• Welcome to AppraisersForum.com, the premier online  community for the discussion of real estate appraisal. Register a free account to be able to post and unlock additional forums and features.

AI Appraisal?

I use Synapse. I like it, but you have to know how to use it and when not to use it.

What worries me are these push button comp selection software products. I think this is where appraisers are going to get disciplined the most. Comp selection is one of the most important aspects of the appraisal.

a couple of AMCs have these products for non-gse work. The comps they feed you/suggest as compared to what you would have chosen from the MLS can be night and day.
 
For $30 a month Synapse is great for getting idiot AMC checklist reviewers off your back about proving your adjustments. What I do is get an MLS CSV output file of ALL sales in my extended market area over the past two year and drop the file into group one for number crunching. Then narrow the sales down to just the subject's market area over the past year, maybe two if it is and small area and drop the file into group two. Finally I make a CSV file of just the comps I would use against the subject on my grid and drop the file into group three. The resulting suggested adjustment ranges are usually large enough to drive a truck through, so I need to narrow them down and usually the number I pick in the middle is what I was going to use anyways, but now I have some fancy language for my narrative and a nice one page pdf to put in my report which keeps the idiot AMC checklist reviewers from going wild with stips. If anyone has a better way to use Synapse then please share.

Of course now AMCs are using AI to review our reports so we get more stupid auto generated stips such as:
A commercial / non residential property was detected within close proximity of the subject. Please ensure all external influences have been reported and adjusted for accordingly, if warranted. If current comparables do not share a similar external influence, then please revisit search parameters to provide a comparable with similar influence to support opinions of marketability or value impacts.
I already said in the report that it is a small town of 1500 people, the subject and all comps are in the same neighborhood with similar locations. The town is literately 10 blocks by 10 blocks! EVERYTHING is within close proximity of commercial / non residential property. A real person who isn't an idiot could read the sentence I wrote on the first line under the market grid, look at the maps I provided and pass the report. But Nooooo, now I have to deal with this AI BS making my life sooooo much easier!
 
For $30 a month Synapse is great for getting idiot AMC checklist reviewers off your back about proving your adjustments. What I do is get an MLS CSV output file of ALL sales in my extended market area over the past two year and drop the file into group one for number crunching. Then narrow the sales down to just the subject's market area over the past year, maybe two if it is and small area and drop the file into group two. Finally I make a CSV file of just the comps I would use against the subject on my grid and drop the file into group three. The resulting suggested adjustment ranges are usually large enough to drive a truck through, so I need to narrow them down and usually the number I pick in the middle is what I was going to use anyways, but now I have some fancy language for my narrative and a nice one page pdf to put in my report which keeps the idiot AMC checklist reviewers from going wild with stips. If anyone has a better way to use Synapse then please share.

Of course now AMCs are using AI to review our reports so we get more stupid auto generated stips such as:
A commercial / non residential property was detected within close proximity of the subject. Please ensure all external influences have been reported and adjusted for accordingly, if warranted. If current comparables do not share a similar external influence, then please revisit search parameters to provide a comparable with similar influence to support opinions of marketability or value impacts.
I already said in the report that it is a small town of 1500 people, the subject and all comps are in the same neighborhood with similar locations. The town is literately 10 blocks by 10 blocks! EVERYTHING is within close proximity of commercial / non residential property. A real person who isn't an idiot could read the sentence I wrote on the first line under the market grid, look at the maps I provided and pass the report. But Nooooo, now I have to deal with this AI BS making my life sooooo much easier!
100% agree.

The issue is that these churn and burn appraisers or these ignorant appraisers do not use the "constrained" feature. It also has a red line if you hoover your mouse over that shows you the error, standard deviation, etc.

It makes them look silly in their reports when the have a crazy wide range for a feature.

Wait, don't flog me...your experience...a half bath could have range from 0-$30k? No! Use some common sense.

For high end custom built homes in urban areas....I don't think so.

It best works in suburban areas of homogeneous subdivisions.

By far, the GLA adjustment is the most accurate.
 
Yes, the initial output may show a stupid range of negative $12,000 to positive $45,000 for an adjustment of an item. There is a scissor tool that you can use to trim the outliers down to a reasonable range of, for example, +$5,000 to +$15,000 and you end up with the +$10,000 that you were going to use anyway.

It looks like the appraisal report from the original post did not trim the stupid outliers from the output.
 
The GSEs have the totally logical opinion that this AI stuff on a macro level is no more inaccurate than the current system.
 
Find a Real Estate Appraiser - Enter Zip Code

Copyright © 2000-, AppraisersForum.com, All Rights Reserved
AppraisersForum.com is proudly hosted by the folks at
AppraiserSites.com
Back
Top