HappyCo turn boards organize the turn; AI review reads the photos that flow through them. Student housing compresses a year of inspections into roughly two weeks, which means far more photos get captured than any human can look at twice. RapidEye analyzes the photos operators already capture in their existing workflow, comparing each unit against its own move-in baseline, and returns damage attribution, vendor verification and dispute-ready evidence. Nothing about how technicians, vendors or the turn board work has to change.
Turn season is a year of inspections inside two weeks
No other residential asset class has a calendar like this. Education Realty Trust's SEC filings describe a portfolio where "substantially all leases commence mid-August and terminate the last day of July," with units prepared "during the first two weeks of August" (sec.gov). Multi-Housing News reported that about 60% of student housing units turn over at the end of July (multihousingnews.com). The industry's own operating documents treat turn not as a workflow but as a season.
The pressure on that season is a leasing number. Preleasing across the Yardi 200 reached an estimated 89.1% in July 2026 (yardimatrix.com). Those beds are sold. The move-in date is fixed by an academic calendar nobody controls. Every unit that is not ready is a resident standing in a lobby with a signed lease.
And the work per bed is not trivial. Multi-Housing News put basic make-ready at roughly $135 to $165 per bed, while Inspection Express's 2026 vendor estimate puts a full turnover at $1,500 to $3,000 per bed (ipropertyexpress.com). Multiply either figure by a few thousand beds and the turn is one of the largest discretionary spends an operator makes all year, executed at the highest speed and the lowest staffing ratio of the year. We collected the rest of the numbers in our student housing turn season statistics.
The turn board is the system of record, and that is the right design
HappyCo built for exactly this. Turn boards are a genuine strength of the platform: every unit in the property becomes a card with a status, the inspections hang off the card, the findings roll into work orders, and a GM can see the whole property's readiness on one screen instead of on a whiteboard in a leasing office. When HappyCo is synced to a PMS, the board reflects what the PMS knows about the lease. HappyCo's integrations page names AppFolio, MRI Software, ResMan, Buildium, Yardi Systems, RealPage and Entrata, and states that 99% of its customers integrated it with their existing PMS (happy.co), which covers essentially every stack a student housing operator actually runs.
The capture layer is strong too. The Happy Property Maintenance app records photos on every inspection item, and it also supports optional video clips of up to one minute per clip, which we audited in detail in does HappyCo support video inspections. Our broader assessment of the platform is in the HappyCo inspections review. The short version: if you run enterprise multifamily or student housing, this is a serious operations platform and the turn board is one of the reasons.
None of that is the thing this article is about. The turn board's job is to organize the work and prove it happened. It does that job. The question worth asking is a different one: what happens to the photographic record after it lands there?
Most turn photos are never looked at twice
This is not a criticism of any platform. It is arithmetic. A 700-bed community running a two-week turn with a handful of items photographed per bed produces a photo count in the thousands. The people who could review them are the same people scheduling vendors, chasing punch items and standing at the door with keys. So in practice, three things happen to a turn photo:
- It is captured and filed. The photo exists, attached to the right unit and the right item, timestamped. It is a record. Nobody reads it.
- It is glanced at in the moment. A technician sees the room while photographing it and notes what stands out. Anything subtle, or anything that only reads as damage next to how the unit looked at move-in, does not stand out.
- It is pulled up later, once, under pressure. A resident or guarantor disputes a charge in September, and somebody goes hunting for the before and after pair. That is the only deep read most photos ever get, and it happens after the money decision was already made.
The gap is not in the capture and it is not in the board. It is that the volume of a two-week turn exceeds any human review capacity, and nobody has ever pretended otherwise. That is precisely the shape of problem worth pointing a model at.
What AI review adds to photos already in the workflow
RapidEye reads the photos operators already capture during turn and compares each unit against its own move-in baseline rather than against a generic idea of a clean room. That baseline comparison is the whole difference. A photo of a scuffed wall is ambiguous on its own; the same photo next to the same wall at move-in is either a charge or it is not.
| What the operator needs | Where it comes from | What it changes during turn |
|---|---|---|
| Damage attribution | Every turn photo compared to that unit's move-in baseline | Findings become "this changed during this tenancy" rather than "this room has a mark on the wall," which is the form a charge has to take. |
| Vendor verification | After photos read against the scope the invoice covers | Paint, clean and carpet work billed across hundreds of beds gets checked as it is submitted rather than spot-checked on the units a GM happens to walk. |
| Dispute-ready evidence | Before and after pairs tied to unit and date | A September dispute is answered from an assembled record instead of a photo hunt, which is where recoverable charges usually get abandoned. |
The recovery side of this is where student housing differs from conventional multifamily, because charges are often split across roommates and backed by parental guarantors who scrutinize them line by line. We pulled that apart separately in our student housing damage charge statistics.
Why this fits the way student housing teams already work
The constraint during the first two weeks of August is not software. It is that nobody has an hour to learn anything. Any change that requires retraining seasonal staff, rewriting inspection templates or asking vendors to submit documentation differently will not survive contact with turn.
So the useful version of AI review is the one that changes nothing upstream. Technicians keep using the same app and the same templates. Vendors keep submitting what they submit. The turn board keeps being the board the GM stares at. The photos flow the way they already flow, and the analysis happens on top of them. An operator's first turn with AI review should look, from the field, exactly like their last one.
The broader case for the vertical, including how turn economics differ from conventional multifamily, is on our student housing page.
Quick FAQ
Does AI review replace the HappyCo turn board?
No. The turn board stays the system of record. HappyCo organizes the inspections, the unit statuses, the work orders and the sign-offs, and that is where the turn is run from. AI review reads the photos that workflow already produces and returns findings against each unit's own move-in baseline. Teams and vendors keep working exactly as they do today.
How many turn photos actually get reviewed by a human?
Far fewer than are captured. Student housing turn concentrates the work into a very short window: Education Realty Trust's SEC filings describe substantially all leases commencing mid-August and terminating the last day of July, with units prepared during the first two weeks of August. When several thousand beds are photographed in that window, the photos are captured for the record rather than examined one by one, and the review that does happen concentrates on units that someone has already flagged.
What does AI review of turn photos actually produce?
Three things an operator can act on during turn: damage attribution, meaning what changed in this unit since the move-in baseline rather than what a room looks like in isolation; vendor verification, meaning whether the paint, clean or carpet a make-ready invoice covers is visible in the after photos; and dispute-ready evidence, meaning a before and after pair tied to a specific unit and date for any charge a resident or guarantor questions.
Does this require changing how technicians or vendors capture photos?
No. RapidEye analyzes the photos operators already capture in their existing workflow. Technicians keep using the same app and the same inspection templates, vendors keep submitting the same documentation, and nothing about the capture step changes. That matters most in the two weeks when there is no time to retrain anyone.
What is the cost of getting a turn unit wrong?
It depends on scope. Multi-Housing News reported basic make-ready costs of roughly $135 to $165 per bed, while Inspection Express's 2026 vendor estimate puts a full turnover at $1,500 to $3,000 per bed. A missed damage finding is either an unrecovered charge or rework inside a window where every day of delay pushes against an August move-in date.
Sources
Sources are named at the publisher level with their root domain, rather than linked or titled; every figure is verifiable at the named source.
- Annual report lease-term and unit-preparation disclosures, Education Realty Trust, filed with the U.S. Securities and Exchange Commissionsec.gov
- Student housing turnover coverage, Multi-Housing News, 2010multihousingnews.com
- Student housing preleasing report, Yardi Matrix, July 2026yardimatrix.com
- Turnover cost estimate, Inspection Express, 2026ipropertyexpress.com
- Product integrations and inspection documentation, HappyCohappy.co