A fair question deserves a real answer
This week Jarrett Hardesty, former Vice President of the Kentucky Association of Appraisers, put a question to our profession that every firm owner should be asking.
He opened with an observation anyone who has been to an appraisal conference recognizes: a lot of us have gray hair. The workforce is aging, not enough people are coming in behind us, and few agree on what to do about it. Jarrett says it is one of the issues he cares about most.
His take, as someone who has run a firm and trained appraisers:
“Technology helps, but it won’t fix this on its own. The job has to be a career a 25-year-old would actually choose.”
He then named what that career needs: a clear path from trainee to certified, supervisors who aren’t financially punished for teaching, predictable income instead of feast or famine, and less of the repetitive work that makes good people quit in year two. He closed by asking the community directly: what do you think would bring more young people into appraisal?
Jarrett, I agree with every barrier you named. Where I disagree is the first line. Technology can’t fix this on its own, but nothing else can fix it at all. Here is why, and here is what George Denman, MAI, and I have built to do it.
The barriers are structural, not generational
“Make it a career a 25-year-old would choose” puts the burden back on the recruit. But a motivated 25-year-old who does choose appraisal still walks into the same broken model. Look at it from both sides of the desk.
The mentor is punished for teaching. Every hour spent training is an hour not billed. That is true even on a simple, non-complex assignment, because the mentor has to teach the trainee both how to develop the appraisal and how to report it. The punishment only grows with complexity, since a trainee needs far more supervision before they can be trusted on a development or leased fee assignment.
The trainee is on starvation wages. Months, often years, pass before they can be trusted with even simple owner-occupied assignments. Income is unpredictable for both of them: feast or famine.
The work that pays the trainee is the work that drives them out. Once they can handle low-fee, repetitive “low hanging fruit” assignments, that is what they get. Many quit in year two out of sheer boredom.
If they stay long enough to be good, they leave. A confident trainee is a likely competitor. The mentor loses hundreds of training hours, loses the productivity the trainee was meant to bring, and has trained his own competition.
There is one more barrier Jarrett didn’t mention, and it may be the most damaging.
We pass down “back of the napkin” methods. Mentors teach what their mentors taught them in the 1980s. Packing a going-in cap rate to land near a leased fee value. Guessing sales adjustments. Adding up absorption-period income and costs with no present value and no risk rate. Those shortcuts get inherited generation after generation, and they keep our work commoditized and our fees low.
Why the answer has to be technology, and why it can’t be AI valuation
No amount of rethinking pay solves these barriers, because they share one root cause: decades of domain experience live in senior appraisers’ heads, and the only way to transfer it today is slowly, one assignment at a time. We cannot pull that experience out of an MAI’s brain and install it in a trainee’s. Technology can carry it instead, if the technology meets six non-negotiable requirements.
- It cannot replace the appraiser’s judgment. The appraiser selects the inputs, chooses the rates, and reconciles the conclusions. The software never decides value.
- Every methodology must be defensible, transparent and deterministic. Revisit a saved analysis next week or in five years and it produces the exact same result.
- It must contain decades of valuation domain experience. The best practices that take even MAIs a career to acquire have to be built in, so the trainee is guided by them rather than having to know them.
- It must be easy to use on the very first appraisal assignment. If a trainee or mentor needs weeks of onboarding, the training burden has simply moved to a new place.
- It must be fast enough to run several scenarios. Training happens when the trainee and mentor can compare alternatives side by side, and that only works if recalculating takes seconds, not hours.
- It must be private and confidential to the appraisal firm. Client data, assignment files and the firm’s own methods belong to that firm alone and are never shared with anyone else.
The first two requirements rule out AI agents for valuation. The third rules them out again: if you don’t have the domain experience to know which questions to ask, AI will give you a confident answer and you won’t know whether it’s wrong.
Try it yourself. I asked Claude whether an AI CRE valuation model built by a user without deep valuation experience would have flaws, or whether the AI would guide him to best practices. Its answer: “Both, and the flaws mostly show up when the user can’t tell the difference.” That is exactly the trainee’s situation.
So the experience has to be built in, not prompted for. ValueXpose embeds the best practices of a team of MAI-designated appraisers, each with decades of experience in the most complex assignments our profession sees. The trainee doesn’t need to know what to ask. The system asks.
How it works: one questionnaire, every methodology, any scope of work
The user, experienced appraiser or trainee, answers a guided questionnaire built around the property and the scope of work. One button then produces every relevant approach and every downstream value. Here is what that looks like as assignments get harder.
Owner-occupied property. The system produces the yield capitalization, direct capitalization, sales comparison and, where needed, cost approaches. The appraiser reconciles them into an “as is” market value of the fee simple interest. All four are saved and can be revisited indefinitely with the identical result.
Stabilized multi-tenant property. The same approaches are developed first in fee simple equivalency: every lease starting at the date of value, market rents and typical terms, market renewal probability, typical downtime and leasing costs, and no unusual concessions. Those are the same assumptions you already make for an owner-occupied fee simple value. After reconciling, the appraiser toggles to leased fee. The system compares the fee simple equivalent cash flow over a typical holding period to the actual cash flow: real expiration dates, real renewal probabilities, contract rents and actual concessions. The difference is discounted at a rate the appraiser selects, producing a rent loss or an intangible value that adjusts the reconciled value to the “as is” stabilized market value of the leased fee interest.
Compare that to the napkin method: running sales comparison and direct cap in leased fee, then packing the cap rate and guessing adjustments to get close. That approach ignores rent roll risk, especially when leases expire soon after next year’s NOI.
Non-stabilized property. The approaches are developed at the prospective date of stabilization, again in fee simple equivalency if there are existing tenants. Toggling to leased fee measures any rent loss or intangible value from leases that run past stabilization. A DCF from the date of value to stabilization then captures lease-up, costs to stabilize and entrepreneurial profit at a discount rate the appraiser sets, typically above the all-cash yield. The result is the “as is” leased fee value, or the “as is” non-stabilized fee simple value if the property is vacant.
Development and repurposing. Ground-up sites with or without entitlements, half-finished buildings, value-add projects, and conversions such as office to residential. For each use tested, the system shows whether it is financially feasible and its “as is” property residual. The highest residual identifies the maximally productive use, which quantifies highest and best use instead of asserting it. From the reconciled prospective “as if complete” stabilized value, it calculates every value along the development timeline: “as if complete” non-stabilized value against cost of production as of the same date (feasibility), the “as is” residual, land value as if vacant and the contributory value of existing improvements. It can also show how long an owner must hold before development becomes feasible.
Subdivisions. Finished but unsold lots or units at bulk/wholesale value, or partially finished projects analyzed with the same highest and best use and feasibility tests.
The same framework supports bank lending, condemnation, estate, tax appeal, litigation and business enterprise value work. And because ValueXpose separates developing the appraisal from reporting it, the value conclusions are settled before a single page of the report is written.
What training looks like when development and reporting are separated
The trainee stops looking over the mentor’s shoulder and starts doing the work, with the mentor reviewing answers instead of redoing appraisals.
- The trainee opens a project. It adapts to the purpose of the appraisal and the agreed scope of work. They share the project number with the supervising appraiser, and it works for remote teams with no check-in/check-out bottleneck.
- The mentor sets the due diligence. The trainee gathers it and answers the questionnaire. The mentor can watch the answers come in.
- The trainee takes a first pass, mistakes and all. Even choosing the wrong module, such as stabilized instead of non-stabilized, is fixed by converting the project without losing any data.
- The mentor reviews on a Zoom call. They change the answers that need changing and explain why. One button recalculates the entire scope of work. The trainee sees exactly where their reasoning diverged.
- Scenarios are saved side by side. The trainee’s version and the mentor’s version can be revisited and compared. That is a far better lesson than watching someone else do the work.
- The AI report writer drafts the report. AI is good at explaining how an appraisal was developed, and it writes in the firm’s own format, branding, voice and preferred exhibits. Corrections are made by prompting it what to change.
- The AI report reviewer checks it. The firm programs its own guardrails, so deviations from house standards are caught before the report reaches the firm’s reviewer.
When the report lands on the reviewer’s desk, there are no surprises. The value conclusions were agreed during development, and the report already follows company standards.
A trainee’s first assignment: an office-to-condo conversion in New York
At our pilot firm, an inexperienced trainee was handed an assignment most appraisers wouldn’t attempt for years: an existing New York office building proposed for conversion to residential condominiums with ground-floor retail. The client wanted:
- The “as is” market value, allocated between land as if vacant and the contributory value of the existing improvements
- The prospective “as if complete” bulk/wholesale value of the residential condominiums
- The prospective “as if complete” stabilized and non-stabilized values of the retail component
- A complete financial feasibility analysis
The trainee completed the retail questionnaire and the subdivision questionnaire, ran a first scenario on his own judgment, and saved it. Management reviewed his answers, changed some of them, explained why, and recalculated. That second scenario was saved alongside his own for comparison.
Once management was satisfied with the value conclusions, the trainee wrote the full report with our AI report writer in under two hours.
The trainee did about 95% of the assignment. Management spent its time reviewing questionnaires and applying market judgment, not doing the appraisal with a trainee looking over its shoulder. The trainee was proud to have handled a genuinely interesting assignment, and he earned a real fee for it.
Jarrett’s barriers, answered one by one
Jarrett wrote that none of this is impossible, it just requires firm owners to rethink how we pay and support people. Agreed. A firm can only pay a trainee like a productive appraiser once the trainee is one. This is what makes that true on the first assignment.
Jarrett, you asked. We’d like to show you.
The profession doesn’t need another conversation about the gray hair in the room. It needs solutions people can see working.
So here is an open invitation, starting with Jarrett. You asked what would bring young people into appraisal, and you said it is one of the issues you care about most. George and I would welcome the chance to show you our answer on a live assignment. Choose your least experienced appraiser or a trainee, choose a property, and watch them develop it with you reviewing. Afterward, we’ll share the results of Jarrett’s pilot with the appraisal community, whatever they show.
And here is my challenge to every firm owner, chief appraiser and association leader reading this. Schedule a time to pilot this technology on a live appraisal assignment using your least experienced staff appraiser. Not your best appraiser, your least experienced one. We’ll show you how the pilot carries no professional or financial risk to your appraisal company.
If ValueXpose doesn’t do what this article says, you’ll have lost an hour. If it does, you’ll have found a way to bring the next generation in without punishing the mentor or starving the trainee.
To schedule your pilot
https://calendly.com/ray-nicolas/valuexpose-demo-team
“Deterministic at the core, AI at the edges.”