Real estate teams: leads, follow-ups and nothing falling through

By the Onsites AI team · Last updated August 28, 2026 · 5-minute read

Real estate support is speed with a commission attached. The inquiry that lands at 9:04 and gets answered at 9:08 books the viewing; the same inquiry answered after lunch competes with three other agents who saw it first; the one answered tomorrow never calls back. Studies of lead response keep finding the same cliff — minutes matter, hours are fatal — and agency life makes the cliff harder: agents are in viewings, on the phone, on the road, while inquiries arrive by portal, email, chat and phone into whatever desk exists. This guide builds the pipeline that catches everything: first reply under five minutes, structured, every lead a thread with an owner, follow-ups scheduled rather than remembered — and a desk shape that prices for it (three seats free, $15 beyond, AI drafts at published credits for the follow-up volume agents actually do).

The response cliff, and what feeds it

The lead-response research is consistent enough to plan around: contacting a lead within five minutes multiplies contact rates; after an hour, contact rates fall steeply; after a day, the lead has usually spoken to a competitor. Feeding the cliff: portal fragmentation — listings push inquiries to mailboxes with no clock, and portals' own lead alerts arrive as another unread tab. Agent shape: viewings and calls occupy the hours inquiries arrive; whoever is free triages the rest. Follow-up decay: the second and third touches — where most deals actually close — decay fastest, because they are remembered rather than scheduled. The desk design meets each: one queue for portal, email and chat (portals forward in as conversations), reply clocks visible to the whole team, follow-ups sequenced rather than scheduled in someone's head. The cliff is a systems problem; systems close it.

The lead's first hour, rehearsed

Because the cliff is won in the first hour, rehearse that hour as a script the desk makes repeatable. Minute 0–2: the inquiry lands (portal forward, email, widget); routing assigns it — the listing agent when the property matches, the least-loaded agent otherwise — and the desk's SLA clock starts. Minute 2–4: the copilot drafts the first reply from the listing card (price, viewing slots, the personal line answering the buyer's actual question); the agent's phone buzzes. Minute 4–5: the agent edits two words on the train and sends; the reply's last line asks the question that moves the deal ("Thursday 5 pm, or Saturday 10?"). Minute 60: if no booking, the desk's clock flags the lead for a call — the second touch, the one manuals always forget. That evening: the sequence schedules day-3 ("two more viewings opened") with the draft ready. The hour costs the agent perhaps four minutes of attention; the desk does the rest; the buyer's review of the agency months later will say "they got back to me immediately." Rehearse it once as a policy, and the desk makes every hour of inquiry-hours behave like the rehearsed one.

The five-minute reply, made structural

Five minutes cannot be a hero's habit; it has to be a system. Route by who can answer: inquiries route to the listing agent by property, or rotate across the team by load — the desk's assignment rules do the routing so triage is not a person at a browser. Draft instantly, review in seconds: the copilot drafts the first reply from the listing's data (address, price, viewing times, the one line of personalization the buyer's question earned) — the agent edits on the phone in the car and sends within minutes; the copilot discipline keeps human eyes on what becomes a client relationship, and the credits line at real lead volume is single-digit dollars (published rates). The reply's shape matters: answer the question asked, state availability ("Thursday 5 pm or Saturday 10 am?"), ask one question back — speed plus a next step is what books viewings; speed alone just earns a "thanks."

The follow-up pipeline: where deals actually close

Most viewings come from the second and third touches; most teams die at the first. Sequence the touches: day-1 (the viewing ask), day-3 ("did Thursday work? two more viewings opened"), day-14 (a new listing matching their brief, with the note that makes personalization real), day-45 (the market update) — the desk's follow-up sequences schedule them with drafts ready, and the agent approves rather than remembers. The thread is the relationship: the lead who asked in March, viewed in April, and is ready in June is one scroll — brief, budget, the viewings they liked, the line they said about schools; agents hand over clients without handover losses. Pipeline stages visible: inquiry → replied → viewing → offered, with aging visible — the two-week-old lead sitting in "replied" is a morning task, not a memory. The pipeline guide's whole argument: leads leak between touches, and the desk is where the leak stops.

The team shape: one queue, many listings

Teams of two to eight agents share the desk's shape. One queue, routed: portals, email, the website's widget and chat flow to one desk; the buyers' accounts carry their threads; the seller's inquiries (their biggest anxiety) route to their own agent with the listing's data beside the reply. Handovers without losses: the agent on leave passes buyers; the desk's threads carry the memory so the covering agent answers like they were there. The hours shape: evening inquiries (the second-peeking work-life-balance crowd) become morning tickets with drafts — the desk holds the hours agents cannot, and publishes them honestly (the channel guide's expectation-setting applies). The pricing for a team: three seats free (a two-person agency rides $0 + credits), $15 per seat beyond — a team of six prices $45/mo billed annually, less than an hour's commission at most agencies' splits. The desk is the cheapest hire a real estate team makes; it is also the only one that never sleeps, forgets or goes to a viewing.

The worked week

One agency's numbers make the case concrete. A three-agent team, 60 inquiries monthly across portals and email, historically answered in a median of 3 hours (viewings booked: 18). After the desk: median first reply — 9 minutes (routing did what heroes didn't); drafted replies reviewed on mobile; follow-ups sequenced to day 45; viewings booked — 31, from the same 60 inquiries. The difference wasn't effort, and the agents didn't work a minute more: the queue caught what the inbox leaked, and the follow-ups that were remembered became scheduled with drafts. The pricing: zero software cost (three free seats), the credits line for 200-odd drafts under $10 monthly. An agency whose pipeline is its memory now has both in one place — and the commission arithmetic on one extra booking a month pays any desk a hundred times over. The cliff is real; the desk is cheaper than falling off it.