Public evidence
Observed website facts and cautious inferences, each returned with its source and evidence identifier.
Agent Business Intelligence
MarginGlow AI Signal gives an authorised agent evidence-based, machine-readable opportunities to improve a specific small business. The calling agent keeps its customer relationship. MarginGlow supplies the specialist commercial intelligence.
What an agent receives
Observed website facts and cautious inferences, each returned with its source and evidence identifier.
Three commercially meaningful opportunities covering revenue, margin, cost, conversion, Google visibility or operations.
A clear first action, confidence level and directional commercial impact for every opportunity.
How it works
The agent reads the MCP tool description and determines whether MarginGlow AI Signal fits the user's business question.
The authorised agent supplies the business name, public website, country and commercial objective.
MarginGlow returns structured intelligence that the calling agent can cite, explain or incorporate into its own workflow.
Machine-readable request
{
"businessName": "Example business",
"website": "https://example.com",
"country": "Ireland",
"objective": "Increase weekday bookings"
}Structured response
{
"evidence": [...],
"opportunities": [
{
"rank": 1,
"title": "...",
"evidenceIds": ["e1"],
"confidence": 0.9,
"commercialImpact": "high",
"firstAction": "..."
}
],
"unknowns": [...],
"limitations": [...]
}Trust boundaries
Beta access
We are inviting agent builders to test MarginGlow AI Signal before the wider credit marketplace opens.