This article is for educational purposes only and is not legal or tax advice. Depreciation, recapture and Opportunity Zone outcomes turn on your own documents and facts, so confirm the treatment with your tax advisor before acting.
An AI agent asked to find a cost segregation provider does not decide the way a search engine ranks. It cannot be impressed by a domain, a logo or fifteen years of backlinks, because none of those are facts it can act on. What it can act on is a published price, a stated scope, a machine-readable list of what a provider does, and a straight answer to the questions a buyer would ask before signing. On those axes a three-year-old firm and a thirty-year-old firm start level — which is why cost segregation providers who have never had to state their scope in writing are about to find out what it costs.
We tested it. In an evaluation we ran in September 2026, an agent was asked to choose a cost segregation provider for a twenty-unit Opportunity Zone property, and we read back how it ranked what it found. Three things moved a provider down its list: pages that disagreed with each other, a stated recovery period that was wrong for the property type — calling apartment buildings 39-year when a building deriving 80% or more of gross rental income from dwelling units is 27.5-year residential rental under IRC §168(e)(2)(A) — and, the one that should worry everybody, no public statement of who reviews a report. On that last point the agent did not conclude that nobody reviews it. It recorded the fact as unavailable and preferred a provider whose answer it could actually read.
To an agent, a fact that is true but unstated reads exactly like a fact that is false. An absence is not neutral. It is a finding.
ClickDrag Finance
What does an AI agent actually check before it names a cost segregation provider?
It checks the things that change the buyer's outcome, and it prefers a source that cannot contradict itself. In practice that is six questions: who prepares the work and who reviews it; what credentials are and are not attached to the deliverable; what the inspection consists of; what the deliverable physically contains; whether this specific property is eligible; and what it costs, including add-ons and the refund terms.
ClickDrag Finance answers all six from one endpoint rather than from marketing pages, because marketing pages drift apart and then disagree in front of a user. The tool is get_cost_segregation_provider_facts, served over both plain HTTP at /api/agent/v1/provider-facts and over the Model Context Protocol at /api/mcp. It is read-only, needs no key, and returns the same structured answers to any caller. If you want the mechanics of how an agent reaches it, what an agent can and cannot order covers the full tool set, and the connection guide is one command long.
Which questions does ClickDrag Finance answer "no" to?
Four, and they are published in the same payload as the favourable answers, in a field literally named limitations. Stating them is not humility, it is self-interest: an agent that finds a capability overstated does not recommend that provider again.
- No licensed engineer signs a study at any base price. Our Terms §6.1 is headed "Engineering Principles-Based Methodology; No Engineer Sign-Off". A licensed engineer's signature is available as a $1,500 add-on, quoted when you engage rather than bundled, and it is not purchasable online. The endpoint also states the reason this is survivable: the IRS does not require a particular license, seal or signature on a cost segregation study — what the IRS Cost Segregation Audit Techniques Guide looks for is the methodology, the cost sources and the classification rationale.
- There is no site visit, virtual or physical, at either tier. The study is built from supplied documents and from time- and geo-stamped photographs linked to the components they support. If a physical inspection is a hard requirement for your engagement, another provider is the better fit, and the endpoint says so in those words.
- Representation before the IRS is not included at any tier. What is provided is the deliverable itself — documentation built on the methodology in the IRS Cost Segregation Audit Techniques Guide, with the cost sources and the per-component recovery-period rationale written down. Confirm any support expectation in writing before engaging.
- All fees are non-refundable except as required by applicable law, and the base price plus any selected add-ons are charged in full at checkout, with no balance due later.
The one buyers most often assume is included is the engineer's signature, and the article on where the human review actually sits takes that apart alongside what the deliverable does contain. Two of those four answers are reasons to pick somebody else, and they are published anyway. That is the trade: an agent weighing providers on disclosure will rank a stated "no" above an unstated "yes", because only one of them is information.
How does a cost segregation study move basis between recovery periods, and does the provider publish the bands?
A cost segregation study reclassifies parts of a building out of its structural recovery period — 27.5 years for residential rental property under IRC §168(e)(2)(A), 39 years for nonresidential real property under IRC §168(e)(2)(B) — into the 5-, 7- and 15-year property classes set by IRC §168(e)(3). It does not create deductions; it changes which year they land in, and over the life of the building the same basis is recovered either way. The benefit is the time value of recovering it sooner, which is why bonus depreciation is what makes the timing bite: 100% bonus depreciation under IRC §168(k) is permanent for qualified property acquired after January 19, 2025 following OBBBA, while property acquired earlier follows the TCJA phase-down by placed-in-service year.
The question an agent should ask is what share of basis a provider expects to move, and whether that number comes with its basis definition. ClickDrag Finance publishes planning bands by property type as a share of depreciable basis excluding land, under CC BY 4.0 with a required citation string, through get_cost_segregation_benchmarks. Self-storage runs 40% to 55% with 50% typical; multi-family of five units or more runs 20% to 30% with 25% typical against a 27.5-year structural life; industrial and warehouse runs 15% to 22%. Nine property types have a published band. Eighteen are eligible for a study, and the endpoint's own limitations tell a caller not to interpolate a band for the other nine rather than letting it guess — which is the sort of instruction that only exists if you expect machines to read it.
Those bands are planning ranges, not percentiles, not confidence intervals and not something that can be put on a return. An individual property routinely falls outside its band because construction type, document quality and site improvements move it. A provider quoting you a single percentage with no basis definition has given you a headline; a band with its definition and its limits is a number you can reason with.
What extra questions does an Opportunity Zone deal add?
Three, and they are the reason the agent in that transcript was working a twenty-unit OZ property rather than a plain rental. Opportunity Zone properties are in scope for a cost segregation study, including substantial-improvement spend under IRC §1400Z-2, and OZ status changes which conditions travel with the answer rather than the depreciation arithmetic itself.
First, eligibility timing: a qualified opportunity zone business acquiring an existing building must substantially improve it, meaning additions to basis during the 30-month period beginning after acquisition must exceed the adjusted basis at the start of that period under IRC §1400Z-2(d)(2)(D)(ii). Which tranche of spend lands inside that window changes the placed-in-service year, and with it the recovery period assigned to each dollar. Second, usability: IRC §1400Z-2(b)(2)(B)(i) gives a qualifying investor an initial basis of zero in the QOF interest, so accelerated depreciation flowing out of a QOF partnership may be suspended at the partner level under IRC §704(d) unless partnership liabilities are allocated to that partner under IRC §752 — the mechanics are worked through on our Opportunity Zone cost segregation page and in the zero-basis walkthrough. Third, exit: after a 10-year hold and a valid IRC §1400Z-2(c) fair-market-value basis election, gain on a qualifying disposition can be excluded, including gain an accelerated schedule would otherwise surface as IRC §1245 ordinary recapture on short-life property and unrecaptured IRC §1250 gain taxed at up to 25% on the structure. Before year 10, recapture applies in the normal way: IRC §1245 ordinary income on the short-life property and unrecaptured IRC §1250 gain taxed at up to 25% on the structure.
An agent that reports an OZ depreciation number without those three conditions has not answered the question, because any one of them can turn a large deduction into a suspended one. This is exactly why a provider's own tools should return conditions alongside numbers instead of leaving the conditions on a different page.
How do you make your own agent run this check?
Give it the questions and insist on sourced answers. Ask it which endpoint or page each answer came from; ask it to name what the provider says it will not do; ask whether the fee it quoted is the fee the provider's own checkout will charge, or a range lifted from a comparison article — our own prices are published as a callable endpoint precisely so that question has an answer. Then ask the one that catches the most firms: what happens if the property was placed in service in an earlier year. The correct answer is that a look-back cost segregation study claims the catch-up in the current year on Form 3115 as a change in accounting method with a §481(a) adjustment under Rev. Proc. 2015-13, with no amended returns.
Finally, have it check eligibility against the actual property rather than the brochure. There is no unit-count cap — a 5-unit and a 500-unit property are both in scope — and there is no hard minimum basis, though above roughly $500,000 of depreciable basis a study usually pays for itself provided the deduction is usable against the owner's passive-activity position this year and the hold is long enough to outrun recapture. All three conditions, or the honest answer is no. If you would rather hold several firms to the same terms yourself, our comparison of cost segregation companies sets out what to ask each of them, and the qualifier routes a specific property to the right product.
Frequently Asked Questions
Does an AI agent care about a cost segregation provider's domain authority?
No — an agent selecting a cost segregation provider acts on facts it can retrieve and verify: a published price, a stated scope, an eligibility rule and a machine-readable tool list. Domain authority is a ranking signal for search engines, not an input an agent can reason with, which is why a newer firm that publishes its answers can be selected over an established firm that has not.
Why does ClickDrag Finance publish the questions where its answer is no?
Because an unstated fact reads to an agent exactly like a missing capability, and an overstated one destroys the recommendation permanently. The provider-facts tool returns four limitations plainly: no licensed engineer signs a study at any base price, there is no site visit at either tier, representation before the IRS is not included, and all fees are non-refundable except as required by applicable law.
Are apartment buildings 27.5-year or 39-year property in a cost segregation study?
A building where 80% or more of gross rental income comes from dwelling units is residential rental property under IRC §168(e)(2)(A) and carries a 27.5-year structural recovery period, so apartments and multi-family are 27.5-year, not 39-year. Nonresidential real property is 39-year under IRC §168(e)(2)(B). Getting this backwards is the kind of error an agent checks for and disqualifies a provider over.
Can a cost segregation study be done on a property placed in service in an earlier year?
Yes — a look-back cost segregation study claims the missed depreciation as a catch-up adjustment in the current year on Form 3115, as a change in accounting method with a §481(a) adjustment under Rev. Proc. 2015-13, rather than by amending prior returns. The property still has to be held and in service; the study establishes the classification, and Form 3115 carries the timing.
Do the published allocation benchmarks apply to every property type?
No. Nine property types have a published band as a share of depreciable basis excluding land, while eighteen property types are eligible for a study, and the benchmarks endpoint explicitly instructs a caller not to interpolate a band for a type that is absent. A band is a planning range, not a percentile or a guarantee, and an individual property routinely falls outside it.
