Automotive dealership tele-sales
Hundreds of customers, three messaging platforms, one inbox, shared among the agents who are online
At Elmongy Cars, an automotive dealership, hundreds of customers a month write in on WhatsApp, Facebook and Instagram, and the tele-sales team works them from one inbox. Each new customer goes, in turn, to an agent who is online; when a customer writes again and their agent is away or slow to answer, a colleague steps in automatically, and a customer left waiting is escalated to a manager. Two audits of the dealership's own data showed what happens when everything depends on one agent's screen and attention.
The Business Problem
Many of the customers who write to Elmongy Cars' tele-sales team come from paid ads, and each needs an owner fast. When a customer belongs to one agent, everything rides on that agent being online, seeing the conversation and getting to it in time; if any of the three fails, a high-intent buyer waits, and many will simply ask another dealer.
Such failures are easy to miss: an agent can have a full screen and a manager can see everyone busy replying, while neither sees the customer still waiting.
Our Approach
We diagnosed the problem with the system's own data before writing the fixes. Two dated audits of the dealership's messaging records showed two different failure modes, and each needed a different answer.
Visibility. Our audit of the dealership's own messaging data found that, in the 30 days to mid-October 2026, 457 conversations in which the customer had written last, across WhatsApp, Facebook and Instagram, had been hidden from the agent handling them; at least 166 were clear questions. Background clean-up jobs hid conversations once a set time had passed, without checking who had spoken last. They were invisible, not abandoned.
They were invisible, not abandoned.
Attention. A second audit at the end of October 2026 found 56 conversations waiting more than an hour; in 49 of them the assigned agent had been replying to other customers in the meantime: an attention problem rather than an availability problem. That calls for prompts, and for a manager when a prompt is not enough.
The answer had to do three things: stop the system hiding work, stop relying on one person's attention, and share customers among whoever is online.
Delivered Solution
The work sits in a purpose-built tele-sales add-on, designed so that a customer does not depend on one agent being online. Live in the dealership's system:
- One inbox for three platforms. WhatsApp, Facebook and Instagram conversations arrive in one inbox in the dealership's Odoo system, and the same rules decide who handles each one.
- New customers shared in turn. Each new customer goes, in turn, to an agent who is online, meaning active in the last five minutes, with a fallback when no one is; assignment is engineered so that two near-simultaneous messages cannot be given to two agents.
- Cover when an agent is away or silent. If a customer writes again and their agent is away, or is online but silent for 15 minutes, an online colleague is assigned automatically. A conversation the agent has been actively replying to in the last 20 minutes stays with that agent.
- Reminders and manager escalation. A customer left waiting triggers a reminder to the agent after 30 business minutes and an alert to the sales manager after two business hours (10:00 to 23:00, Cairo time); polite sign-offs such as تمام or شكراً do not trigger it. Twenty hours after a customer's last WhatsApp message, the agent is warned that Meta's 24-hour reply window closes in four hours.
- A visibility guard. A conversation where the customer spoke last stays on the agent's screen for seven days.
- Health checks. If no messages arrive from a platform for two hours during working hours, the team is alerted; conversations with no assigned agent, customers waiting in a hidden conversation and slow message deliveries are flagged too.
Built for a live sales floor
The handover timings were set from 30 days of the dealership's real reply times, not from defaults. The cover automation reassigns at most five conversations a minute, so a backlog cannot flood colleagues' inboxes at once; it has an off / shadow / live switch and logs every routing decision.
Evidence-Backed Value
- A measured, dated diagnosis on the dealership's own data, separating a visibility problem from an attention problem. After the fix, a production check from 17 October to 6 November 2026 found no conversation awaiting a reply hidden within seven days.
- Estimate: at one minute each, finding and handing back by hand the 457 conversations our audit found hidden in 30 days would take a supervisor almost eight hours a month; the system now keeps them visible for seven days.
- Expected benefit: the system is designed so that enquiries, many of them from paid advertising, reach an agent who is online, and so that agents can still reply freely before Meta's 24-hour window closes rather than having to re-open the conversation with a template message.
- Asked back: Elmongy Cars has since asked us for further additions to the system.
What This Means for Your Organization
If your customers reach you through messaging, three questions matter: who gets each new customer, what happens when that person is offline or busy, and who notices a customer still waiting. Some of those customers may be waiting where your team cannot see them, and the usual dashboards may not show it. Where a live system exists, we measure it before we fix it; for the automation that reassigns live conversations, we built in a shadow mode and a log of every routing decision.
Bring one process that frustrates your teams.
In a first working conversation, we will walk through:
- where it waits unseen
- where its numbers are re-keyed
- which of its rules live only in people’s heads