AI Agents · Jordan
AI Agents & Agentic AI Systems in Jordan.
We build AI agents that read your documents, decide inside your rules, act in your software, and stop for a person when the stakes require it. Built in Amman, for companies in Jordan.
Your data stays where you choose: on your own servers, in your own cloud account, or with a contracted model provider. We confirm which in writing.
An agent is defined by what it is allowed to do
Many products sold as “AI agents” are a language model behind a chat box. That is a useful thing to own, and it is not an agent. An agent is a system that holds state across steps, takes actions with consequences in your software, and has an explicit boundary around what it may do without asking.
Which means the interesting engineering is not the prompt. It is the part nobody demos: what happens on the third retry, what happens when the system is only partly sure of what it read, what happens when two agents disagree, and how you prove afterwards what the system did and why. We build that part first and treat the model as one component inside it.
Put plainly: firm rules and checks around a model that can be wrong. The rules decide what is allowed; the model reads, drafts and suggests.
What we build
Document agents
Read contracts, invoices, policies and official letters — including Arabic scans — extract structured fields, check them against your records, and flag what they are not sure of instead of guessing.
Conversational agents
Handle real customer conversations on WhatsApp, voice or your website: answer from your actual policies, take the booking or the order, and hand to a human the moment the conversation needs one.
Back-office agents
Sit inside the workflow rather than beside it — reconcile ledgers, chase renewals, route internal requests, prepare the proposal, and update the systems of record directly.
Multi-agent systems
Several specialised agents coordinating under one supervisor, for processes too varied for a single prompt. This is where most of the hard engineering lives.
From our catalogue
Agent systems we've built
Travel and tourism
Travel enquiries and live prices on WhatsApp
AI agents answer a travel agency's customers on WhatsApp and social media day and night, suggest packages and quote hotel and flight prices from live rates. Serious requests reach the sales team with the details already collected.
Hospitality and leisure
WhatsApp guest service for hotels
AI agents answer hotel guests on WhatsApp, take reservations and service requests, send each request to the right department and keep managers informed. Complaints and compensation stay with the duty manager.
Each card says where that work stands today. See the full catalogue, or read how Nour answers Dallas Travel's travellers on WhatsApp.
Why a Jordan-based team changes the build
Arabic is a first-class requirement. Arabic document extraction and Arabic conversation are not an afterthought bolted onto an English system here. With Arabic documents the failures usually come from scan quality and inconsistent formatting, not vocabulary, so we test on real or anonymised samples of your documents before anything goes live. You can try the basic step yourself with our free Arabic OCR tool.
Where it runs is a design input. Some organisations in Jordan, banks among them, have rules on where data may be processed. You choose: your own servers, your own cloud account, or a contracted model provider. See where your AI runs for how each option works.
Same timezone, same week. In our experience, agent projects stall on slow feedback more often than on the model. Working from Amman means the people who own the process can be in the room when we map it.
How an engagement runs
- 1
Scope one process
A call where you describe the manual work. We come back with what an agent can take, what it should not, and where the approval gate goes.
- 2
Build against real data
Your documents and your edge cases, not a synthetic demo set. This is where most of the actual requirements surface.
- 3
Ship gated
The agent goes live with a person approving consequential actions and a log of everything it does. Approval steps are relaxed only when measured results justify it and you agree; decisions about money always stay with a person.
Not sure which process to start with? Run the operations audit or request a scoped quote.
More in Jordan
AI agents in Jordan — questions we get asked
What is the difference between an AI agent and a chatbot?
A chatbot answers. An agent acts. A chatbot returns text and leaves the work to a person; an agent reads the document, queries your database, writes the record, sends the message and stops for approval when it should. The engineering difference is state: an agent has to know what it already did, what failed, and what it is not allowed to do on its own. That is most of the build.
Do you build AI agents for companies based in Jordan?
Yes. We are based at the 7th Circle in Amman and work with companies in Jordan and across the region. Being in the same timezone matters more than it sounds: agent projects need short feedback loops with the people who actually own the process.
Can an AI agent work in Arabic?
Yes. We design agents for Arabic from the start, in conversations and in documents. Arabic document work is harder than English — joined letters, diacritics, right-to-left layout and inconsistent scan quality all cause misreads — so we don't trust a single model call: results are checked against your records, and anything uncertain goes to a person.
How long does it take to deploy an AI agent?
It depends on the process, the systems the agent has to connect to, and how much checking it needs. We agree the timeline in writing with the scope, before work starts, and it includes the approval steps and the log. We would rather put one agent fully to work than pilot five.
What happens when the AI agent gets something wrong?
It is designed to be wrong safely. Every action an agent takes is logged, reversible where the underlying system allows it, and anything with a financial or legal consequence waits for a person to approve it. An agent we can't audit doesn't go live.
Do our data and documents leave our infrastructure?
That is your decision, not ours. Your agents can run on your own servers, in your own cloud account, or with a contracted model provider. If nothing may leave your network, your own servers are the option that fits. Tell us the constraint before design starts — it changes the architecture, not just the paperwork. Your data is never used to train shared models.
Map one process with us.
Every project starts with one piece of work, mapped with your team, before anything is built.
Talk to usYou decide. We answer for it.