What I actually work on.
Nine areas, arranged the way they actually relate: AI product work at the centre, and the delivery expertise that makes it shippable underneath it.
The shape of it
The main category is AI product building: products, agents and automation, for businesses in Kuwait and the UAE. Everything else on this page is either a specialism inside that, or a delivery capability that makes the AI work possible.
That ordering is deliberate. Tools change — the eCommerce platform, the site builder, the model provider. What doesn't change is the ability to take a business problem, decide what should exist, build it and put it in front of users. The tools underneath are how, not what.
AI Product Development
Taking an AI idea from a business problem through scoping, prototype, integration and production. MVPs, internal products, SaaS, workflow tools and AI features added to systems already in production.
Evidence: AhlanAI, an Arabic-first AI tools directory covering 500+ tools across 20+ categories, built and shipped rather than specced. Fifteen years of product delivery across Kuwait and the UAE behind it.
AI Agents
Agents that own a bounded workflow — support triage, lead qualification, research, operations, internal knowledge — with defined permissions, tool access, monitoring and a human approval path.
The part that takes the experience isn't the agent loop; it's deciding what an agent may do unsupervised, and building the escalation and observability that make that decision safe to keep.
AI Automation
Removing repetitive work across CRM, email, WhatsApp, support, content operations, reporting, eCommerce and internal admin. Usually the first phase of an AI programme, because it pays for the more interesting work.
Evidence: this site runs its own editorial pipeline — a multi-stage system that briefs, drafts, fact-checks, scores against a rubric and gates publication on quality thresholds, with human review on anything that fails. It's a working example of automation that keeps a person in the loop where judgement is required.
Generative AI
LLM features inside real products: drafting, classification, extraction, summarisation, structured output, retrieval over private data, and the evaluation work that tells you whether any of it is actually accurate.
Practical focus rather than research: model selection against constraints, prompt architecture that survives handover, cost per operation, and fallbacks for when a provider changes behaviour.
Digital Product Strategy
Deciding what should exist. Scoping, sequencing, and the harder half — deciding what not to build and being able to defend it.
Evidence: CTO, COO and CEO seats across Wavai, Al-Qabas, Menasa, Bleep and SHASHA. Strategy here comes from having lived with the consequences of the decisions, not from a framework.
Arabic-first Product Design
Bilingual products where Arabic is the primary design target: RTL as a structural property, Arabic typography set independently, dialect-tolerant input, normalised Arabic search, and prompting authored in Arabic.
Evidence: AhlanAI, and this site's full Arabic route tree with its own typography stack and an Arabic version of the article archive.
eCommerce
Storefront builds, product data, bilingual catalogues, checkout, and the post-purchase operations that decide whether a store is actually profitable.
Evidence: Wavai, a Kuwait-based web and eCommerce practice building Shopify stores, WordPress sites and Webflow experiences for brands across the GCC.
Notion & Business Systems
Workspace design, documentation systems, and the connective tissue between where a team plans and where the work actually happens — increasingly the substrate an internal AI assistant reads from.
Evidence: Notion Certified Expert, with a dedicated consulting practice covering custom workspaces, templates and automations.
AI-assisted Development
Building software with AI tooling — Lovable, Claude and modern AI-assisted workflows — which is what makes a real prototype in week one a reasonable commitment rather than a sales claim.
The time saved goes into the parts that don't compress: understanding the workflow, integration, and handling the cases where the model is confidently wrong.
Supporting product and delivery expertise
The tools underneath. Each has a dedicated page with the detail.
Lovable — AI-assisted product development
Notion — Notion Certified Expert
Webflow — Marketing sites and CMS
Shopify — eCommerce builds
WordPress — Publishing and content platforms
AI Solutions — Automation and intelligent workflows
The full stack — What I actually use to ship
Services
AI Product Development
AI Agents
AI Automation
AI Consulting
Arabic-first AI Products
Recent writing
Need one of these applied to a real problem?
Need one of these applied to a real problem?
Bring the problem rather than the category. The right expertise usually becomes obvious in the first conversation.