No AI leasing agent vendor publishes a price. We checked all ten of the major ones in July 2026 and every single site routes you to a demo request. Independent buyer guides put the enterprise going rate at roughly $3 to $6 per unit per month, which puts a 300-unit property near $900 to $1,800 a month, usually with a portfolio minimum attached.
Last updated July 2026.
That opacity is the single most frustrating part of buying leasing AI, and it is not an accident. Per-unit pricing is negotiated against portfolio size, so publishing a number would anchor every deal. The practical effect is that you cannot build a budget line before you have sat through three discovery calls.
This article gives you what the vendors will not: the pricing models actually in use, the math for a portfolio your size, the costs that show up after signature, and the questions that get you a real number faster.
How much does an AI leasing agent cost?
Expect roughly $3 to $6 per unit per month from the established multifamily vendors, with a minimum contract that typically rules out portfolios under a few thousand units. On a 300-unit property that is about $900 to $1,800 a month, or $10,800 to $21,600 a year. Those figures come from independent buyer guides rather than the vendors, because none of them publish.
Two things widen that range more than anything else. Voice is the first: an assistant that answers the phone costs meaningfully more than one that handles chat, text and email, because call handling is expensive to run and expensive to get right. The second is whether the AI is sold standalone or bundled into a property management system you already pay for.
What we found when we checked every vendor
We went through ten vendor websites in July 2026 looking for a published price. We found zero.
| Vendor | Assistant | Published price? | Pricing model in use |
|---|---|---|---|
| EliseAI | Elise | No, demo request | Per unit per month, portfolio minimum |
| RealPage | AI Leasing Agent, with Knock CRM | No, demo request | Quoted against the wider RealPage contract |
| AppFolio | Lisa | No | Sold alongside AppFolio Property Manager |
| Yardi | RentCafe Chat IQ | No, demo request | Quoted inside the Yardi stack |
| Entrata | ELI+, built on the 2024 Colleen AI acquisition | No, demo request | Quoted inside the Entrata stack |
| Zuma | Kelsey | No, demo request | Per unit per month |
| Respage | ResMate | No | Per property rather than per unit |
| LeaseHawk | ACE | No, demo request | Per property, voice-led |
| BetterBot | BetterBot | No | Per unit, reported lower than the field |
| Funnel | Funnel, CRM-first | No, demo request | Quoted as part of the CRM |
Checked against each vendor website in July 2026. We have deliberately not filled that third column with numbers scraped from competitor blog posts. Where a figure is genuinely well corroborated, like the $3 to $6 per unit band or BetterBot's reported $1.50 to $3, we say it came from third-party buyer guides and not from the vendor, because that is the truth and it changes how much weight you should give it.
Why is AI leasing agent pricing per unit?
Because it maps to the customer's own revenue and it grows automatically. A per-unit fee scales with your portfolio, so the vendor's revenue rises every time you take on a property, without a renegotiation and without any relationship to how much work the AI actually does.
Look at what that means in practice. If you add 200 doors that lease up quickly and generate almost no inbound questions, your bill rises by the same amount as if you had added 200 doors with heavy turnover and constant inquiries. You pay for units, not for conversations. In a stabilized portfolio with low turnover, that is the wrong meter: the AI is answering a fraction of the volume you are being charged for.
The model works in your favor in exactly one situation, which is high-turnover, high-inquiry lease-up. There, the per-unit fee is fixed while the AI absorbs a genuinely large amount of work.
What does an AI leasing agent cost for a small portfolio?
Often nothing, because you cannot buy one. This is the part of the market nobody writes about honestly: the enterprise vendors attach portfolio minimums that make an operator with 200 to 2,000 doors uneconomical to sell to. You will book the demo, answer the portfolio-size question, and stop hearing back.
Below the minimum, the realistic options are the assistant bundled into your PMS if you run AppFolio, Yardi, RealPage or Entrata, a chat-only tool like BetterBot, or a general messaging agent priced flat. Our own AI leasing agent sits in that last group at $79 to $499 a month regardless of door count, which is a different trade: no native PMS writeback, no outbound calling, but a price you can read and a bill that does not move when you add a property.
What costs show up after you sign?
The per-unit fee is rarely the whole number. Four line items catch buyers out.
- Implementation and onboarding. Often a one-time fee, and always a real internal cost in staff hours. Someone on your team has to define the prescreening questions, the escalation rules and the tone, then test it against real inquiries.
- Telecom pass-through. Text messages cost money. Check whether SMS and voice minutes are included, billed at cost, or resold at a margin, and whether 10DLC registration for business texting is handled and billed to you.
- Integration fees. A two-way sync into Yardi or RealPage sometimes carries its own charge, on either the AI vendor's side or the PMS side, and occasionally both.
- Contract length. Annual is the norm and multi-year is common at the enterprise end. If the AI underperforms in month three you are usually still paying in month eleven, so weigh the term as heavily as the rate.
Is an AI leasing agent worth the cost?
The honest answer is that it depends on how many inquiries you currently miss, not on how good the AI is. The value is almost entirely speed to lead and after-hours coverage. A prospect who messages five communities on a Saturday tours the ones that answered, so if your leads already get a reply within minutes during the hours they arrive, an AI leasing agent has little left to capture and the per-unit fee is hard to justify.
Run the arithmetic on your own leakage before you run it on the vendor's case study. Take your monthly inquiries, the share that currently get no reply within an hour, your historical inquiry-to-tour and tour-to-lease rates, and your average rent. If recovering even a handful of those leaks covers the annual fee, the case is real. If it does not, no amount of demo polish changes the answer.
Be skeptical of vendor-published lift figures on this point. Any single precise percentage in a leasing AI advertisement is marketing, not research, and the direction is the only part you should treat as reliable.
How do I get a real price out of a vendor faster?
Lead the first call with the four facts that determine your quote, before they run discovery: unit count, property count, which PMS you run, and whether you need voice. That collapses three calls into one and gets you a number in the first meeting.
Then ask these directly, and get the answers in writing:
- What is the rate per unit per month, and what is the minimum contract value?
- Is the fee per unit, per property, or per conversation, and what happens when I add or sell a property mid-term?
- What is the implementation fee, and what does onboarding require from my team in hours?
- Are SMS and voice included, passed through at cost, or marked up?
- Is the PMS integration included, and does it write guest cards back automatically?
- What is the contract term, and what are the exit terms if it underperforms?
The pricing question is only half the diligence. The other half is compliance, and it is the half that carries real risk in US multifamily. Any automated system touching prospect conversations has to ask every applicant the same objective questions, must never ask about or act on protected characteristics under the Fair Housing Act, and must not make the approval decision. That is a configuration problem rather than a software problem, which means the burden sits with you. Property managers already running structured vendor compliance work, where insurance certificates get tracked and verified before a contractor sets foot on the property, tend to handle this well, because it is the same discipline: define the rule, apply it uniformly, keep the record. Have counsel review your prescreening question set before the AI sends its first message.
AI leasing agent cost compared to a leasing consultant
This comparison gets made constantly in vendor decks and it is mostly misleading. A leasing consultant tours units, reads a hesitant prospect, handles exceptions and closes. An AI leasing agent answers repetitive questions and books appointments. They are not substitutes, so a cost-per-head comparison is measuring the wrong thing.
The defensible framing is coverage. An AI leasing agent buys you response coverage in the hours and volumes a staffed office cannot reach: evenings, weekends, and the twenty simultaneous inquiries that arrive after a listing refresh. Priced against that, a few thousand dollars a year for a 300-unit property is reasonable if those hours are currently dark, and poor value if they are not.
The short version
Budget $3 to $6 per unit per month for an enterprise multifamily AI leasing agent, expect a portfolio minimum, and expect to sit through a demo to confirm it. Add implementation, telecom and integration on top, and read the contract term as carefully as the rate. If your portfolio sits under the minimum, your realistic paths are the assistant bundled with your PMS or a flat-priced messaging agent.
Whichever way you go, the number that decides the purchase is not the vendor's rate. It is how many inquiries you are currently losing to silence. Work that out first, because it is the only figure in this decision that is genuinely yours.
MessageAgent · Get started
Put one AI on every channel
One agent that answers, qualifies, books, and upsells across SMS, WhatsApp, Instagram, web chat, and email, in one inbox. AI is always disclosed, with human handoff built in.