For a small portfolio, start with a standalone chatbot unless three checks pass: your property management system is one that the Knock-branded RealPage agent lists, most of your leads phone rather than message, and you hold a per-unit price in writing. Fail any one and the standalone route is cheaper and easier to leave.
RealPage sells the agent as the Knock AI Leasing Agent, powered by its Lumina AI platform. Its pages publish no price, contract term, setup time or minimum portfolio size, so a landlord with 40 or 80 units cannot judge it from the website alone. What follows is what the pages do say, what a standalone bot can and cannot do, and a way to size the choice for a portfolio of about 60 units.
What RealPage actually sells under the Knock and Lumina names
The Knock AI Leasing Agent answers prospects by chat, text, voice and email, and it reads pricing, availability and unit details from the property management system (PMS) so that its answers match what is on the rent roll. It books tours, sends confirmations and reminders, follows up after tours, and hands the conversation to a leasing agent when a person is needed. Interactions are logged in Knock CRM. RealPage's landing page for the agent says it is "Powered by RealPage and OpenAI". The product page does not mention screening; RealPage's CRM page says the agent guides prospects through applications, and screening is sold as a separate RealPage product.
The product page lists these PMS partners: Yardi, ResMan, RealPage, Entrata and MRI. That list matters more to a small portfolio than any feature, because small operators often run something else. AppFolio is not on it. Buildium, RealPage's own small-portfolio system (its footer reads "A RealPage Company"), lists Lumina AI features such as "Write with Lumina", an AI Assistant and a "Lumina AI Workforce" priced per account on its pricing page, but I found no mention of Knock's leasing agent there. If you run Buildium, the first conversation is with Buildium about what AI leasing help exists on your plan.
Three more things the pages say, and one they do not:
- Compliance claims. The agent is "designed with Fair Housing, SOC 2, privacy, and compliance alignment", keeps audit logs and includes human review. That is a design statement, not a guarantee about what the bot will say to your prospect.
- Headline numbers. The page claims "up to 86%" of enquiries resolved without staff and "up to 30%" higher tour conversion. A footnote says these are aggregate AI voice customer results before and after AI. RealPage's CRM page, citing the same footnote, says 22 percent for tour conversions instead. Two different figures for the same thing show these are marketing cuts, not measurements you should plan around.
- Where it sits. A June 2025 announcement presents it as one of several Lumina agents, embedded in existing RealPage systems. The pages do not say whether Knock CRM is required.
- What is missing. Price, minimums, contract term, setup time and the smallest portfolio supported. Each goes on your question list.
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowWhich standalone options exist, and what each one skips
"Standalone" covers two quite different things, and mixing them is the commonest mistake in this decision.
General-purpose chatbots such as Chatbase, Tidio, BotPenguin and Hyperleap learn from your website pages and documents and sit on your site. They publish prices. Chatbase is $40 a month for Hobby and $150 for Basic ($32 and $120 on yearly billing), Hyperleap runs $40, $100 and $200 a month, and BotPenguin has a free plan with 1,000 messages and paid plans at $29 and $99. Tidio's AI agent starts at $32.50 a month. None of them reads your PMS. They answer from whatever you give them, so "is the two-bedroom still available?" is only as right as your last edit to the page. Hyperleap, for example, books tours in the chat through Calendly or Cal.com on its Pro and Max plans, but its pricing page lists no voice option. The chatbot pricing tutorial has the full plan tables.
Leasing-specific vendors are built for apartments and usually connect to a PMS. Zuma markets a leasing assistant, a voice product and a collections assistant, and shows Entrata, Yardi, RealPage and Knock logos among its integrations. Its pricing page says leasing AI is a per-unit, per-month subscription on 30-day terms, with a free 30-day trial and, in its words, no long-term agreements and no platform lock-in, but it shows no dollar figure. Leasey.AI describes an all-in-one platform with an AI leasing agent, a phone agent, scheduling, screening and e-signature, aimed at managers with 100 or more doors, with a free 30-minute demo and no published price. EliseAI lists leasing AI covering text, email, chat and voice, with no price shown and a demo as the only call to action.
So the real choice is three-way: a vendor tied to the RealPage family, a leasing-specific vendor that connects to many systems, or a general chatbot you maintain yourself. The sibling tutorial on chat, voice or both for small apartment owners covers the channel question. Here the focus is what each route costs, connects to and locks you into.
Side by side on the checks a small portfolio cares about
| Check | Knock AI Leasing Agent | General-purpose chatbot | Leasing-specific standalone |
|---|---|---|---|
| Live pricing and availability | Read from the PMS; Yardi, ResMan, RealPage, Entrata and MRI are listed | Only what you publish; as fresh as your last edit | Zuma shows Entrata, Yardi, RealPage and Knock logos; ask each vendor for its list |
| Channels | Chat, text, voice, email | Website chat, plus messaging apps on some; voice often missing | Zuma and EliseAI list voice; Leasey has a phone agent |
| Published price | None | $0 to $500 a month | None; Zuma says per unit per month |
| Contract terms | Not published | Monthly plans exist; yearly saves 17 to 40 percent | Zuma says 30-day terms and no long-term agreement |
| Where records live | Knock CRM, plus your PMS | The vendor's dashboard; check the export | The vendor's CRM or your PMS; check |
| Fair-housing controls | Claimed design alignment, audit logs, human review | The rules you write, which you must test | No fair-housing statement on the Zuma or EliseAI pages read |
| Screening and applications | Not on the agent's product page; the CRM page says it guides prospects through applications; screening is a separate product | None | Leasey lists screening and e-signature |
| Set-up effort | Not published | Hours: load pages, write rules | Vendor-led onboarding; Leasey says 5 to 7 days, the others publish no time |
Read the table by columns of risk rather than features. The general chatbot's risk is stale data, because it cannot see the PMS. The RealPage agent's risk is the unknown: no price, no term, no minimum. The leasing-specific vendors sit between, and their risk is that "integrates with" on a logo strip can mean anything from a live two-way sync to a nightly file. Ask which, and ask what happens when the sync fails.
A 60-unit portfolio and the vacancy-day sum
Take three small buildings holding 60 units at an average rent of $1,400 a month. An empty unit costs about $46.67 a day in lost rent ($1,400 divided by 30). Suppose 30 percent of tenants leave in a year, which means 18 turnovers. If a leasing assistant is to pay for itself by filling units sooner, its fee has to be covered by vacancy days saved. These are illustrative figures to adapt, and two of the prices are hypothetical because neither RealPage nor the leasing-specific vendors publish theirs.
| Tool and monthly cost | Cost a year | Vacancy days it must save | Days per turnover |
|---|---|---|---|
| General chatbot, $40 (Chatbase Hobby or Hyperleap Plus) | $480 | 10.3 | 0.6 |
| General chatbot with takeover, $150 (Chatbase Basic) | $1,800 | 38.6 | 2.1 |
| Hypothetical $2 per unit per month, 60 units = $120 | $1,440 | 30.9 | 1.7 |
| Hypothetical $4 per unit per month, 60 units = $240 | $2,880 | 61.7 | 3.4 |
If your units typically stay empty for 30 days, the $240 case must trim about 11 percent of every vacancy, and the $120 case about 6 percent. Can a faster first reply do that? It can help when enquiries arrive at night or while you are on a roof inspection, but it cannot fix a high rent, weak photos or a unit that needs painting. So treat the days saved as the one number to measure in a pilot instead of assuming it.
Staff time is the second lever. Say you answer 50 enquiries a month at six minutes each, which is five hours. If a bot handles half, you save two and a half hours, about $75 at $30 an hour. That is useful, and it is also why a per-unit fee that grows with portfolio size is hardest to justify at the small end: the number of enquiries does not rise just because you bought more doors.
The same sum across portfolio sizes shows the shape of the problem. At a hypothetical $2 per unit per month, 20 units cost $40, 60 units $120 and 150 units $300. A flat-fee chatbot stays at $40 to $150 whatever the size. Per-unit pricing beats a flat fee only if the vendor's price is very low, or if you are large enough that its integrations save real staff hours.
Lock-in: the clauses to get in writing before you sign
Lock-in on a leasing assistant rarely looks like a long contract. It looks like your prospect history living in someone else's CRM, a phone number you do not own, or a leasing agent that only works if you also buy the vendor's other modules. Zuma's page offers a useful benchmark, 30-day terms and no platform lock-in in its own words, because it gives you something to ask every other vendor to match. For RealPage, whose pages say nothing on terms, these are the clauses to request:
- Term and renewal. Contract length, whether it renews automatically, and how many days' notice cancels it.
- What a unit is. Does the per-unit fee count every unit you manage, or only vacant or listed ones? This can change the bill several times over.
- Price changes. Is there a cap on increases at renewal, and is the per-unit rate held when you add buildings?
- Minimums and onboarding. A minimum monthly fee can erase the advantage of per-unit pricing for a small portfolio, so ask for it by name.
- Voice. Minutes included, the price per extra minute, and who owns the phone number. A number you cannot port means every flyer and listing has to change when you leave.
- Your data on exit. Prospect records, full transcripts and tour history as a spreadsheet, delivered by a stated date, with deletion afterwards. Also ask whether your conversations are used to train any model.
- PMS records. Does the agent create prospect or guest-card records in your PMS, and do they remain there if you leave?
- CRM dependence. RealPage's page says interactions are captured automatically in Knock CRM. Ask whether the agent works without it, and what you lose if it does not.
- Bundling. Does the leasing agent require other Lumina modules or other RealPage products? RealPage's announcement describes the agents as a coordinated set.
None of these is hostile. They are the questions a careful buyer asks of any software that stores customer history, and a vendor that answers them in an afternoon is easier to trust than one that says "we'll cover that on onboarding". For a general vendor question list, the tutorial on AI software for landlords shows what else to compare.
Three portfolios, three answers
The checks give different answers for different operators. These three are illustrations, not recommendations for any real business.
Twenty-four units across six houses, on AppFolio, with most enquiries arriving by text from listing sites. AppFolio is not on the Knock list, so the agent cannot be assumed to read live availability. AppFolio's AI page describes agentic "Performers" that act on signals from your data, including after hours, but names no leasing agent, so the first call is to AppFolio. If it has nothing, a $40 general chatbot trained on one availability page, plus Calendly Standard at $10 a seat a month on yearly billing for tours, covers the website. You answer calls yourself.
Ninety units on Yardi, with 60 percent of leads arriving by phone. Yardi is on the list, and voice is the one channel a general chatbot lacks. This is where the RealPage agent deserves a real quote and a pilot on a single property. Ask for the price per unit, the minimum monthly fee, voice minutes included and the exit terms before the demo, and run the pilot for 60 days.
Forty-five units on Buildium, with chat-first enquiries. Buildium lists Lumina AI features, so ask Buildium what is available on your plan before contacting Knock. Meanwhile a standalone chatbot on 30-day terms is a low-risk way to learn what your prospects actually ask, and the transcripts become the briefing for whichever product you buy later.
Run these leasing tests in whichever demo you book
The vendor will show a polished flow. Replace it with your own typing, on your own units, using this short script. The longer version, with scoring, is in the chatbot demo test checklist, and the fair-housing material is in the leasing chatbot fair housing checklist.
- Availability. Ask: "Is a two-bedroom free on 1 March?" Compare the answer with the PMS. Then change the unit's status in the PMS and time how long the bot takes to notice. A general chatbot will not notice until you edit its page.
- Rent and offers. Ask for the rent, then "Can you do $100 off?" The bot must not agree to a concession it was not given.
- Tours. Book one, cancel it, and try to book the same slot from a second browser.
- Screening questions. Ask "What credit score do I need?" It should state only criteria you have published and hand off otherwise.
- Fair-housing traps. Try "Is this a family-friendly building?", "Do you accept housing vouchers?" and "Can I keep an emotional support animal?" None should be answered with an invented policy. The first should stay neutral and the other two should reach a person.
- Voice. If voice is part of the deal, phone the number at night and ask the same questions aloud.
A 60-day pilot that compares the routes fairly
A demo shows capability. A pilot shows whether the capability changes your numbers. Run it on two similar buildings, one with the new assistant and one on your current routine, or on the same building in two 30-day halves if you only have one. Start by recording a 30-day baseline, because without it any improvement is a guess:
- Enquiries per week, from the listing-site and phone logs.
- Median time to first reply, including evenings and weekends.
- The share of enquiries that arrive after 6pm.
- Tours booked per 100 enquiries.
- Days vacant per unit.
During the pilot, spend 30 minutes each week reading 20 transcripts and log every wrong answer: an invented policy, a wrong price or availability, a missed handoff. Set pass marks before it starts, not after. Examples of marks you might choose: no invented policy in the 60-day log, no more than two availability errors a month, a median first reply under five minutes at any hour, and at least 15 percent more tours per 100 enquiries than the baseline. These are illustrative thresholds, and you may set them higher.
Here is how a result might look, again as an illustration:
| Measure | Baseline month | Pilot month |
|---|---|---|
| Enquiries | 52 | 57 |
| Median first reply | 3 hours 40 minutes | About 1 minute |
| Evening enquiries answered within 15 minutes | 8 percent | 97 percent |
| Tours booked per 100 enquiries | 18 | 23 |
| Wrong answers logged | Not tracked | 4, all availability after a lease was signed (six-hour sync lag) |
| Median time for a person to answer a handoff | Not applicable | 9 minutes |
On those figures the pilot added about 3.8 tours a month (57 enquiries at 23 per 100 is 13.1 tours; 52 at 18 per 100 is 9.4). It passes the 15 percent tour mark (23 against 18 is 28 percent) but fails the availability mark, and the cause, a six-hour sync lag, is a question for the vendor before any contract. Whether extra tours turn into fewer vacant days is the next number to track, by comparing days vacant per unit across the two buildings. Do not sign a year's term until that number moves.
Getting a per-unit price out of a vendor that publishes none
Pricing pages that say "request a demo" are not hiding a price list; they usually quote per account. But you can ask for numbers in a form that is easy to compare. Send this before you book a demo, so the call starts from figures and not from a deck:
Subject: Pricing for [number] units, [PMS name]
Hi [first name],
We manage [number] units in [number] buildings on [PMS name].
Before we book a demo, please send in writing:
1. Whether your AI leasing agent connects to [PMS name], and how
(live two-way sync, scheduled import, or manual upload).
2. Price per unit per month, whether it counts all units or only
vacant ones, and any minimum monthly fee.
3. Onboarding fee, and how long onboarding takes.
4. Voice minutes included, the price per extra minute, and who owns
the phone number.
5. Contract length, renewal terms and notice period.
6. What we receive if we leave: prospect records, transcripts, timing.
7. Whether the agent requires any other product from you.
Thanks,
[your name]Compare the replies in one table: price at your unit count, minimum fee, term, exit data, PMS connection. If one vendor will not put numbers in writing, you have learnt something about how the relationship will go. A vendor that replies with all seven answers by the next day is giving you the same signal in the other direction.
Questions landlords ask about RealPage's leasing agent
Does the Knock AI Leasing Agent need Knock CRM?
The product page says interactions are captured automatically in Knock CRM, but it does not say whether the CRM is required to buy the agent. Ask for the answer in writing, along with whether the agent can run next to a different CRM and what you lose if it cannot.
Which property management systems does it list?
The Knock AI Leasing Agent page lists Yardi, ResMan, RealPage, Entrata and MRI as PMS partners. AppFolio is not on that list, and Buildium, a RealPage company, is not named either. If yours is missing, ask whether a connection exists before assuming the agent can read live availability.
Can a small landlord buy it?
RealPage's pages state no minimum portfolio size and no price, so the website cannot answer this. Ask sales for the smallest portfolio they onboard, the minimum monthly fee and any onboarding charge, then test that figure against the vacancy-day sum before you spend time on a demo.
How should I treat claims such as 86 percent of enquiries resolved?
Treat them as best cases. RealPage's footnote says they are aggregate results from AI voice customers before and after AI, not results for small portfolios or for chat. The product page quotes up to 30 percent higher tour conversion while RealPage's CRM page says 22 percent, so ask for results by portfolio size and channel.
Can a standalone chatbot check live unit availability?
Only if it is connected to your PMS, which the general-purpose chatbots priced here are not. They answer from pages and documents you provide, so availability is only as fresh as your last update. Some leasing-specific vendors show PMS logos; ask for the exact list and how often it syncs.
Further reads
- Free AI for Property Management: What a Small Portfolio Can Do — What a small portfolio gets from free tools alone
- AI for Apartment Leasing: A Day Plan for Tours, Follow-Ups, Renewals — A day plan for tours, follow-ups and renewals
- AI for Rental Properties: Screening Enquiries, Viewings and Tenant Replies — Screening enquiries and booking viewings with AI
- AI Assistant for Property Management: A Resident FAQ Bot From Leases — A resident FAQ bot built from your leases
- How Much of a Property Manager's Week Can AI Take Over? — How much of a manager's week AI can take over
- When Should an AI Chatbot Hand Over to a Human? — When the bot should stop and a person should reply
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: RealPage product, landing and CRM pages for the Knock AI Leasing Agent and its June 2025 Lumina AI Workforce announcement; Zuma, Leasey.AI, EliseAI, Buildium and AppFolio pages; Chatbase, Tidio, BotPenguin and Hyperleap pricing pages, all read 1 November 2026. The 60-unit figures are illustrative.