Skip to content
BXGLNDR

Digital solutions

Expertise

The expertise behind my websites, apps and automation: their uses, technical choices and the work that illustrates them.

The pillars

  1. 01

    Acquisition

    Being found: a clear local presence, fast pages, readable by search engines and by AI assistants alike.

    • Local presence and business listings

      Rebuilding the Google listing: description, service areas, menu, posts, replies to reviews.

    • SEO and AI-assistant readability

      Structured data, sitemap, linked language versions, llms.txt file.

    • Multilingual sites (FR, EN, AR with right-to-left)

      Translations whose parity is checked automatically, including right-to-left Arabic.

    Related work

  2. 02

    Sales

    Taking the order with no friction: an installable app, basket, payment, and the kitchen screen that receives it all.

    • Installable ordering app (PWA)

      Menu, product configurator, persistent basket, takeaway, dine-in or delivery orders.

    • Online payment and pay-on-delivery

      Card payment (Stripe), pay-on-delivery, prices recomputed on the server.

    • Kitchen screen and printed tickets

      A queue designed for a 5-inch terminal in the kitchen, with automatic ticket printing.

    Related work

  3. 03

    Loyalty

    Bringing customers back: loyalty cards, messaging journeys, rewards that are verified rather than claimed.

    • Loyalty cards and programmes

      Tiers, rewards chosen by the customer, unambiguous rules (one reward, one use).

    • WhatsApp journeys (official API)

      One QR code, one message, a guided conversation; opt-out and GDPR notice built in.

    • Push notifications

      Announcements sent to installed customers, with explicit consent.

    Related work

  4. 04

    Operational intelligence

    Seeing what happens: a back office your team understands, useful alerts, automatic checks that warn you before the customer does.

    • Custom back office

      Orders, menu, offers, copy and settings editable without a developer or a redeploy.

    • Automated tests and checks

      Strict types, tests, palette and translation guards, checks on the page as actually served.

    • Dashboards and alerts

      Today’s figures, late orders flagged, kitchen-screen monitoring.

    Related work

  5. 05

    Innovation

    Using AI where it helps, with guardrails: audits, content generation, assistants, and a controlled cost.

    • Guarded generative AI (Claude)

      Batch generation with caching, a capped cost, content adapted to the declared age.

    • AI-assisted audits and sales preparation

      Analysing a company from public sources, with a call script and email prepared.

    • Working method with AI

      Project constitution, lessons log, anti-slop catalogue, verification by measurement.

    Related work

  6. 06

    Technical foundation

    What carries everything else: typed and tested code, well-organised data, controlled hosting.

    • Next.js, React and TypeScript

      Fast server-rendered interfaces, code typed end to end.

    • PostgreSQL and Supabase

      Controlled schemas, row-level access rules, backups.

    • Hosting and deployment

      Private server (VPS), Cloudflare, verified deployments, handover documentation.

    Related work

Technical details

Each project has its own constraints. These are the technologies actually used in the public projects; their features, checks and limits are detailed in each case study.

What I do not do

Four principles, so things are said before we start.

  • 01

    Promise a figure I cannot measure.

    Google rating, footfall, revenue: I commit to what I deliver, not to what the market will do.

  • 02

    Invent reviews, testimonials or figures.

    My pieces show what is verified. The rest is marked "to complete", or not written.

  • 03

    Deliver without saying what is left to do.

    Every project leaves with the list of what is done, what is not, and what depends on you.

  • 04

    Set advertising trackers by default.

    No tracking cookie by default: this site uses none, and it is my starting position for yours.

Method: AI and rigour

AI writes fast. What decides quality is the checks around it.

  1. Scope

    Who uses the tool, where, under what conditions? A kitchen screen is not designed like a home page.

  2. Build

    Typed code, automated tests, palette and translations guarded by checks that fail when you drift from the frame.

  3. Measure

    The render as actually served is measured on desktop and phone: contrast, overflow, text sizes.

  4. Hand over

    Handover documentation, access, a list of what is left. Nothing is hidden behind an "it works".

7-month register

Each square is a day on which code was recorded in my working repositories, confidential projects included, without revealing their content. Part of that code is written with AI assistants: that is deliberate, and it is the point of the method.

138 active days out of 215, 2186 code records (commits).Dates and totals only. Confidential projects are folded into an anonymous total.
See the month-by-month detail
Month by month
MonthActive daysRecords
March 202621 / 31161
April 202616 / 30121
May 202611 / 31127
June 202625 / 30422
July 202627 / 31338
August 202626 / 31654
September 202611 / 30315
October 20261 / 148
Breakdown by project
ProjectActive daysRecords
HoneyCold70660
Le K’Pital33460
Ciao Bella18357
AURA22282
Let’s Go Pizza24184
AM Market7134
Vantora816
Other projects (confidential)1293

A project to build?

Describe your situation in a few lines. I will tell you honestly whether I can help, and where to start.

Write a message