What is AI tenant renewal risk scoring? AI tenant renewal risk scoring is the use of machine learning to predict whether a commercial tenant will renew its lease, built from building operations data such as badge-in frequency, amenity usage, and service requests rather than from the lease document alone. On October 27, 2026, HqO launched REX, an agentic AI platform that productizes exactly this for office portfolios across 24 countries. That makes one question urgent for office landlords: can engagement data actually predict a renewal, and what should you do when it says no? See also our buyer's guide to AI property management platforms.
Key Takeaways
- HqO launched REX on October 27, 2026 in 24 countries after a nine month beta, scoring tenant health and lease renewal risk from building engagement data.
- The economics favor acting early: JLL analysis puts re-letting costs at roughly 2.5 times the cost of retaining a sitting office tenant.
- Badge-in data became a weaker satisfaction signal in 2026, because 89 percent of employers now mandate at least three office days per week, per CBRE.
- Backtest any renewal score against your own last 24 months of lease expirations before letting it trigger concessions or capital spend.
- False positives carry real cost: flagging a tenant who was never leaving converts a free renewal into unnecessary tenant improvement dollars.
What HqO Launched on October 27
HqO, a Boston-based real estate operating system for commercial owners and operators, announced the full launch of REX on Tuesday, October 27, 2026, opening it in 24 initial countries after nine months in beta. The name is a loose abbreviation of "real estate experience." REX is built on HqO's Digital Grid, which the company claims is the industry's first actionable portfolio-wide ontology for commercial real estate, putting building data onto one shared record.
The platform is agentic, meaning users set guardrails for what it may do on its own. Within those limits it routes service requests, automates preventive maintenance, clears work queues, answers natural-language questions, and writes reports spanning leasing, operations, and experience. "Commercial buildings have long run on dozens of systems that do not share a record," co-founder and CEO Chase Garbarino told Commercial Observer. "REX is an AI operator that makes these systems actionable." HqO also announced an AI builder group that embeds its staff inside client teams.
HqO published no accuracy rates, reference customers, or pricing tiers, which shapes how you should evaluate the renewal claim. Our analysis of Yardi Virtuoso and buying AI from your system of record covers the broader vendor question; this piece focuses on the renewal signal itself.
AI Tenant Renewal Risk Scoring Explained
A renewal risk score converts behavioral telemetry into a ranked probability that a tenant leaves at expiry. Instead of reading the rent roll, it reads the building. HqO says REX provides early signals of engagement drop-off, renewal risk, and experience gaps drawn from the shared record, so teams can intervene before a decision hardens.
The inputs are the operational exhaust an office building already generates: badge-in frequency, conference room and amenity bookings, parking utilization, tenant app engagement, service request volume and resolution satisfaction, and survey responses. Garbarino framed the logic plainly: "From a tenant health score perspective, are they engaged? Are they coming to the building?"
Critically, REX measures health "based on data and variables that you tell it are important." That configurability is both the selling point and the catch: a score whose weights you choose is a structured heuristic, not a validated model, until someone tests it against outcomes. It is also blind by design to the lease-side variables that usually dominate renewals, including in-place rent versus market, remaining term, options, and the tenant improvement package a competing landlord will offer.
Why the Retention Math Justifies an Imperfect Signal
Even a mediocre score can pay for itself, because the cost of losing an office tenant is wildly asymmetric to the cost of keeping one. A JLL analysis of tenant turnover costs works a Zurich example: a 500 square meter office at CHF 500 per square meter per year, producing CHF 250,000 of annual net rent. Retaining the sitting tenant with three months free rent and a CHF 50,000 fit-out contribution costs about CHF 90,000. Letting that tenant walk costs CHF 228,125 once you add refurbishment, vacancy during works and marketing, marketing expense, a brokerage fee of 8 to 12 percent of annual rent, and incentives for the replacement. That is more than 90 percent of a year of net rent, roughly 2.5 times the retention case.
Run that through a valuation lens. Vacancy flows straight through to net operating income, gross revenue minus operating expenses, which excludes debt service and capital expenditures. Because a cap rate is NOI divided by value, a permanent $100,000 NOI reduction removes roughly $1.43 million of value at a 7 percent cap rate. Six months of advance warning is cheap against that. CRE investors looking for hands-on AI implementation support can reach out to Avi Hacker, J.D. at The AI Consulting Network.
The 2026 Attendance Mandate Problem
Here is the caveat that should govern how you read any engagement-based score this year: in 2026, attendance is largely mandated, so badge-ins measure policy compliance rather than tenant satisfaction. CBRE's 2026 Americas Office Occupier Sentiment Survey found 89 percent of employers now require at least three in-office days per week, up from 78 percent a year earlier, while only 14 percent of organizations are pursuing meaningful workplace reinvention.
That breaks the naive inference in both directions. A tenant whose badge-ins jumped because headquarters imposed a four day policy is not a happier tenant, and may still be running a relocation study. One with flat attendance under a deliberate two day policy may be perfectly satisfied. The fix is to normalize engagement against each tenant's own attendance policy rather than a portfolio average, and to weight signals that are harder to mandate: amenity spend, voluntary event attendance, and survey trends.
The same survey shows why renewals are getting harder to read from behavior alone. Twenty-three percent of organizations say AI already influences space planning and another 30 percent expect impact within two years, while 37 percent anticipate AI-driven headcount cuts yet only 4 percent of those plan significant space reductions. Occupiers are rethinking what space does, not just how much they need, which is a decision no badge reader captures. See also our work on AI office obsolescence and the flight to quality and how corporate occupiers build elastic portfolios.
How to Validate a Renewal Risk Score Before You Act On It
Treat a vendor renewal score as a hypothesis, not a finding. Five steps separate an early warning system from an expensive dashboard.
- 1. Establish your base rate first. Calculate your actual renewal rate by asset, tenant size, and industry. If you renew 80 percent of tenants, a model that always predicts "renew" is 80 percent accurate and useless.
- 2. Backtest against your own history. Run the score against the last 24 months of expirations where you know the outcome, and measure how many departures it caught versus how many renewals it wrongly flagged.
- 3. Check the lead time. Large occupiers often begin internal relocation analysis well over a year before expiry. A score that fires 90 days out is reporting history, not predicting it.
- 4. Keep the lease math in the loop. Never let a behavioral score override in-place rent versus market, remaining term, or option structure. Engagement is one input into a mostly economic decision.
- 5. Price your false positives. Decide in advance what action each threshold triggers and who authorizes it. Handing concessions to a tenant who was never leaving is pure margin leakage.
The AI Consulting Network specializes in exactly this kind of model governance work for CRE owners.
Frequently Asked Questions
Q: What is HqO REX?
A: REX is an agentic AI platform for commercial real estate that HqO launched on October 27, 2026 in 24 countries after a nine month beta. Built on HqO's Digital Grid ontology, it scores tenant and building health, flags lease renewal risk, routes service requests, and answers natural-language questions about a portfolio.
Q: Can AI really predict whether an office tenant will renew its lease?
A: AI can rank tenants by departure risk using engagement data, which is useful for triage, but it cannot see the lease economics or boardroom decisions that usually drive the outcome. Treat it as a prioritized watchlist that earns its keep by buying lead time, not a forecast.
Q: Why are badge-in numbers a weak signal for tenant satisfaction in 2026?
A: Because attendance is now largely compulsory. CBRE found 89 percent of employers require at least three office days per week, up from 78 percent a year earlier, so rising badge-ins can reflect a corporate mandate rather than a happier tenant. Normalize against each tenant's own policy.
Q: What does it actually cost to lose an office tenant?
A: JLL's worked example puts re-letting at CHF 228,125 against CHF 90,000 to retain, on a property generating CHF 250,000 of annual net rent. Re-letting exceeded 90 percent of a year of net rent once refurbishment, vacancy, marketing, brokerage fees, and new-tenant incentives were included.
Q: Should office landlords buy a tenant experience platform just for renewal scoring?
A: Rarely. The renewal score is a derived output of operational data you should collect anyway, so justify the platform on service request routing, maintenance automation, and reporting, then treat the renewal signal as upside you validate. For personalized guidance, connect with The AI Consulting Network.