In MO, property values for taxation are currently determined by County Assessors using geospatial data.
There are many algorithmic models for AI and machine learning assessment of real estate that consider a wide variety of variables.
Data quality, human oversight, and use in appropriate scenarios heavily impact AI and machine learning models’ assessment accuracy and usefulness.
Can AI be used to accurately assess home values?
County Assessors determine property values to calculate the amount of property taxes owed (MO State Tax Commission 2019). In MO, personal property values are assessed annually, while real estate property is assessed every odd year. For more information on the assessment process and the impact on property taxes for residential real estate, see the Science Note Property Taxes and Home Values.
County Assessors in MO use three approaches to determine the value of a property (MO State Tax Commission 2019):
In MO, Assessors currently use geospatial data and algorithms to determine home values for property tax assessments (personal communication Lukasek 2025). Geospatial data can include property boundaries, satellite images, topographic data, land use information, proximity to roads and other transportation hubs, distance to amenities such as schools and parks, and environmental data such as flood zones, fire risk, or proximity to water (BLM n.d., OHK n.d.). In addition, online images and street view of properties can be used for review (personal communication Lukasek 2025, Massachusetts Department of Revenue n.d.).
Algorithms and intelligent software systems have been utilized in property valuation since the 1990s (Rossini 1999, McCluskey et al. 1996). These include geographic information systems (GIS) (Faheemuddin 2025, Ostrikova & Selyutin 2024). All programs aim to integrate data from a variety of sources to produce standardized, unbiased property valuations with minimal errors.
The source of data and methodology varies among programs. Typically, they integrate information about home size, age, number of beds and baths, location, and similar home values (Root et al. 2023, Mortgage Bankers Association 2019). Some programs allow for additional data to be used in evaluations, including neighborhood characteristics such as landmarks, amenities, or lighting, or home features such as construction material, lot size, or kitchen area (Ostrikova & Selyutin 2024). Programs also vary in the speed, accuracy, and transparency of their processes and results (Root et al. 2023).
Assessment programs face several challenges. All models require a large amount of high-quality, up-to-date data (Ali et al. 2025, Faheemuddin 2025, Root et al. 2023). Typically, this results in lower valuation and tax revenue (Bollum 2021). In addition, models consistently required human oversight and training, and none of the models independent from human supervision (McCord 2022).
While GIS is already widely used, around data privacy, model interpretability, appropriate use cases, model accuracy, and transparency (Ali et al. 2025, Faheemuddin 2025, Lee et al. 2024, McCord 2022).
Some large counties have implemented AI toolsof the real estate assessment process. New York City has used AI to manage and process forms using optical character recognition (McCord 2022). This allows the city to digitize paper records quickly and uniformly, increasing access to information for analysis and processing. OCR still requires some human oversight but has streamlined processes for senior citizen tax exemption applications. This type of AI can be used to digitize permitting, land record, and property information that is used for assessments.
Figure 1. Process for assessing home value using algorithmic models. Considerations for choosing the algorithms used in assessing the value of properties. Many algorithms exist which utilize artificial intelligence (AI), machine learning (ML), and geographic information systems (GIS), along with statistical frameworks to determine property values using a variety of data. Data quality and human oversight have the greatest impact on assessment accuracy.
References
Ali W, Samarasinghe DAS, Feng Z, Rotimi JOB (2025) Assessing AI Techniques for Precision in Property Valuation: A Systematic Review of the Four Valuation Methods. CIB Conferences. 1(322). https://doi.org/10.7771/3067-4883.1790
Bollum T (2021) Public revenue leakage from real estate non-disclosure laws. Thesis. Montana State University - Bozeman, College of Agriculture. https://scholarworks.montana.edu/items/d540ff30-25b7-4025-aecd-2192b9f53f58
Bureau of Land Management (BLM) (n.d.) Geospatial Business Platform. https://gbp-blm-egis.hub.arcgis.com/
Faheemuddin S (2025) Exploring the Role of Artificial Intelligence in Predicting Property Value Trends: A Systematic Review of Machine Learning Applications in Real Estate Pricing and Risk Assessment. American Journal of Multidisciplinary Research and Innovation. 4(5): 2832-4854. https://doi.org/10.54536/ajmri.v4i5.4815
Lee H, Han H, Pettit C, Gao Q, Shi V (2024) Machine learning approach to residential valuation: a convolutional neural network model for geographic variation. Annals of Regional Science. 72: 579-599. https://doi.org/10.1007/s00168-023-01212-7
Lukasek X (2025) Constituent Services Manager. St. Louis County Assessor’s Office. Personal communication.
Massachusetts Department of Revenue (n.d.) Guidelines on CAMA System Acquisitions. https://www.mass.gov/doc/guidelines-on-cama-system-acquisitions/download
McCluskey W, Dyson K, McFall D, Singh S (1996) Mass appraisal for property taxation: an artificial intelligence approach. Australian Land Economics Review. 2(1). https://wrap.warwick.ac.uk/id/eprint/60982/
McCord M (2022) A Review of the methods, applications, and challenges for adopting artificial intelligence in the property assessment office. Belfast School of Architecture, Ulster University. https://pure.ulster.ac.uk/ws/portalfiles/portal/101555812/Review_of_AI_in_Property_Assessment_v2_1_.pdf
MO State Tax Commission (2019) Property Reassessment and Taxation. https://stc.mo.gov/wp-content/uploads/sites/5/2019/05/Property-Reassessment-Pamphlet-5-23-19.pdf
Mortgage Bankers Association (2019) The State of Automated Valuation Models in the Age of Big Data. Valuation Analytics Working Group. https://www.mba.org/docs/default-source/uploadedfiles/member-white-papers/stateofautomatedvaluationmodels-final.pdf?sfvrsn=6cd37b1_0
OHK (n.d.) Case Study: Using Geographic Information Systems (GIS) to Map Property Values and Assess Property Taxes. https://ohkconsultants.com/case-study-gis-to-value-property-and-taxes-part1
Ostrikova A, Selyutin V (2024) Machine Learning for Mass Valuation of Residential Real Estate. Advances in Information and Communication. Proceedings of the 2024 Future of Information and Communication Conference (FICC). 1: 570-578. https://link.springer.com/chapter/10.1007/978-3-031-53960-2_37
Root TH, Strader TJ, Huang YU (2023) A Review of Machine Learning Approaches for Real Estate Valuation. Journal of the Midwest Association for Information Systems. 2(2): 9-28. https://jmwais.org/wp-content/uploads/sites/8/2023/07/V2023.I2.A2.pdf
Rossini P (1999) Accuracy Issues for Automated and Artificial Intelligence Residential Valuation Systems. International Real Estate Society Conference 1999. https://www.prres.org/uploads/460/1044/Rossini_Accuracy_Issues_For_Automated_and_Artificial_Intelligent_Residential_Valuation_Systems.pdf
Sevgen SC and Tanrivermis Y (2024) Comparison of Machine Learning Algorithms for Mass Appraisal of Real Estate Data. Real Estate Management and Valuation. 32(2). https://doi.org/10.2478/remav-2024-0019
