 to S147 alone moved roughly 500 strings, on top of 4 metadata catalogs at 146 keys each, 144 boat translations and 48 translated guides. That is what turned an English-only site into a 384-URL sitemap with an hreflang matrix and x-default across all four languages.

The discovery layer is deliberately wide: JSON-LD for AutoDealer, CollectionPage, Product with Offer, BreadcrumbList, WebSite, BlogPosting, Person, Service and FAQPage, an Atom feed, llms.txt and llms-full.txt, an AI-crawler group in robots.txt naming 12 bots (GPTBot, ClaudeBot, meta-externalagent and others), plus an IndexNow key file and ping script. Lead capture writes to an insert-only Supabase table behind RLS, with Turnstile and a honeypot live and Upstash rate limiting wired but fail-open until its keys are set in production. Analytics is self-hosted: a site_events table and a navigator.sendBeacon client, with no third-party analytics as a dependency (Vercel's own Analytics and Speed Insights tags stay mounted as a free cross-check, but nothing reads from them). A password-gated /admin, unlinked, disallowed in robots.txt and excluded from the sitemap, carries a leads worklist, traffic, boats ranked by enquiries and CSV export. A separate social pipeline renders posts and slides from live catalogue data through satori and resvg and pushes them to Buffer as drafts.