Manual maintenance intake is a decision system, and the decisions it relies on are made by whoever picks up the call. Is this urgent? What category does it fall into? Where does it go? Every incoming request requires someone to answer those three questions, and the answers vary with the person. AI maintenance intake replaces those decisions with defined logic that runs the same way on every request, regardless of channel or time of day. What the AI executes is the classification. What stays with you is the logic it runs on and every decision that involves committing money or dispatching a contractor.
Here is the problem the insight behind this article describes. It is a Tuesday afternoon and two residents in the same building call in within thirty minutes of each other. Both describe water under the kitchen sink. The morning coordinator took the first call, recognized the phrasing as a possible supply line issue, and escalated it immediately. The afternoon coordinator took the second, heard the same description, and logged it as routine follow-up. Same building, same week, same problem. One became an emergency work order. One became a note in the queue.
The building did not change between calls. The intake process did.
What the manual decisions actually are
When a maintenance request arrives, three decisions happen before the work order exists: classification (what kind of issue is this), urgency (does it need attention now or later), and routing (who receives it and in what form). In a manual intake system, all three are made on the fly by the person handling the contact.
At small scale this works reasonably well. A small team builds shared context. Staff learn the building's quirks. They recognize the resident who always escalates. They know the difference between a dripping faucet and a burst supply line even when the resident describes both the same way. The variability is real but contained because the same people are making the calls consistently.
As portfolios grow, the number of people making intake decisions grows with it. Consistency becomes a function of training, and training is never complete. The result is not a broken intake system; it is a variable one. And variable intake means variable outcomes: inconsistent work orders, uneven SLA enforcement, and maintenance teams that arrive with incomplete information because whoever took the call gathered it differently.
What AI intake replaces
Alex, the AI built into Scaalr, handles intake across phone and email. The replacement happens at each of the three decision points:
Information gathering. For phone intake, Alex conducts a structured conversation: the nature of the issue, the unit, access instructions, the resident's contact. For email intake, Alex reads the message as it arrives, extracts what is there, and flags what is missing. Either way the work order is built from a consistent set of fields rather than whatever the resident chose to mention and whatever the staff member thought to ask.
Classification and urgency. Alex applies the urgency logic you define: what qualifies as an emergency, what falls into routine, what category each issue type belongs to. The same burst-pipe description receives the same classification every time. A resident who phrases it as "water under my sink" and a resident who says "my kitchen is flooding" are evaluated against the same criteria rather than the intuition of two different staff members. How AI Detects Emergency vs Non-Emergency Maintenance Requests covers the classification logic in detail.
Routing. Once classified, Alex routes the work order to the queue you have defined for that combination of property, category, and urgency level. Routing rules are set by you; Alex executes them without discretion. The maintenance coordinator for Building B receives urgent HVAC requests for Building B, not a general inbox that someone sorts in the morning.
| The intake decision | Manual | AI |
|---|---|---|
| Information gathering | Whatever the staff member asks | Structured fields, consistent every call |
| Urgency classification | Staff judgment on the day | Defined criteria applied identically |
| Routing | Whoever the staff member knows to call | Predefined rules, no discretion |
| After-hours handling | Voicemail or answering service | Same logic, 24/7 |
| Dispatch decision | Staff or coordinator | Staff or coordinator |
What the operator defines
Replacing manual decisions with AI does not remove the operator from the intake system. It changes where the operator's judgment applies. Instead of being applied request by request by whoever happens to be working, it is applied once to the logic layer and then executed automatically on every request that follows.
You define what information Alex collects. You define what urgency thresholds mean: whether a gas smell is always an emergency or whether Alex should ask a follow-up question first. You define how requests route: which queue receives HVAC emergencies at the downtown portfolio, which coordinator handles plumbing at the suburban properties. You define what gets escalated after hours and what waits for morning.
The discipline this requires is different from training a coordinator. Instead of explaining what to do in a hundred situations, you write the rule that covers all of them. When the rule is wrong, you fix it once and every future request reflects the correction immediately. When your portfolio adds a building, you apply the same logic to it without onboarding another person into your intake culture.
What stays yours
The intake decisions that AI handles are the repetitive ones: gather, classify, route. The decisions that require judgment about a specific situation, a specific contractor, or a specific dollar amount stay with a person.
Dispatch stays yours. Which contractor to send, when to schedule them, whether to call the resident before sending someone: those decisions involve trade-offs that depend on context AI does not have. The work order arrives with complete, consistently structured information. The decision about what to do with it stays with the coordinator.
Spending stays yours. Any commitment of money, whether approving a repair estimate or authorizing emergency overtime, is a human decision. Alex does not spend money and does not commit your contractors to anything.
Edge cases stay yours. Alex flags requests that do not fit cleanly into defined categories and routes them to the person who can make the call. The goal is not to remove human judgment from intake; it is to stop applying human judgment to the cases that do not need it so that judgment is available for the ones that do.
What this looks like at portfolio scale
The operational argument for AI intake is not about the occasional misclassified call. It is about what variability costs when it compounds across hundreds of requests per week. Inconsistent urgency classification means SLAs are enforced on some requests and not others. Incomplete information means technicians arrive without what they need. Inconsistent routing means the right person sometimes finds out too late.
When intake is consistent, everything downstream is more predictable. Work orders arrive with complete information. Urgent requests reach the right person reliably, including at 2 AM when no coordinator is watching the queue. SLA performance becomes a function of your defined thresholds rather than whoever was working that shift. Reports reflect actual patterns rather than the artifact of how different staff categorized the same problems differently.
The AI-Powered Maintenance Triage and SLA Enforcement article covers the downstream enforcement model in full. This page is narrower: what happens at the moment of intake, which decisions move to AI, and which remain yours.
Key questions
What does AI maintenance intake actually do?
AI intake handles the three decisions that manual intake assigns to staff: what is this issue, is it urgent, and where does it go. It gathers the required information from the resident, applies the urgency and category logic you define, creates a structured work order, and routes it to the right queue. The same logic runs on every request regardless of channel, time of day, or who is working.
How is AI maintenance intake different from manual intake?
Manual intake distributes the classification decisions across people. Different staff interpret the same request differently, which means urgency and routing vary with whoever is working. AI intake concentrates those decisions into defined logic you set once. The variation moves from request outcomes to the logic itself: when the logic is wrong, you adjust it and every future request reflects the fix immediately.
What information does Alex collect from a maintenance request?
Alex gathers the information your workflow requires: the nature of the issue, the unit and location, access instructions, and contact details. For phone intake, Alex asks structured questions and confirms the answers before closing the call. For email intake, Alex reads the message as it arrives, summarizes it, and flags anything missing. Either way the work order arrives with complete information rather than whatever the resident chose to mention.
Can AI dispatch a contractor automatically?
No. Alex handles intake: gathering information, classifying urgency, creating the work order, and routing it to the right queue. The dispatch decision stays with a person. That is an intentional boundary: committing spend and scheduling a vendor involves judgment that the operator keeps.
Which Scaalr plan includes AI maintenance intake?
AI-powered intake and triage are available on Growth and up, with AI enabled on the account. Growth is $99 a month for the first 50 units, then $1.49 per additional unit. At 400 units that is $620.50 a month. Starter covers core operations without the AI features.