Expenses and utilities
A bill or an expense line that is well above the same property's history, or above similar properties of the same size.
Use Case
Find out when something in your portfolio needs attention
You should not have to read the same reports every week to find the one line that matters. Exception reporting checks every property for you and flags what is unusual, so you only look at the data when something needs your attention.
Anomaly detection, which many operators call exception reporting, means your portfolio data is checked for you and you are told when something is unusual. Propexo brings every system's data into a warehouse you own, and an agent compares each property with its own history and with similar properties. It flags what is outside the normal range and sends it to the person who can act. You look at the data when there is something worth looking at.
For asset managers
What does Monday morning look like for an asset manager?
It is Monday morning and the weekly reports are in. You oversee thirty properties, and each one has a rent roll summary, a delinquency report, a budget variance report and a maintenance summary. You open the first one.
Most of what you read looks the same as last week. Occupancy moved by a fraction of a point. Expenses are close to budget. You keep reading anyway, because the one line that matters could be on any page, and you will not know which until you get there.
Sometimes you find it. A water bill has doubled at one property, or delinquency at another has been rising slowly for six weeks. Sometimes you do not find it, and you hear about it on the owner call, a month after it started costing money.
Then there is the weekly call with your managers. It is booked for an hour and runs longer, because the only way to be sure nothing was missed is to go through every property in order.
What does reading the same reports every week cost you?
Example inputs. Replace with yours.
- Cost a year
- —
- Hours a year
- —
- Hours a week
- —
- Reports read a year
- —
Why are problems easy to miss in weekly reports?
A report shows you everything, every week, with the same emphasis. The property that is fine and the property with a problem take up the same space on the page. Finding the problem is left to the reader, and the reader has many more reports to get through.
- 01 Property 1: nothing to review
- 02 Property 2: nothing to review
- 03 Property 3: nothing to review
- 04 Property 4: nothing to review
- 05 Property 5: nothing to review
- 06 Property 6: nothing to review
- 07 Property 7: Delinquency above its normal range
- 08 Property 8: nothing to review
- 09 Property 9: nothing to review
- 10 Property 10: nothing to review
- 11 Property 11: nothing to review
- 12 Property 12: nothing to review
- 13 Property 13: nothing to review
- 14 Property 14: nothing to review
- 15 Property 15: nothing to review
- 16 Property 16: nothing to review
- 17 Property 17: nothing to review
- 18 Property 18: nothing to review
- 19 Property 19: nothing to review
- 20 Property 20: nothing to review
- 21 Property 21: nothing to review
- 22 Property 22: Water cost per unit above similar properties
- 23 Property 23: nothing to review
- 24 Property 24: nothing to review
- 25 Property 25: nothing to review
- 26 Property 26: nothing to review
- 27 Property 27: nothing to review
- 28 Property 28: nothing to review
- 29 Property 29: nothing to review
- 30 Property 30: nothing to review
In most weeks, most properties need nothing from you. What you need to know is which ones do. People tell us they would rather not read the same reports week after week, and would rather open the data only when something noteworthy is happening.
Exception reporting
Every property is checked, and only the exceptions are sent.
Once every system's data is in your warehouse, each property can be compared with its own history and with similar properties. Nothing is sent for a number inside the normal range. A number outside it goes to the person who can act on it.
- PMS Leasing and rent occupancy, delinquency, renewals
- Accounting and utilities Expenses invoices and bills, by property
- Maintenance Work orders opened, assigned, closed
Your warehouse · your account Connect keeps the pipes running
Data Map · one definition of each metric Each property against its own history and its peers an agent checks every metric on the schedule you set and flags what is outside the range you chose SourcePMSAccountingMaintenance
- Exception report to the asset manager DefinitionSource
- Alert with the evidence to the property manager DefinitionSource
- Answer in Claude or Copilot through Propexo MCP DefinitionSource
How it decides
What counts as an anomaly?
Each metric at each property is compared with two things: that property's own recent history, and similar properties in your portfolio. A number is flagged when it is outside the range that both comparisons would lead you to expect.
Water cost per unit · Property 22
1 2 3 4 5 6 7 8 9 10 Flagged 11 12
This is different from a fixed threshold. A rule that flags any property above a set delinquency rate fires every week for the property that always runs high, and says nothing about the one that has doubled from a low base. Anomaly detection looks at whether the number is unusual for that property.
Because every system is in one warehouse, a flag can draw on more than one of them. A flag for a rise in water cost comes with that property's occupancy and open work orders, so you can see what else changed at the same time.
What it flags
What can it flag across the portfolio?
You choose the metrics. These four are common starting points.
Delinquency and bad debt
Delinquency rising at one property while the rest of the portfolio is flat, or a group of new residents who stop paying in their first months.
Renewals at risk
Leases coming up for renewal where the resident has open work orders or late payments, flagged while there is still time to act.
Work that has stalled
Work orders, unit turns or vendor jobs that have been open longer than the limit you set.
Accountability
Who was told, and what did they do about it?
A flag is only useful if someone acts on it. Each one goes to a named person, such as the property manager, a regional manager or a vendor, with the evidence attached. It stays open until that person responds.
| What was flagged | Where | Sent to | What happened |
|---|---|---|---|
| Delinquency above its normal range | Property 07 | Property manager | Acknowledged |
| Water cost per unit above similar properties | Property 22 | Regional manager | Leak found and fixed |
| Work orders open longer than your limit | Property 15 | Maintenance vendor | No response yet |
Each exception leaves a record: when it was raised, who received it, and what happened. On the weekly call you go through the open flags, not every property. Over time the record shows which properties, managers and vendors resolve problems quickly and which do not.
Act
What does the agent do, and what stays with you?
The agent does the checking. It runs on the schedule you set, across every property and every metric you chose, and writes an exception report of what it flagged. In a quiet week the report is short.
You set the ranges, choose who receives each kind of flag, and decide which alerts go out on their own and which wait for your approval. The connection to each of your systems is read-only. See Agents for how an agent runs, or the Agent Lookbook for more examples.
Exception reporting works once your portfolio data is already in one dataset. If your reports are still built by hand from several systems, start with one portfolio view.
How it gets done
How to set up exception reporting and anomaly detection with the Propexo Platform
There are five steps, and you choose what is watched and who is told.
-
Choose the metrics to watch and who owns each one
List the numbers you check every week, such as delinquency, expenses against budget, utility cost per unit and open work orders. For each one, name the person who should hear about a problem first.
-
Bring each system's data into your warehouse with Connect
Connect has pre-built connectors for the major PMS platforms and for accounting, utility and maintenance systems, 127+ proptech systems in all. Each system's data arrives unchanged in a warehouse you own, so every flag can be traced back to the record behind it.
-
Agree definitions in Data Map
Data Map recognises the same property, unit and resident across every system and holds one definition of each metric, approved by your team. A property is then compared with others on the same basis, whatever PMS it runs on and whoever manages it.
-
Set the ranges and the recipients
For each metric, decide how far from normal counts as unusual and who receives the flag. You can start with a few metrics and wide ranges, then add more once you see what is useful.
-
Read the exception report and review the open flags
The agent checks every property on the schedule you set and sends you an exception report of what it flagged. On the weekly call you review the open flags and what was done about each one.
Anomaly detection, which many operators call exception reporting, means your portfolio data is checked for you and you are told when something is unusual. Propexo brings every system's data into a warehouse you own, and an agent compares each property with its own history and with similar properties. It flags what is outside the normal range and sends it to the person who can act. You look at the data when there is something worth looking at.
Frequently asked questions
What is anomaly detection for a multifamily portfolio?
It is an automated check of your portfolio data that tells you when a number is unusual. Each metric at each property is compared with that property's own history and with similar properties. When a number is outside the normal range, it is flagged and sent to the person responsible.
Is exception reporting the same as anomaly detection?
They do the same job. Exception reporting is the longer-standing name for a report that lists only the items outside a set limit. Anomaly detection sets that limit from each property's own history and from similar properties, in place of one number for the whole portfolio. This page uses both terms for the same thing.
How is this different from a dashboard or a scheduled report?
A dashboard or a report shows you everything and leaves you to find what matters. Exception reporting does the looking for you and sends only the exceptions. You can still open the full report when you want it. You no longer have to read it every week to be sure nothing was missed.
How is an anomaly different from a threshold alert?
A threshold alert uses one number for the whole portfolio, such as any property above a set delinquency rate. It fires every week for a property that always runs high and misses a property that has doubled from a low base. An anomaly is measured against what is normal for that property.
Will we get too many alerts?
You choose which metrics are checked and how far from normal counts as unusual, so you control how many flags are raised. Each flag names one property and one metric. A week with nothing unusual produces a short exception report.
Who receives the alerts?
The people you name. A flag can go to the asset manager, the property manager, a regional manager or a vendor, depending on the metric. Each flag keeps a record of who received it and what they did.
Can an owner use this for properties a third-party manager runs?
It depends on the access your management agreement gives you. Connect needs credentials for each system, and at a managed property the manager holds them. Where you have that access, the data goes into a warehouse account in your name and the same checks run on it.
What do we need in place to start?
Credentials for the systems you want landed, and a warehouse account or the managed Snowflake option inside Connect. No data team, and no custom build on your side.