Off-market means the listing is not on the MLS, not on Zillow, and ideally not visible to anyone who is not a serious, financially capable buyer. The whole point is to spare the seller from showings, low-ball offers, and neighborhood gossip. But a generic site form is a sieve. You either publish enough detail that the address is guessable from Google Street View, or you publish so little that nobody knows whether to inquire. Either way, your inbox fills with tire-kickers and you spend Tuesday afternoon explaining to a 19-year-old that no, the $2.4M townhouse is not available for student rental.

SleekAI maps a private listings custom post type with fields for price_band, neighborhood, beds, baths, square_feet, and seller_motivation, plus a buyer-side preapproval_status and budget_ceiling. Display conditions scope the bot to the off-market page. Visitors describe what they want. The bot matches against criteria, asks the qualifying questions a junior agent would ask, and only after pre-approval is confirmed does it share neighborhood and rough price band. Full address never leaves the database.

Generic bots fail because they cannot read a gated CPT, do not understand pre-approval as a gating step, and tend to either spill everything or refuse outright. SleekAI handles the in-between, treating the chat as a qualifying conversation that ends with a calendar link or a polite deflection, not a database leak.

Off-market real estate is a relationship business, but the first touch is rarely a relationship. It is a stranger on a website typing a question. The cost of treating that stranger like a qualified buyer is wasted hours, leaked address detail, and irritated sellers asking why their pocket listing is being shown to unqualified passersby.

The cost of treating that stranger like a tire-kicker is missing the one in twenty visitors who actually has the budget and timeline you need. A bot that runs the same qualifying script every senior agent runs, in the same warm tone, solves both ends. It treats every visitor seriously enough to ask the right questions, then it routes based on the answers.

Sellers stay happy because their address never appears in a casual reply. Buyers stay engaged because the bot does not behave like a contact form. Agents reclaim the hour a day they spent on phone screens that went nowhere.

The qualifying questions are the same five questions in every market, so once the system instruction is dialed in for your brokerage it requires almost no maintenance. New off-market listings appear in the bot the moment the post is published, and disappear when the post is unpublished. There is no separate spreadsheet to maintain.

The CRM gets a cleaner pipeline of buyers with stated budgets, lenders, and neighborhoods, which makes the next pocket-listing introduction five times faster.