Live preview boundaryLocal/mock review mode: no Supabase/Postgres runtime, no email, no analytics, no outreach, no export/copy/send/social/live-provider actions, no production DB mutation.

external actions: falsepreview only

Approval preview · read-only

Madrid Dental Clinic

Requested decision: approve manual next action

risk: lowsource: locallead: lead-mock-opp-001readiness: ready for approvalnoExternalAction: true

Detailed proof

“I called three times and nobody answered, then had to book somewhere else.”

https://example.com/reviews/madrid-dental-clinic

Recommended manual action

Send a short audit showing the repeated missed-call complaint pattern and a 7-day AI front-desk trial offer.

Blocker / limitation notes

No blocking issue detected, but this remains preview/read-only and still needs human approval.

This approval preview does not execute approvals, send messages, copy text, export data, or contact any external system.

Back to approval queue

Local score summary

fit93strong

intent64promising

confidence93strong

Decision options

previewOnly: true · noExternalAction: trueApprove for manual follow-up prep

Preview-only approval to prepare a human-reviewed follow-up asset later. It does not send, copy, export, or contact anyone.

previewOnly: true · noExternalAction: trueMark needs more research

Keep the Lead Card in review because more public proof or source context is needed before a human action.

previewOnly: true · noExternalAction: trueReject / not fit

Reject the lead locally because the fit, timing, or evidence is not strong enough.

previewOnly: true · noExternalAction: trueBlock due to risk or missing evidence

Block the lead locally when risk is too high or evidence is incomplete. No external action is possible from this queue.

Decision history

2026-07-06T11:00:00.000Z · local reviewer · previewOnly: true · noExternalAction: trueApprove for manual follow-up prep

Preview decision: evidence and score are strong enough to prepare a manual follow-up asset later, but no external action is executed.

Score reasons

fit · +18Opportunity language matches the product ICP.

evidence: ev-001

fit · +18Opportunity contains pain language from the buyer-intent map.

evidence: ev-001

fit · +12Opportunity maps to the target outcome language.

evidence: ev-001

intent · +24Visible buying/pain signal appears in summary or evidence.

evidence: ev-001

confidence · +35Source evidence is attached before scoring.

evidence: ev-001

confidence · +15Evidence includes source URL and quote.

evidence: ev-001

confidence · +8Multiple scoring signals agree.

evidence: ev-001

risk · +5Risk is low because confidence is strong and channel is email.

evidence: ev-001

Evidence references

local · missed-call complaintPublic reviews

“I called three times and nobody answered, then had to book somewhere else.”

https://example.com/reviews/madrid-dental-clinic

Limitations

Deterministic local scoring only; no LLM, source hunter, outreach, export, or production database mutation runs.Lead card is normalized from deterministic local fixtures only; no outreach, export, live source call, browser automation, or production database mutation runs.Approval queue is preview/read-only: no approval execution, outreach, send, export, copy, live source call, browser automation, or production database mutation runs.