Review analyzer
Paste any review and get a risk score against a recall-targeted threshold, the complaint aspects it contains and a recommended action, in under a second.
Product
Everything between a review and a decision.
HomeLens AI is a risk command center for home-service providers. It reads reviews, explains complaints, monitors providers over time and drafts the next step for a person to review.
Provider risk leaderboard
Rank 254 anonymized providers by a percentile-based risk score. Filter by tier, search by code and open any row for its yearly trajectory.
Complaint-aspect radar
Eight aspects (workmanship, reliability, pricing, communication, timeliness, professionalism, warranty, safety) sized by lift. Select one to filter the leaderboard.
Explainable evidence
Every flag shows the sentence behind it, the phrases that matched and the term weights from the served model. If nothing matches, it routes to manual review.
Action playbooks
Draft a triage response email or a quality-audit checklist for a provider, an aspect or a single review. Drafts are for human review; nothing is sent.
Human in the loop
Signals are review-based and relative. Every Elevated or High-concern tier is an invitation for a person to investigate, never a verdict.
Command center
One screen for the whole risk picture.
Three headline numbers, a live leaderboard, the aspect radar and the evidence feed share a single view, so a manager can go from “what changed?” to “what do I do?” without switching tools.
Aspect radar
See what customers keep complaining about.
The radar sizes each aspect by lift, frequency or negative rate. Hover for its keyword cluster; select it to filter providers whose top signal matches.
Triage
From signal to next step.
Select a provider, an aspect or a flagged review and open the triage dock. HomeLens drafts a response email or an audit checklist that a manager can edit, copy and own.
Intended use
Built for the people who own service quality.
Operations and quality teams
Spot providers drifting toward repeat complaints before they become a pattern, and decide where to audit first.
Customer-experience leads
Understand which complaint aspects are rising across a service category and what to coach on.
Marketplace trust teams
Prioritize human review of providers whose reviews carry safety, conduct or billing language.
Status
What is in the demo today.
Available now
- Paste-a-review analyzer (TF–IDF model, DistilBERT in the notebook)
- Provider leaderboard, trends and tiers on anonymized data
- Aspect radar, evidence feed, model weights
- Triage email and audit checklist drafts
- Documented /api/analyze endpoint
On the roadmap
- Upload your own review export
- Hosted DistilBERT inference
- Saved provider watchlists and alerts
- User accounts and audit logs
- Additional markets beyond Tucson, AZ
