Services · AI for your business

2 min read ·

For owners and managers of small and mid-sized companies, this page explains how to use AI on your real work, such as answering customers, reading documents and summarising reports, without customer data leaving your control. Most teams already use AI, often in tools nobody approved. I help you decide where it actually saves time, build it into the systems you already run, and keep customer data where it belongs. I work with the popular models (ChatGPT and the OpenAI API, Claude, Gemini) and with open models that run on your own servers when data must not leave.

Book a free 20-minute call (opens in a new tab)Message me on WhatsApp (opens in a new tab)
01

  • Answering the same customer questions on your site or WhatsApp, in English and Arabic, and handing over to a person when it is unsure
  • Pulling fields out of invoices, contracts and forms into your system instead of retyping them
  • Summarising long reports, tickets or meeting notes so managers read one page, not forty
  • Searching your own documents and policies in plain language
  • Automating the repetitive steps around that work, such as orders, bookings and weekly reports, in Zapier, Make, n8n or code
02

The chat on this site is built the way I build AI for clients. Try it and judge for yourself.

Every message is screened before the model sees it and every reply is screened before you see it, so it can't be talked into leaking instructions or inventing prices. It runs under a hard monthly spending cap, it answers in the visitor's language, and I read every conversation.

03

If you don't know yet where AI fits, the AI Use Check shows what your team already does with it and the three rules to set now. If you do know, the AI Readiness Audit maps your data and the uses worth paying for. Building a feature or automating a workflow is AI Integration & Automation; owning AI use month to month is AI Governance.

04

The popular hosted ones (OpenAI's GPT models, Anthropic's Claude, Google's Gemini) and open models such as Llama or Qwen running on your own infrastructure. I pick per job, based on quality, cost and where your data is allowed to go.

Not if it is set up properly. I use paid API tiers that don't train on your data, redact personal details before anything leaves your servers, and use your own infrastructure when the rules require it. For the legal side in Saudi Arabia, the UAE and Egypt, see the insight "PDPL and data residency: what to ask your software vendor" (youssefmaged.com/insights/pdpl-data-residency-for-software-vendors).

Often more than for a large one: one assistant or one extraction job can save a small team hours a week. The first call is free and I'll tell you honestly if AI isn't the answer.

Updated: