AI Data Verification Layer for Property Data: Rules First, a Person on Borderline Cases
Big Data •Data Analytics •Machine Learning •Real Estate

AI Data Verification Layer for Property Data: Rules First, a Person on Borderline Cases

A property data provider working in 20+ countries collects listings from many portals, normalizes them into one schema and merges duplicates. Checking that work means reconciling records by hand, field by field, so only a small share is ever checked. The solution is a verification layer beside the pipeline that takes the routine reconciliation off people and leaves every judgement call with them.

Facing a similar challenge? Tell us about the task, and we will show how we would build it for your business.

Build Something Like This for Me

Free 30-min call · One-page solution outline within three business days · No commitment

Reconciliation by Hand

The company collects listings from the largest real estate portals in each market and sells the combined data on. The pipeline has three steps. Collection brings in raw pages. Normalization maps each one onto a standard schema: property type, address, postcode, rooms, area, price. Deduplication merges the records that describe the same property, and the same agent, across portals.

Checking that work is reconciliation by hand: open the record, open the source, compare field by field, move on. It reaches only a small share of the data. Hard cases are settled by whoever happens to review them, and the reasoning leaves with that person.

Shortcuts That Cost Trust

  • Handing the whole check to a model. It is faster, but the borderline cases get settled by a machine nobody can question.
  • Putting a model inside the matcher. Every merge has to be reproducible and explainable to a customer. A model that can answer differently on the same input breaks that promise.
  • Routing on the model's opinion of itself. A model reports its own confidence just as firmly when it is wrong, so routing on that number lets mistakes through.
  • Re-visiting every page. Opening the live listing again for every record is slow and expensive. Verification that costs as much as collection will not be run often enough.
  • A verdict with no trail. "Checked, correct" with nothing behind it cannot be confirmed or overturned.

Three Levels of Checking

The layer sits beside the pipeline: the crawlers, the normalization and the matcher stay as they are. It reads their output and checks every record, and each check stops at the cheapest level that can resolve it.

  1. Rules. Deterministic checks compare each extracted value with the raw page stored at collection: exact and normalized matches, units, ranges, references. There is a versioned rule pack per portal and country, and what the rules settle never reaches a model.
  2. Model. Only what rules cannot settle. The relevant fragment of the page is selected first, and the model returns a verdict, a corrected value and the exact source text it relied on.
  3. Routing. Evidence decides what goes to a person: whether the cited text is really in the source, whether repeated runs agree, whether two different model families agree, whether the value departs from that portal's usual pattern. Where the signals disagree, the case is borderline by definition.
  4. Reviewers. Borderline cases reach people in a queue ranked by disagreement and field impact, along with a blind random sample that keeps the automatic levels honest.
  5. Trail. Every verdict is stored with what was compared, the source fragment, the deciding level and the reason.

Drift monitoring runs alongside, per source and with no model calls, so a portal redesign is flagged before the data goes wrong.

Technology Stack

Large language modelsText embeddingsPerceptual image hashingStatistical drift detectionDeclarative data validation

Who Decides What

Routine comparisons belong to the layer; judgment calls belong to reviewers.

  • Borderline and complex cases always go to a person. However confident the automation looks, it leaves them to a reviewer.
  • Any case can be checked by hand. A person can open any record, escalated or not, see the reasoning in one place, and confirm or overturn it.
  • The model verifies and the matcher merges. The matcher stays deterministic. A contradiction between portals goes to a person instead of being silently corrected.
  • People are measured too. The reference set uses two independent reviewers per record and a third to settle disagreements.

Coverage Targets

What follows is the coverage the layer is built to provide.

  • Every record checked, not a sample.
  • No borderline case settled without a person.
  • Any automated verdict can be opened with its evidence and reason, and overturned.
  • A matcher with no model components.
  • Accuracy measured against the human reference per country, portal and record type, so weak groups stay visible instead of hiding inside an overall figure.

Other Data Businesses With the Same Risk

A business that sells data assembled from third-party sources it does not control carries the same risk:

  • reconciliation is done by hand, record by record;
  • only a small share of the data ever gets checked;
  • some cases need a person's judgment rather than a rule;
  • every decision about the data has to be explainable afterwards.

Job boards, vehicle marketplaces, multi-seller product catalogs and company registry aggregators are typical cases. The same layer can sit beside other pipelines, and the share of cases sent to people is set with the business that owns the data. See human-in-the-loop and LLM evaluation in our glossary for the methods, and our big data consulting and data analytics services for how such projects are run.

100%

of records checked, not a sample

20+

countries in the provider's data

Other cases

AI Agent Platform With Governed Model Access: Many Agents, One Runtime

AI Agent Platform With Governed Model Access: Many Agents, One Runtime

Read more ›
AI Candidate Screening That Checks CV Claims Against Public Evidence

AI Candidate Screening That Checks CV Claims Against Public Evidence

Read more ›
ChatGPT-Powered Content Generation Platform

ChatGPT-Powered Content Generation Platform

Read more ›
AI Data Analyst in Microsoft Teams: Answers From Live Operational Data

AI Data Analyst in Microsoft Teams: Answers From Live Operational Data

Read more ›
AI Delegation Agent for Executive Teams: Assignments, Statuses and Scores in One Loop

AI Delegation Agent for Executive Teams: Assignments, Statuses and Scores in One Loop

Read more ›
AI Image Annotation Solution for Medical Assessment

AI Image Annotation Solution for Medical Assessment

Read more ›
ML/Computer Vision + Data Annotation for ADAS

ML/Computer Vision + Data Annotation for ADAS

Read more ›
AI/ML-Driven Telecom Network Optimization

AI/ML-Driven Telecom Network Optimization

Read more ›
Custom GenAI Medical Drug Search Accelerator

Custom GenAI Medical Drug Search Accelerator

Read more ›
ML/Computer Vision-Powered Data Annotation for Sports

ML/Computer Vision-Powered Data Annotation for Sports

Read more ›
OCR Retail Receipt Data Extraction Tool

OCR Retail Receipt Data Extraction Tool

Read more ›
Computer Vision + Image Annotation for Insurance

Computer Vision + Image Annotation for Insurance

Read more ›
AI Presale Assistant for a Marketing Agency: From Meeting Transcript to Sourced Draft

AI Presale Assistant for a Marketing Agency: From Meeting Transcript to Sourced Draft

Read more ›
AI Pricing Engine for an Equipment Distributor: The Model Proposes, a Person Approves

AI Pricing Engine for an Equipment Distributor: The Model Proposes, a Person Approves

Read more ›
AI Trade Contractor Sourcing: Verified Profiles and Pre-Screening for a Home Builder

AI Trade Contractor Sourcing: Verified Profiles and Pre-Screening for a Home Builder

Read more ›
Cinema Scheduling and Pricing Optimization: A Week of Showtimes, Computed

Cinema Scheduling and Pricing Optimization: A Week of Showtimes, Computed

Read more ›
Computer Vision + ChatGPT-Powered Recommendations

Computer Vision + ChatGPT-Powered Recommendations

Read more ›
Data Annotation for AI-Driven Sports Analytics

Data Annotation for AI-Driven Sports Analytics

Read more ›
Generative AI Bot for Customer Service

Generative AI Bot for Customer Service

Read more ›
Generative AI Solution for Code Compliance

Generative AI Solution for Code Compliance

Read more ›
SaaS Marketing Content Generation Platform for Healthcare

SaaS Marketing Content Generation Platform for Healthcare

Read more ›
Interactive Chatbot with an Animated Avatar

Interactive Chatbot with an Animated Avatar

Read more ›
Managed Video Annotation Pipeline for a Client's Own Human-Motion Footage

Managed Video Annotation Pipeline for a Client's Own Human-Motion Footage

Read more ›
ML/AI Technical Due Diligence for Startup Acquisition

ML/AI Technical Due Diligence for Startup Acquisition

Read more ›
OCR Solution for Invoice Processing

OCR Solution for Invoice Processing

Read more ›
Property Management Platform With an AI Voice Assistant for Multiple HOAs

Property Management Platform With an AI Voice Assistant for Multiple HOAs

Read more ›
AI-Powered Platform for Real Estate

AI-Powered Platform for Real Estate

Read more ›
AI Assistant for Report Generation & Visualization

AI Assistant for Report Generation & Visualization

Read more ›
Sitecore Migration

Sitecore Migration

Read more ›
ML/AI Technical Due Diligence for SporTech

ML/AI Technical Due Diligence for SporTech

Read more ›
Voice Interview to Social Posts: AI Content in the Professional's Own Voice

Voice Interview to Social Posts: AI Content in the Professional's Own Voice

Read more ›

Let's talk about your project

We'll help you harness AI and ML technologies to create innovative software solutions

File requirements: pdf, doc, docx, rtf, ppt, pptx — max 15 MB

No file selected