0

Technology you buy?  Or technology you sell?

EBV, Evidence Based Valuation© is applied technology.  The modern – today – version comprises AI, Data Science, and micro-economic theory.

There are two main types of artificial intelligence:

  • Generative AI – Used for new content creation, creative pursuits, and personalized experiences.
  • Analytic AI – Used for data-driven decision-making, optimization, and predictive analytics.

Generative AI is handy when you have no comps.  (Just kidding.)

Analytic AI is needed for valuation — “appraisal is substantially predictive in nature.”  The Appraisal of Real estate.

So let us consider “evidence” as it works with analytic AI.  What is evidence?

Can we consider fake evidence, real evidence, or “belief-based” evidence?

Fake evidence is false evidence.  It generally reflects some bias on the part of the purveyor.  The bias can be personal, whether conscious or unaware.  Or the bias can be analytical, whether on purpose, or from incompetence.

Real evidence can be data, if that data shows to be valid, authentic, and in context.  Now it becomes relevant information.  Real evidence can be demonstrative, like tables, graphs, maps, and photos.  It can be documentary, like public records, listings, or even other professional opinions, (if trustworthy and authentic. Or it can be testimonial evidence, of seeing or witnessing something, like inspecting or measuring a property.

Belief-based evidence is of credible opinion.  More specifically, in USPAP, “credible” means worthy of belief.  The “worthiness” part refers to the opinion of the reader/reviewer.

Given the advent of AI, we must also consider what AI, analytic AI, bases its results on.

Analytic AI can only build on data given and processes gathered, delimited by guardrails provided by the analyst.  Data given is given by the analyst.  But process and (appraisal) theory come from years of accumulated texts, courses, and journal articles.

Unfortunately, the gathered “appraisal process” knowledge is built on years of sparse and difficult data, with limited (or no) analytic power.  We can call this the “pick comps, then adjust” paradigm.

AI does not do well with the “pick comps, then adjust” model.  The model is too vague, and does not deal well with the reality of belief-based data selection, and improbable adjustment “support.”  That model, a historical legacy, worked well some 50 years ago.  It is obsolete today.

The “pick comps, then adjust” model does not give AI the clear-cut algorithms it needs.  It does not comport AI to the sharp and clean appraisal-expert judgment it needs.

EBV©,  Evidence Based Valuation, provides the crisp (and simpler) process AI needs.  The team of AI and the appraiser/analyst can only work with a clear path.

“Trust me, I know a good comp when I see one” doesn’t work.  The future is the AI/Appraiser EBV team!

Public policy and professional survival demand nothing less.