How to automate lead distribution in a real estate agency
In a real estate agency with no distribution rules, a new lead lands in a shared inbox or a Telegram group, someone writes "who's taking it?", two agents answer at once or nobody does, and the Friday-evening lead gets read on Monday morning. Meanwhile the client has called another agency. Automatic distribution solves exactly this: the lead has an owner in seconds, not hours, and someone finds out when that owner doesn't react. Below are the rules we usually configure, the trade-offs of each, the escalation ladder, the edge cases and what it all looks like in Telegram. Every figure here is an example parameter, not a result.
The problem, more precisely
"Who's taking it?" isn't an organisational question, it's wasted time with three effects: the lead waits, agents debate instead of calling, and the manager can't say afterwards who was supposed to respond and when. Any automatic rule — even the simplest — beats having none, because it removes the ambiguity. Everything else is fine-tuning.
Distribution rules that work — and their trade-offs
| Rule | How it works | The trade-off |
|---|---|---|
| Round-robin | Each new lead goes to the next agent on the list. | Fair and simple, but ignores expertise: the Botanica specialist gets the Ciocana lead. |
| By area | A lead marked "Botanica" goes to the agents covering that area; round-robin among them. | Needs the area filled in at the source. Leads without an area fall back to the default rule. |
| By transaction type | Sale, rental, purchase — different teams or agents. | A client who is "selling and buying" enters two flows; you have to decide where the main card lives. |
| By language | A lead from the Russian-language form goes to agents who prefer working in Russian. | Sources don't always state the language; it's inferred from the form or the first message. |
| With a workload cap | An agent with, say, 15 active leads receives no more until they close some. | Protects quality, but can leave leads waiting when everyone is full — hence the escalation to the manager. |
| "First who taps" | The lead is sent to 2–3 agents; whoever taps "Take it" first gets it. | Fast, but favours agents who are always on their phone, and needs escalation if nobody taps. |
| Owner of the existing client | If the number already exists in the CRM, the lead goes to the agent who owns it, regardless of the other rules. | You have to decide what happens when that agent has left or is on leave. |
In practice the rules are combined, in order: first "existing client", then "area" or "transaction type", then round-robin among the eligible agents, with the workload cap as a filter. The order is the agency's — we configure it from Web Admin, not in code, for the common cases.
The escalation ladder
Distribution without escalation is half a solution: the lead has an owner, but if the owner doesn't react, nobody knows. The ladder we usually set up looks like this (the minutes are an example; the agency sets them to its own rhythm):
- 0 min — the lead reaches the chosen agent in Telegram, with a "Take the lead" button.
- 10 min with no reaction — a reminder to the same agent.
- 15 min with no reaction — the lead is offered to the next eligible agent; the first one gets a message that they've lost it.
- 30 min with no reaction from anyone — the manager (or team lead) is notified and can assign it manually.
- Outside working hours (say after 20:00 or at weekends) — the lead is assigned, but the escalation timer starts at the beginning of the next working day, so nobody is penalised for a Saturday-night lead. Alternatively, an "on-call" agent receives everything that comes in.
What matters here isn't the numbers but the fact that every step is recorded: who received the lead, when, who took it, when. Which brings us to the next point.
Response time — the metric management gets for free
When taking a lead is a tap with a recorded time, "response time" stops being an estimate and becomes a column in a report: per agent, per source, per day. The manager sees, without asking anyone, who takes leads within five minutes and who within two hours, and whether Meta leads wait longer than website leads. We don't promise that this number improves by itself — but it becomes visible, and what's visible gets discussed at the meeting with data rather than impressions.
The edge cases the rule has to anticipate
- Agent on leave. An "unavailable" status in the CRM removes them from the rotation; their existing leads stay with them or are temporarily redistributed, as the agency decides. Whoever forgets to set their status creates a gap in the chain — which is why escalation to the manager is mandatory as a safety net.
- Duplicate lead from two channels. The same client fills in the website form and then calls. Without de-duplication, two agents receive "two" clients. The "owner of the existing client" rule, applied to the phone number, handles most cases; the rest we cover in the article on integrating the sources.
- A client returning after 6 months. The number exists, but the card is closed ("Lost" or "Sold"). Does it go to the old agent or come in as a new lead? There's no universal answer; the agency chooses and the rule respects the choice (for example: to the old agent if they still work here, otherwise round-robin).
- A lead with insufficient data. The form only has a phone number, no area or type. The area rule can't apply; the lead falls back to the default rule (round-robin), and the first call fills in the rest.
What it looks like in Telegram
The agent sees none of the logic above. They see a message:
New lead · Meta ("Botanica" campaign)
Looking for: 2–3-room apartment, Botanica or Centre
Budget: up to €75,000
Phone: +373 •• ••• •••
[Take the lead] [Call now] [Not for me] [Open card]
"Take the lead" records the time and creates the "First contact" task with a due time. "Not for me" passes the lead on immediately, without waiting the 15 minutes — and asks for a short reason, so the manager can see whether the distribution rule is systematically wrong. The manager, in Web Admin, sees the list of unassigned or overdue-untaken leads, with the current owner and the escalation history. The same flow, with the same buttons, appears on the real estate CRM page and in the general description of Telegram CRM.
How to start: one rule, then measure, then adjust
The classic mistake is trying to encode every exception from day one: "by area, but if it's a rental then team B, but if it's an old client then…, but on Fridays…". The result is a rule nobody understands any more and agents work around. Our recommendation:
- Weeks 1–2: a single rule (round-robin or by area) plus escalation to the manager. That's all.
- Week 3: look at response time and at how many leads were declined with "Not for me", and why.
- Then: add a second rule only if the data shows a real problem (for example, rental leads always ending up with sales agents).
Rules change from Web Admin, without a developer, for the common cases — so adjusting isn't a project, it's a setting. More on how we build this kind of automation beyond real estate on the process automation page.
What is NOT worth automating
Some agencies prefer that obviously large leads — an investor asking about a whole building, a company looking for ten apartments for staff — be assigned by the manager, not by a rule. That's a legitimate choice. The CRM can flag a lead as "VIP" by simple criteria (budget above a threshold, a specific source, keywords in the message) and send it to the manager for manual assignment instead of putting it in the rotation. Automation is for the daily volume; the valuable exceptions can stay with a human's judgement, if that's what the agency wants.
Read on: CRM for real estate agencies · Telegram CRM · Meta + telephony + CRM: how a lead reaches the right agent directly.