Services

Three practices, one accountable team.

We take on a small number of engagements at a time and stay with them from the first workshop to years of production. Below is what each practice delivers, what it is typically built with, and when it is the right fit.

01

Software & Web Development

Custom platforms, built to be run for years.

Most of our client work starts here: a business process that has outgrown spreadsheets, a legacy system nobody dares to touch, or a product idea that needs a real engineering team behind it. We design the data model first, then the interface, then ship in weekly increments you can use.

What we deliver

  • Web platforms and portals

    Back-offices, customer and supplier portals, dashboards and internal tools built on modern React and Node stacks.

  • ERP and CRM implementation

    RedERP rollouts and customisation, data migration from legacy systems, training and go-live support.

  • E-commerce

    Storefronts connected to live stock, orders and payments — Stripe, ClicToPay and regional gateways.

  • Mobile and point of sale

    React Native apps for field teams and counters, with offline tolerance and real-time sync.

  • Integrations and APIs

    REST APIs, webhooks, payment and logistics integrations, single sign-on, document generation.

  • Run and support

    Cloud hosting on Azure, monitoring, backups, security updates and a named engineer for support.

Typical stack

  • React
  • Next.js
  • Node.js
  • Express
  • PostgreSQL
  • React Native
  • Tailwind CSS
  • Azure
  • Vercel
  • Stripe

A good fit when

You need a system your business will depend on for years, want a single team accountable for design, build and operations, and value clear scope and weekly visible progress over big-bang releases.

02

Agentic SaaS Applications

AI agents that do the work, inside the system.

Agentic software is where our own product, RedERP, has taken us. We build SaaS applications where AI agents carry out real operations — with the same permissions, validations and audit trail as a human user — and we know what it takes to run them reliably for many tenants at once.

What we deliver

  • Multi-tenant SaaS foundations

    Tenant isolation, subscriptions and billing, onboarding flows, roles and permissions, usage metering.

  • Agent workflows

    Specialised agents with tool access to your domain — create, read, update — behind explicit guardrails.

  • Natural-language operations

    Chat and voice entry points in English, Arabic and French that resolve to real actions, not summaries.

  • Evaluation and safety

    Scenario test suites, human-in-the-loop approvals, rate limits and full audit logs of every agent action.

  • Product analytics

    Instrumentation of agent success rates, cost per action and adoption, so the product improves on evidence.

  • Go-to-market surfaces

    Public sites, documentation hubs, in-app onboarding and pricing pages that convert.

Typical stack

  • OpenAI and open-weight models
  • Tool calling
  • Node.js
  • PostgreSQL
  • Redis
  • Next.js
  • Stripe Billing
  • Azure

A good fit when

You are building a software product rather than an internal tool, you want AI to perform tasks inside it rather than sit beside it, and you need multi-tenant reliability, billing and compliance from day one.

03

Innovation & Research

From open question to working prototype.

Some problems are not ready for a product roadmap. For those we run applied research engagements: a clear question, a short timeline, a prototype that can be measured, and an honest report on what works. Recent topics include applied computer vision, accessibility, retail automation and industrial tooling.

What we deliver

  • Feasibility studies

    Technical due diligence and architecture reviews with a written recommendation and cost envelope.

  • Prototypes and pilots

    Two- to eight-week builds that put a working artefact in front of real users.

  • Applied AI and machine learning

    Computer vision, language models and classic ML applied to a concrete dataset and a concrete decision.

  • Accessibility and inclusive technology

    Interfaces and models designed with and for users with disabilities, from screen-reader-first layouts to assistive computer vision.

  • Data pipelines and analytics

    Collection, cleaning and modelling of operational data to answer the question you actually have.

  • Partnerships

    Joint projects with universities, labs and public institutions, including grant-funded programmes.

Typical stack

  • Python
  • PyTorch
  • TensorFlow
  • OpenCV
  • LLM APIs
  • Jupyter
  • PostgreSQL
  • FastAPI

A good fit when

You have a promising idea with real uncertainty, need evidence before committing a budget, or want a partner who can move from a paper to a running prototype without a hand-off.

Engagement models

Three ways to work together.

Every engagement starts with a short discovery phase and a written proposal. From there, pick the model that matches the shape of the work.

  1. 01

    Fixed scope

    A defined deliverable, a fixed price and a date. Best for websites, prototypes, integrations and well-understood implementations.

  2. 02

    Dedicated team

    Designers and engineers embedded with your product, billed monthly, steering with you sprint by sprint. Best for products and platforms that keep evolving.

  3. 03

    Run and improve

    Hosting, monitoring, support and a monthly improvement budget for systems already in production — ours or someone else's.

Work with us

Have a project in mind?

Tell us what you are trying to build or fix. We reply within one business day with a clear next step — usually a 30-minute call.