Check several locations
See where the business is visible, weaker, failed, or not measured—without turning one point into a city-wide claim.
5 of 11 valid points in the top threeInspect the location view →Measure selected Google locations and sampled AI answers in one evidence-led view—then see the competitor, source, or location gap worth fixing next.

Local positionsObserved at selected points
AI recommendationsMentions and citations separated
Named competitorsSame tracked set and denominator
Next useful moveFinding first, suggestion second
Northstar Coffee appears at number 2 at one checked point. Radar widens the view: where positions weaken, which sampled answers mention it, and whether those answers cite its website.
Open the full customer preview ↗Fictional Manchester example · collected 14 October 2026 · local and AI checks retain separate coverage.
Move from a headline number to the location, prompt, source, comparison, or report behind it. All examples below are fictional.
See where the business is visible, weaker, failed, or not measured—without turning one point into a city-wide claim.
5 of 11 valid points in the top threeInspect the location view →Change the phrase while keeping the business set defined, so a service gap stays understandable.
7 coffee mentions · 3 breakfast mentionsCompare the phrases →See the exact prompt, supported method, valid answers, and which mentions include a usable website source.
7 mentions · 4 citations · 12 valid answersInspect the AI evidence →Compare observations only when the phrase, market, engine, location settings, and tracked set still match.
2 points worsened in compatible checksReview the history rule →Exact promptquiet client meeting
✓02Named sourcemenu page cited
↗03Collected14 Oct · 14:32
✓Inspect the prompt, method, source, collection date, and missing coverage behind a finding.
One finding · a visible evidence trailOpen the source example →Keep observed evidence separate from the suggested action, then carry a bounded report into the workspace.
3 priorities · 1 selected-business reportSee the action handoff →Google already offers nearby discovery, AI recommendations and Business Profile performance and insights. Radar's planned value is a consolidated, repeatable research workflow for your business—not privileged access to Google's ranking algorithm.
| Your task | An ordinary nearby search | Radar's planned workflow |
|---|---|---|
| Find a place | Lists, pins, business details and recommendations | Start with discovery, then inspect a selected business |
| Compare locations and services | Repeat searches and inspect each context | One defined set of points, phrases and competitors |
| Understand AI presence | Google's own AI experiences | Supported AI research methods, exact prompts and sources |
| Track and improve | Consumer discovery and separate profile insights | Compatible snapshots, explainable alerts, evidence and reports |
This comparison concerns the ordinary nearby-search journey, not every Google product or every SEO platform. Radar is not affiliated with Google.
Pick a business, area and the customer questions you want to examine.
Supported research services return observed local results and AI answers. Coverage stays visible.
Compare named businesses, inspect sources and see compatible changes over time.
Review evidence-backed suggestions, save your report and use your existing VisibleGen workspace.
Search → useful free preview → request the full comparison → verify your email.
Find your business → preview → verify email → request your bounded report. Ownership is a separate check.
Sign in → choose your authorized workspace and project → inspect evidence → prepare client reports.
See agency workspaceStart with bounded findings. Expanded coverage, retained history, monitoring and branded reporting will follow your existing VisibleGen plan and applicable data permissions.
Radar access and numerical allowances are still being calibrated. This preview does not introduce new pricing, unlimited checks or a separate subscription.
These explanations describe the approved design and planned integration. All interactive examples remain fictional.
Radar will collect observations at selected coordinates and group them into H3 Resolution-8 geographic cells. Each point keeps its phrase, market, method, date and valid/failed/not-measured status. The preview map is an illustration, not a live Google Map.
H3 is geographic indexing, not semantic vector retrieval. No embedding engine or vector database is claimed. Unmeasured cells are never filled with predicted rankings, customer traffic or revenue.
Inspect location records →GEO asks whether your business appears in the AI answers that were checked and which sources support those answers. Mention rate = answers mentioning the business ÷ valid answers. Citation rate = answers citing its domain ÷ valid answers. Share of voice = its mentions ÷ all mentions in the named tracked set, at most once per business per answer.
Failed answers are excluded and reported. API research is not automatically equivalent to a consumer AI app. Prompts, engine/model, method, supported geographic context and dates must remain visible. These checks cannot guarantee future inclusion.
Inspect AI coverage and denominators →The planned system validates scope and allowance, checks a compatible permitted cache, queues bounded requests to supported provider endpoints, normalizes returned businesses and attaches sources and collection status. Existing Laravel jobs dispatch report snapshots to the existing Python report workers.
This is not an unrestricted crawler or an agent roaming the web. Verification checks response structure, business identity and context; it does not independently prove every upstream claim. Failures stay visible, costs stay bounded and a report may be ready while research is partial.
See the sample activity trail →No. It can show an observed difference in a defined sample. A missing citation does not prove a particular website omission caused it. Suggested edits are testable hypotheses, not ranking or revenue guarantees.
Coordinates, phrase, engine/model, market, language, device, grid settings, prompt/measurement version and competitor set must match. Otherwise Radar labels the results Not directly comparable. Monitoring, alerts and notification controls here are simulated.
The planned private workspace retains only permitted data for each source's allowed period. Cache reuse requires matching measurement dimensions and tenant scope. Email verification is separate from ownership, and optional marketing consent is separate from access. Customers will be able to pause or cancel a monitor without deleting existing permitted reports; no real schedule or email is active in this preview.
Manchester, United Kingdom · English · Desktop style
Northern Quarter, Manchester · Prepared place sample
Verify an email to open the bounded comparison and save this prepared search.
Business menu and location details
The public preview stays useful; the bounded report needs verified email.
Verified email unlocks this prepared comparison or report. It does not prove business ownership, change a plan, or add marketing consent.
Manchester · English · 13 local samples · collected 14 Oct
The prepared search is preserved. No comparison, report, or ownership access is granted yet.
Nothing was lost. The prepared search can return to the request step without running research again.
Your bounded sample is unlocked. Business ownership remains Not verified.