MAYDAN
Memberships, class booking, gradings and kids' terms across two branches — with five interfaces, two of them entirely inside WhatsApp.
AI agents · Automation · Products · IT delivery
BASES builds AI agents that take real steps inside your systems, automation that removes the manual handoffs between them, and mobile and web products people use every day. Five of our own products are live below. Open one and use it.
Both channels reach the same people — the ones who would do the work.
We work across several markets. An engagement is delivered in your working hours, under your rules, into your environment.
What we build
The list is short on purpose. Each of these is something we have built end to end, not a service line we are willing to attempt.
Systems that read what comes in, decide inside limits you set, act in your software, and hand the case to a person when they should.
AI agents in detailThe manual steps between systems: intake, matching, approvals, reconciliation, the report someone rebuilds every morning.
Automation in detailProducts that hold up in daily use — Arabic first, right-to-left as the default, fast on a mid-range phone.
Mobile and web products in detailThreads where a booking, a payment or a reschedule actually completes, writing to the same records as the app.
WhatsApp systems in detailIntegration with what you already run, deployment into your environment, and operating the system after launch.
IT delivery in detailOur work
We designed and built each of these ourselves, end to end — the domain model, the seeded world, every screen, both languages. They are not screenshots and not videos. Every control does something.
Memberships, class booking, gradings and kids' terms across two branches — with five interfaces, two of them entirely inside WhatsApp.
An own-fleet, cash-on-delivery parcel network across merchant, driver and operations — with cash tracked as an object from the door to the payout.
Five clinic branches and their own laboratory: booking, live queue, a lab worklist from accessioning to validated release, and critical-value callbacks.
A guest layer over an existing point of sale: booking against real table availability, an allergen passport, coursing, split bills, and the floor.
Three branches where the specialist is the bookable unit, privacy is machinery rather than a promise, and at-home service is a mode of the same calendar.
These are our own product builds, shown as our work. We do not name businesses and we do not present them as client engagements.
We would rather put one small thing into real use than deliver a document about six large ones.
The spreadsheet, the phone calls, the paper receipt. What the work is now, before any talk of a model.
Acceptance criteria written and signed before the first line of code. If we cannot state them, the scope is not ready.
One narrow piece, working, inside your systems. Fixed scope, fixed price, a date.
Code, documentation, and the training to run it without us. Your team should be able to change it.
Where an agent stops
Autonomy is not the goal. A bounded system that a supervisor trusts is worth more than a clever one nobody switches on.
Permitted actions are enumerated. Anything outside the list halts and asks rather than improvising.
Input, decision, output, and who or what approved it. Readable months later by someone who was not there.
Money moving, a contract sent, a record deleted — those pass through an approval gate by design.
Per capability, without a deployment, by someone who is not an engineer.
Most WhatsApp work in this market is a notification pipe. It sends a message and then sends the customer somewhere else. That is where the drop-off happens.
Booking, rescheduling, paying, confirming a delivery address — done inside the thread, in the customer's own words.
A booking made in a thread is the same record the app lists and the staff screen prints. One ledger, not a parallel one.
In one of our products, two of the five interfaces are WhatsApp: the member's and the coach's. Rosters, attendance and cover are handled without opening an app.
Notes
Short pieces on the parts of this work that decide whether a system survives contact with a real operation.
On the distance between a demo that impressed the board and something the operations team depends on.
On why most WhatsApp automation quietly becomes a person typing.
On what breaks when Arabic is added at the end of a project.
One message is enough to start. If it is not something we should build, we will say so.
Our WhatsApp line is being connected. Email reaches us today.