01 How Ai Comps Work

Data Sources

The AI pulls from:

• MLS data — Recent sold and active listings from multiple listing services.
• Public records — County assessor data, tax records, and deed filings.
• Market feeds — Aggregated real estate data from partner sources.

These sources are combined to build a property database that the AI searches against for each comp request.

How the AI Selects Comparables

When you enter a property address, the AI analyzes the subject property's characteristics and searches for properties that match:

• Location — Properties within the same neighborhood or radius, weighted by proximity.
• Property type — Same type (single family, condo, multi-family, etc.).
• Size — Square footage range, lot size, number of beds and baths.
• Sale recency — More recent sales are weighted higher.
• Condition and attributes — Age, upgrades, and unique features.

The AI compares each potential comp against the subject property across these dimensions and calculates a similarity score.

Scoring Methodology

Each comp receives a score between 0 and 100:

• 80-100 — Strong comp: very similar property, recent sale, close proximity.
• 60-79 — Good comp: similar with minor differences in size, age, or distance.
• 40-59 — Fair comp: notable differences that require adjustment.
• Below 40 — Weak comp: significant differences, use with caution or exclude.

The score reflects overall similarity, not whether the comp supports or challenges your price. A high score simply means the properties are comparable.

What Makes a Comp "Good"

A good comparable shares these characteristics with your subject property:

• Same property type and similar square footage (within 20%).
• Located within 0.5 miles (urban) or 2 miles (suburban/rural).
• Sold within the last 6 months.
• Similar age, beds, baths, and lot size.
• Similar condition and upgrades.

When a comp matches on most of these dimensions, the AI assigns a higher score and you can use it with more confidence in your analysis.

Confidence Indicators

The results include confidence indicators:

• Number of comps found — More comps generally mean higher confidence.
• Score distribution — Tight cluster of high scores indicates a reliable set.
• Data freshness — Comp data is updated regularly. Timestamps on each result show data age.

Adjusting Results

Use the filter bar to refine results:

• Adjust distance radius if your area has limited sales.
• Expand the sale date range to include older comps in slow markets.
• Filter by property type if mixed inventory.

Next steps: - Reading a Comp Result - Multi-Agent Exit Strategies - Demographics and Market Data - How to Run Comps