Update: This post is a work in progress as each step takes quite some time. As of early February, I have filed a written appeal based on what ChatGPT provided, and I will update this post with the results as they come in.
In Maryland, your home is assessed every three years and your property taxes are based on that assessment. We last dealt with this in 2023 and the appeal was successful.
I challenged it based on the qualifying factor, a multiplier used in our basic assessment. The state found us to be a 6, which meant multiplying your base value by 1.37. We lowered it to a 5, which is a multiplier of 1.17. That resulted in a large saving (20%!).
This year we cannot do that because we are still at a qualifying factor of 5.
Three years ago I spent a few hours learning how it all worked, but this time I wanted to see if ChatGPT could give me a helping hand. I know the process of an appeal, but ChatGPT was able to give me insight into how it all really worked… and it delivered results. And within ten minutes I filed an appeal.
First, I requested our Housing Worksheet and Comparable Sales List from Maryland that I used for my assessment. I then entered the documents into ChatGPT and let it analyze my chances and tell me exactly what to do.
📒 If you want to dispute your taxes but don’t want to do it yourself, Ownwell may be able to help. If available in your state, they can dispute it for you and you only pay a percentage of what they save you in taxes. Nothing in advance.
Table of contents
ChatGPT has identified objection targets
The Housing Sheet gave ChatGPT enough information about our property to compare it to the sales list and see if there were any areas we could dispute.
This is what it said about comparable sales:
Your Area sales list (neighborhood XXXXXXXXX.XX) is gold
This is what immediately jumps out of that sheet:
A. Recent sales cluster lower than your estimate
A lot of Sales at arm’s length 2023–2025 in your exact neighborhood code show:
- Sales prices usually between:
$775k – $1.05 million- Several large plots (5-10 acres) to sell well below your estimated value
- Properties with:
- Similar vintage (1970s-90s)
- Similar or larger living space
- Similar area
- Often equal or better quality ratings
Yet SDAT places you on $1.3 million.
That gap is the core of your call.
For context, we live in an area where large lots are often much older, as they predate the vast amount of buildings built over the last twenty to thirty years. Most new single-family homes are now on lots of half an acre or three acres at most.
We have a house built in 1990 on 10 acres and it is only a limited number of acres because it is uneven, wooded and (fortunately for us) too expensive to subdivide and build.
So if you know the area, this is all true and the valuable argument is plausible.
ChatGPT also identified likely errors:
3. Likely assessment errors/issues
These are the arguments that usually end up:
🔹 1. Overestimated market adjustment factors
SDAT applied:
- Province multiplier: 1.60
- Quality factor: 1.17
- Neighborhood adjustment: 1.10
👉 Combined, this one dramatically increase basic costs.
Your compositions suggest market prices did NOT rise at that rate for similar homes.
ChatGPT also picked out a few line items that seemed aggressively high, such as a completed cellar valuation and pool.
ChatGPT even gave you an idea of what could be saved:
5. What a realistic victory looks like
Typical Successful Results in Howard County:
- 5-15% reduction at Supervisor level
- About your review:
- 10% = ~$136,000 discount
- That is real tax savings every year
If the supervisor review fails, you can escalate to:
- Real Estate Tax Appeals Board (PTAAB)
…but most people never need that.
Submitting a written appeal
If you don’t know how to appeal your property taxes, ChatGPT can give you step-by-step instructions.
The first step in Maryland’s appeals process is to go to the online appeal portal and submit a written appeal, known as a Supervisor’s Review.
In Maryland, it is an online form with a simple text box. ChatGPT can compose a cover letter with attachments (such as the sales list with marking to identify actual comps) if your state allows it.
For now, I simply ask ChatGPT what to paste into the form and this is what it suggested:
I am requesting a supervisory assessment because the estimated market value of my property appears to be exaggerated.
The current total appraised value is $1,300,000 ($400,000 land and $900,000 improvements). Based on recent sales within arm’s length Neighborhood code XXXXXXX.XXproperties of similar age, quality, size and living area have sold for significantly less than my estimate.
Several comparable sales from 2023–2025 in this neighborhood show market values significantly lower than my estimated value, including similarly sized homes (approximately 3,500–4,500 square feet) on large acreage (5–10+ acres). These sales suggest that the applied provincial multiplier and neighborhood adjustment exceed observed market prices.
Additionally, the contributory value assigned to certain features (finished basement and in-ground pool) appears higher than supported by market data, as comparable properties do not demonstrate proportionate price premiums for these items.
Based on these comparable sales, I believe the market value as of the appraisal date is more accurate $__________. I respectfully request a review and adjustment to bring the review in line with actual market value.
Upon request, I am happy to provide you with a detailed comparable sales overview, photos and supporting documentation.
Notice the blank: ChatGPT suggested different values based on how conservative or aggressive I wanted to be. I said aggressive and mention 3-5 compositions as examples… because you don’t get what you don’t ask for!
Here’s how ChatGPT updated the note:
I am requesting a supervisory assessment because the estimated market value of my property appears to be exaggerated.
The current total appraised value is $1,300,000 ($400,000 land and $900,000 improvements). Based on recent sales within arm’s length Neighborhood code XXXXXXX.XXproperties of similar age, quality, size and living area have sold for significantly less than my estimate.
Comparable neighborhood sales from 2023–2025 include:
• Indoor sales March 2024 for approx $775,000 (improved property by acreage)
• Indoor sales June 2024 for approx $850,000
• Indoor sales September 2024 for approx $900,000
• Indoor sales March 2025 for approx $1,050,000
• Indoor sales June 2025 for approx $1,100,000These properties are within the same neighborhood code and reflect market prices well below my estimated value, despite similar features. The applied provincial multiplier and neighborhood adjustment therefore appear to overestimate current market conditions.
Additionally, the contributory value assigned to certain features (including the finished basement and in-ground pool) appears higher than supported by market data, as comparable properties do not demonstrate proportionate price premiums for these items.
Based on these comparable sales, I believe the market value as of the appraisal date is more accurate $1,160,000and I respectfully request an adjustment to reflect true market value. Supporting comparable sales data and documentation are available upon request.
And that’s what I submitted. We’ll see what the board says!
Results of appeal [TBD]
It will take some time for Maryland to respond to my request, especially given all the other calls, but I will update this section with their response and what I do next.
If you are not already subscribed to the newsletter and would like to receive updates as they happen, please sign up below.
#ChatGPT #Dispute #Property #Taxes