Real estate is one of the largest asset classes in the world, and every transaction involves data: prices, locations, features, financing and risk. AI is increasingly used to value properties, market them, match buyers and tenants, and manage buildings.

Automated valuations

Automated valuation models (AVMs) estimate property values using recent sales, property characteristics, location and market trends. Consumer-facing estimates such as Zillow’s Zestimate made AVMs familiar to millions of homeowners. Lenders also use AVMs to support mortgage decisions, subject to regulatory standards.

AVMs work best where many similar homes sell often, and worst for unusual properties or thin markets.

A cautionary example: in 2021, Zillow shut down its home-flipping business, Zillow Offers, which bought homes based partly on algorithmic pricing, after it ended up overpaying in a fast-changing market and reported large losses. Even good models can fail when markets shift quickly.

Listings and marketing

  • AI drafts property descriptions from features and photos.
  • Virtual staging adds furniture to photos of empty rooms.
  • AI enhances photos and creates floor plans and virtual tours.

Listings must stay accurate: virtually staged photos should be disclosed, and AI shouldn’t invent features. Misleading listings can break advertising and consumer protection rules.

Search and matching

Property portals use AI for natural-language search (“three-bedroom near good schools with a garden”), recommendations and price alerts.

Tenant screening and lending

AI tools score rental applicants and borrowers. These uses carry legal risks: in the US, the Fair Housing Act prohibits discrimination in housing based on protected characteristics, and US housing authorities have warned that tenant screening algorithms and targeted housing ads can produce discriminatory outcomes. Data such as eviction or criminal records can be inaccurate and have disparate impacts.

Property management and investment

  • predicting maintenance needs;
  • optimizing energy use;
  • analyzing markets and forecasting rents for investors;
  • automating document processing for leases and due diligence.

Rent-setting software that uses competitors’ data has faced antitrust lawsuits in the US, alleging it helped landlords coordinate prices.

Tips for buyers and renters

  1. Treat online valuations as rough estimates; ask for comparable sales and professional appraisals.
  2. Visit properties or request unedited photos when listings are virtually staged.
  3. Ask landlords what screening tools they use and how to dispute errors.
  4. Use AI assistants to understand contracts and questions to ask, but get professional advice for major decisions.

How an automated valuation model works

An AVM typically:

  1. gathers data on the property — size, bedrooms, bathrooms, age, lot size, condition indicators;
  2. finds comparable recent sales nearby;
  3. adjusts for differences between properties and for market trends since each sale;
  4. combines several modeling approaches, from regression to machine learning, into a single estimate;
  5. produces a confidence score or range, which is often more informative than the single number.

AVMs know nothing about a renovated kitchen, a noisy neighbor or a leaking roof unless the data captures it. That is why appraisers and agents still inspect properties for most mortgage and sale decisions.

AI for agents and brokers

Real estate professionals use AI assistants to draft listings and emails, summarize disclosures and inspection reports, prepare market reports for clients, schedule viewings and answer common questions from buyers. The same responsibilities apply: listing claims must be accurate, advertising must comply with fair housing rules, and personal data must be protected.

What to watch next

  • Regulation of tenant screening algorithms.
  • Antitrust cases involving algorithmic rent-setting.
  • Disclosure rules for AI-edited listing photos.
  • Quality standards for automated valuation models in lending.

Key takeaways

  • Automated valuation models estimate prices but are weaker for unusual properties and fast-changing markets.
  • Zillow Offers closed in 2021 after algorithm-supported purchases led to losses.
  • AI drafts listings and stages photos; accuracy and disclosure are essential.
  • Tenant screening and housing ads must comply with fair housing laws.
  • Algorithmic rent-setting has faced antitrust lawsuits in the US.

Frequently asked questions

How accurate are online home value estimates?

They can be close for typical homes in active markets but can differ significantly for unusual homes. Use them as a starting point.

Generally yes, but it should be disclosed and must not misrepresent the property’s condition or features.

Can I contest a tenant screening decision?

In many jurisdictions you have rights to know what report was used and to dispute inaccurate information.