Local AI Presence Monitor — Changes & Alerts

Monitor a local business over time instead of repeating a one-off audit. This Actor compares each new snapshot with persistent state, detects new public Google reviews, rating and review-count changes, website AI-readiness changes, and score drops. It can also benchmark the target against up to three competitors with the same Google/review scoring method. It returns structured JSON, retains up to 50 snapshots and can send high-severity alerts to your webhook.

Use this tool when

• You already know which business, store, hotel, restaurant or agency client you want to monitor.
• You need a durable baseline and a clear answer to “what changed since the last run?”
• You want new-review detection without receiving the same review as new on every run.
• You want an automation webhook only for important changes, while keeping every observation in JSON.
• You need a scheduled monitoring step that an AI agent, CRM or reporting workflow can call through API or MCP.
• You want to compare a business with up to three competitors using an explicit, comparable Google-profile and recent-review score.

Use Local AI Presence Audit instead when you need a one-time full audit with an action plan, llms.txt and LocalBusiness JSON-LD deliverables. Use Review Radar when your primary workflow is Slack alerting for Google reviews.

What changes are detected

The first successful run creates a baseline. It does not mark all existing reviews as new. Subsequent runs can detect:

• New review IDs in the recent Google review window, with Bayesian urgency classification.
• Google rating and total review-count changes.
• Presence-score changes and configurable score-drop alerts.
• Loss or addition of LocalBusiness, FAQ and other JSON-LD signals.
• Meta description, H1 and Open Graph changes.
• /llms.txt availability changes.
• robots.txt AI-crawler rule changes.

The website scan measures technical AI-readiness signals. It does not claim that a business is cited by ChatGPT, Gemini or Perplexity.

Quick start

1. Enter a business name and city in query, a public websiteUrl, or both.
2. Keep dryRun enabled for the first integration test. Dry-run skips billing events and webhook delivery; platform compute and proxy usage can still apply.
3. Run the Actor once to create the baseline, then run the same input again.
4. Read verdict, changes, newReviews and alertDeliveryStatus in the dataset.
5. Add a schedule in Apify only after the two test runs produce the expected stateful behavior.

For a website-only monitor, leave query empty. This avoids browser and Google proxy usage. For Google Maps monitoring, the low-cost default uses datacenter proxy. If Google blocks the request, select the RESIDENTIAL proxy group and account for separate residential-traffic costs.

Input

Field	Purpose
query	Optional business name and city used to resolve the Google Maps profile.
websiteUrl	Public website scanned for technical AI-readiness signals.
maxReviews	Recent reviews inspected per run, from 1 to 30.
competitorQueries	Optional list of up to three competitor names with city. Produces comparable scores and a target rank.
scoreDropThreshold	Point drop that makes a score change high severity. Default: 5.
ratingDropThreshold	Star-rating drop that triggers high severity. Default: 0.2.
webhookUrl	Optional secret URL receiving a JSON POST for high-severity alerts.
slackConnector / slackChannel	Optional MCP connector and channel for high-severity Slack alerts outside dry-run. Credentials stay in the connector proxy.
language	Google Maps interface language, such as fr, en or de.
dryRun	Skips event charges, webhook and Slack delivery.
proxyConfiguration	Apify proxy settings used only when query is present.

Output

Each run writes one main dataset item:

• verdict: baseline, unchanged, changes-detected or error.
• presenceScore and scoreComponents: transparent scoring, not a third-party ranking promise.
• rating, reviewCount and reviewsAnalyzed.
• newReviewCount and newReviews.
• changeCount, changes and alertCount.
• alertDeliveryStatus: no-alerts, not-configured, delivered, failed or dry-run-skipped.
• websiteAudit and googleProfile for the current observation.
• historySize and a stable monitorStateKey derived from the target.
• competitors with one comparable result per successful competitor.
• competitorComparison with target score, average competitor score, difference, rank and the exact comparison method.

The default key-value store also contains SUMMARY, HISTORY and ALERTS. Persistent named storage keeps state between runs; HISTORY retains the newest 50 snapshots.

Webhook payload

When alerts exist, dry-run is disabled and webhookUrl is configured, the Actor sends a JSON POST containing text, business, capturedAt, alerts and newReviews. Use your own authorized automation endpoint. A delivery failure is reported in the dataset; the monitoring snapshot is still retained.

Pricing

Pay-per-event pricing:

Event	Price	When
run-started	$0.25	Once per run, outside dry-run.
business-monitored	$0.25	When at least one source is successfully analyzed.
competitor-compared	$0.10	Per competitor successfully analyzed outside dry-run.
alert-delivered	$0.10	Only after successful webhook delivery.

Apify compute and proxy usage are separate. A website-only run avoids Google browser/proxy work. The public prefilled scenario is website-only and uses dry-run for inexpensive QA.

API and AI agents

Run by name through the Apify API, or use the shared MCP server at https://mcp.apify.com. An agent should read the input schema, preserve the same target between runs and use the returned change objects rather than inferring changes from prose.

Example input:

{
  "query": "restaurant Esch-sur-Alzette",
  "websiteUrl": "https://example.lu",
  "maxReviews": 5,
  "scoreDropThreshold": 5,
  "ratingDropThreshold": 0.2,
  "dryRun": true,
  "proxyConfiguration": {"useApifyProxy": true}
}

This Actor supports Apify agentic payments through x402/USDC when called through the enabled payment flow. Standard authenticated API and MCP calls remain available.

Scheduling and limits

Schedule the same saved task daily or weekly. Changing query or websiteUrl creates a different target state key. Keep the input stable for meaningful comparisons. A recent-review window can miss a review if it disappears from that window between two infrequent runs; use a frequency appropriate for the business. Public web data and third-party availability can vary. Only collect data you are authorized to process and review applicable privacy and platform terms.

Related Actors

• Local AI Presence Audit — one-time audit and ready-to-publish fixes.
• Review Radar — Google review monitoring and Slack workflows.
• AEO AI Visibility Checker — website-only technical AI visibility checks.

Built by Adaga solutions. For support, open an issue on the Actor page with the run ID, input and expected behavior. Include enough detail for an honest review of the problem; do not include webhook secrets or personal credentials.