Can AI generate a reliable cost segregation report?

AI is useful for some tasks around a study. An owner might use it to draft a document request, organize questions for a CPA, or summarize a source that the owner then checks. A provider might use software to make an intake easier to complete.

Those tasks do not make the model the report authority.

A general chatbot should not decide the land allocation, establish depreciable basis, estimate component quantities from missing facts, classify property for tax purposes, or complete the depreciation schedules. Those decisions depend on records, physical details, tax authority, and a method that remains available for review.

The boundary matters more than the label. Calling a product “AI-powered” says little about whether its output is supportable. The useful questions concern what the model does, which facts it receives, how its output is checked, and whether the report stands on evidence outside the generated text.

Language models generate likely responses based on their training and the information available during a request. Depending on the tool, that information may include retrieved web pages, uploaded documents, and instructions from the user. Access to more material can help, but it does not guarantee that the answer fits a particular property. OpenAI explains that hallucinations remain a problem even as models become more capable.

A peer-reviewed Nature study on semantic entropy also describes plausible falsehoods and unsubstantiated answers. The researchers developed a way to detect one class of errors known as confabulations. They did not claim that the method guarantees factual output or catches every kind of mistake.

That distinction is important in cost segregation. A model can repeat a real tax citation and still apply it to the wrong asset. It can perform correct arithmetic using an unsupported land value. It can produce a detailed component schedule from facts that the owner never supplied.

A clean PDF can still be a bad study.

Suppose a rental-property owner gives a chatbot the purchase price, address, square footage, and several exterior photographs. The owner asks for a complete cost segregation report.

The model assumes that land represents 20 percent of the purchase price. It applies a generic percentage to shorter-life assets, estimates component costs from broad construction averages, and produces depreciation schedules with citations to IRS publications.

The numbers add up. The report looks finished. It may still fail for several concrete reasons.

• The land allocation did not come from the filed depreciation schedule, assessor ratio, appraisal, or another documented source.

• The starting basis may omit capitalized acquisition costs, later improvements, prior depreciation, or facts that change basis.

• Exterior photographs do not establish concealed systems, interior finishes, quantities, materials, or installation details.

• A general citation does not prove that a particular component received the correct tax classification.

• Generic percentages do not explain how the component costs were measured or estimated for that property.

• The allocated costs may not reconcile to the owner's actual depreciable basis and renovation records.

• No identified preparer takes responsibility for the methodology, assumptions, or final report.

The problem is not the model's writing style. The problem is that the report has no dependable path from the owner's records to the final schedules.

The IRS Cost Segregation Audit Technique Guide, Publication 5653 describes principal elements of a quality study and report. Its guidance covers preparer expertise, methodology, supporting documents, legal analysis, unit costs, engineering take-offs, cost reconciliation, assumptions, certification, and exhibits.

A property owner or CPA should be able to trace the main parts of the analysis.

Not every report will present these items in the same format. The underlying work should still be visible. A reviewer should not have to trust a paragraph simply because it sounds certain.

Saving the prompt and chat history may show what the user asked. It does not establish that the inputs were complete or that the answer was correct.

Useful workpapers connect the report to records outside the model. Those records may include the closing statement, existing depreciation schedule, assessor record, appraisal, invoices, photographs, measurements, and renovation documents. The right set depends on the property and the scope of the study.

Owners who need help with the starting number can review how to determine depreciable basis. Owners reporting later work can use the renovation documentation guide to organize dates, costs, invoices, and category details.

Use this checklist whether the report comes from a traditional firm, a software-based service, or a provider that advertises AI.

• Does the report identify the property facts and records used in the analysis?

• Does the depreciable basis reconcile to the owner's records and separate nondepreciable land?

• Does the report explain how component quantities and costs were determined?

• Are material estimates and assumptions easy to find?

• Does each significant classification have support tied to the actual component?

• Do the allocated costs reconcile to the total cost being studied?

• Does the report identify the preparer and the responsibility behind the work?

• Can a CPA follow the schedules without reconstructing the analysis from a chatbot conversation?

• Is there a clear process for answering questions or correcting a confirmed error?

A report that cannot answer these questions needs more support, no matter how professional it looks.

Basis does not use AI to calculate or complete the cost segregation report.

The practical boundary is simple. Property facts, cost calculations, tax classifications, and the completed study must stand on evidence and responsible review, not generated text.

The report also does not decide how a taxpayer files a return. The owner and tax professional must evaluate how the depreciation schedules fit the return, elections, timing, and broader tax situation.

If you are still deciding whether a study makes financial sense, start with the cost segregation eligibility guide. A planning estimate can help with that decision. It is not a replacement for the records and analysis behind a completed report.

Before paying for a study, compare the provider's documentation, cost method, assumptions, classification support, and review process. Ask for clear answers when any part of the analysis is vague.

You can also start with a free Basis property estimate to decide whether a full study is worth considering. There is no need to order a report before the expected benefit and study cost make sense for the property.

© 2026 Basis Cost Seg. For informational purposes only and not tax or legal advice. Depreciation treatment depends on the property and the taxpayer. Consult a qualified tax professional regarding your situation.