Stop hand-picking couriers — let your agent assign Bosta, Mylerz, ShipBlu, or Aramex per order

The decision nobody makes well at 80 orders a day

You run more than one courier on purpose. One is reliable in Cairo and Giza, another reaches Upper Egypt without the failed-attempt loop, a third is cheaper on light parcels, a fourth you keep for the governorates the others won't service well. The problem is that this knowledge lives in one person's head, and that person assigns couriers by eyeballing each order in a dashboard. When they're out, assignment gets random. When volume spikes, everything goes to whoever's fastest to click, not whoever delivers that zone.

The cost shows up later as failed deliveries in zones the wrong courier was never good at, and as overpaying a premium courier for a 0.5 kg order across town.

What the agent does

Connect your couriers and your store to the Wassel MCP endpoint and hand the agent your assignment rules as a standing instruction:

On every paid order, choose the courier by these rules: Cairo and Giza under 2 kg go to courier A; Upper Egypt and remote governorates go to courier B; anything over 5 kg goes to courier C; if the order is COD, prefer the courier with the best collection record for that zone. Create the shipment with the chosen courier, write the courier name and tracking number back to the order, and message the customer.

What runs:

• The store signals the paid order. The agent reads destination governorate, weight, and whether it's COD.
• The agent applies your rules in order and picks one courier. The logic is your operational knowledge written down once, not a guess.
• The chosen courier gets the shipment with the COD amount from the order total.
• The store gets the courier name and tracking number written back, so support and finance see which courier carried which order without asking anyone.
• WhatsApp Cloud sends the customer the tracking link.

When an order matches no rule cleanly, the agent flags it for a human instead of guessing. You review the few edge cases, not the whole queue.

Why it matters

1. Zone fit, every time. Orders go to the courier that actually delivers that zone well. Failed-attempt rates in the regions you were assigning by guess drop, and every recovered first attempt is a fee and a return you avoid.
2. You stop overpaying on light parcels. A 0.5 kg order across Cairo doesn't go to your premium nationwide courier just because someone clicked fast. Over a month at volume the courier-cost line moves visibly.
3. The knowledge stops being a single point of failure. The rules are written down and applied identically whether your ops lead is in or on leave. New courier or renegotiated zone — you edit the rule, not retrain a person.
4. COD-aware routing. COD is most of your volume and collection reliability varies by courier and zone. Letting that factor into assignment protects cash you'd otherwise chase for weeks.

Point your agent at the endpoint

Open the Wassel dashboard, copy the MCP endpoint into your Claude, ChatGPT, or Cursor config, and connect your store and the courier keys you use. Start with two couriers and two or three rules, run it on a day of orders, check the assignments match what you'd have done by hand, then expand the ruleset.

Keep the ruleset honest: it should reflect how your couriers actually perform by zone, not the order you signed them. Revisit it when performance shifts. If a courier you use isn't connected yet, tell us which one.