Why Does the Same WhatsApp Question Repeat Every Day?
The question repeats because staff lack authority to respond, so every refund or discount case gets routed to the owner even when it is simple and recurring. The issue is not staff hesitation but the absence of a clear delegation that lets them act without asking.
A customer service agent at a clothing store messages the owner four times in one day about refund requests under 150 riyals, while the owner is in a meeting and does not open WhatsApp for two hours. The customer waits, the agent waits, and a simple, predictable decision sits unresolved.
This delay costs more than time. Tabby's 2025 report notes that Saudi customers consider fast support among the most important factors for trust and loyalty,3 so every hour lost on a simple message reduces repeat purchase likelihood, much like when the same shipping-status question keeps repeating without an immediate answer.
What Separates a Case That Needs Approval From One That Doesn't?
The distinction rests on recurrence and the absence of dispute: if a case happens constantly with a predictable, undisputed outcome, it does not need approval every time. If it is rare, carries real financial risk, or is ambiguous, it belongs with the owner.
Reviewing three months of chat logs for a home goods online store typically shows most requests fall into three categories: refunds for a clear product defect, size or color swaps that do not match the order, and shipping delays past the announced date. These three categories recur without meaningful dispute, meaning the decision does not require fresh thinking every time.
A case where a customer requests a refund after a month of use, or a large discount over an undocumented complaint, remains a genuine exception worth the owner's time. This distinction between routine and exception shapes the table in the next step and keeps delegation from turning into uncontrolled authority.
How Do You Write the Limits Table in Numbers, Not General Words?
A practical table has only three columns: case type, maximum value in riyals or percentage, and the exact required action. General wording like refund in reasonable cases does not work, because it recreates the same problem once each employee interprets reasonable differently.
For a kitchenware online store, the table might read: full refund if order value is under 200 riyals, the product is unused, and receipt was within seven days; a discount of up to ten percent to compensate for shipping delays past the announced date; immediate swap for a wrong size or color without approval as long as the product is in original condition. Every line has a clear number and a defined action, leaving no room for personal interpretation.
This kind of table resembles what sales teams in other sectors need when a decision gets stuck between the field and the office; the problem is identical: no written limit that gives frontline staff the authority to act immediately within a known scope.
How Do You Apply the Table With the Same Team Within One Week?
Applying the table needs no new system and no change in team, just three sequential steps within one week using the tools and people you already have.
Spend one full day reviewing old WhatsApp or support ticket logs and extracting the recurring cases where staff previously consulted the owner. Then spend a day drafting the table with the staff themselves, since they know the real cases they face daily better than anyone else in the store. A children's goods online store team might discover in this session that ninety percent of their owner consultations repeat around just three cases, which alone narrows the table to a few clear lines.
After drafting, pin the table as a fixed message in the internal work group for at least two weeks before any revision, so the team gets used to referring to it instead of the owner. This pinning mirrors the idea behind a live handover log instead of a verbal note during shift changes; the principle is the same: written documentation stops the same question from repeating every time.
Where Does the AI Agent Fit, and Where Must a Person Decide?
An AI agent executes pre-written decisions without waiting for anyone, but it does not invent new decisions outside the table's limits. That distinction is exactly what determines where humans intervene and where automation is enough.
At an electronics online store, an AI agent on WhatsApp and the web receives a refund request, automatically checks the order value and purchase date against the store's records, and if the case falls within the pre-written limits, executes the decision immediately and informs the customer without anyone waiting. A refund request above the limit, or a case involving a customer who has complained repeatedly, gets escalated instantly to a human staff member because these cases need judgment and negotiation that automation cannot provide.
This split protects the store from two opposite mistakes: forcing a customer to wait on a simple case, or letting a system decide a sensitive case that deserves a human touch. Setting up this division requires business automation built on the same table the team drafted, not on general assumptions about customer behavior.
How Do You Measure the Impact, and When Doesn't This Fit?
The sharpest measure is the average reply time for refund and exchange requests, compared across two weeks before the table and two weeks after, not vague satisfaction indicators that are hard to track.
Log the time each refund request arrives and the time it gets a final reply, for two weeks before applying the limits table, then repeat the same measurement for two weeks after. If average reply time drops noticeably while error or complaint rates stay the same or fall, the table is working. A skincare store that ran this measurement found most delays happened during midday hours when the owner was away, confirming the real problem was missing delegation, not request volume.
Still, this idea does not suit every store. If your daily order volume stays under ten requests, a simple verbal agreement with your one employee may be enough, without the time cost of drafting and updating a formal table. The table becomes necessary once the team grows or multiple people handle replies, a common scenario as Saudi e-commerce registrations exceed 39,000,1 alongside near-total reliance on the internet for daily customer communication.2 If you are planning to grow your store or support team, the logical first step is contacting the ASLS team to review your recurring cases and draft a table that fits your actual business size.
Sources
- 1Monsha'at report: Saudi commercial registrations hit 1.7 million · Saudi Gazette (2025-09)
- 2السعودية تعزز حضورها الرقمي مع وصول انتشار الإنترنت إلى 99.6% وتوسع استخدام الخدمات الحكومية · The Saudi Times (نقلاً عن تقرير إنترنت السعودية 2025 - CST) (2026-08)
- 3تقرير تابي 2025: اكتشف سلوك المتسوقين في السعودية والإمارات · Akhbar24 (نقلاً عن تقرير Tabby) (2025-10)
