Your application understands when a user is preparing to buy, book, hire, compare, or subscribe. AIAdsExchange gives that intent a destination, returning highly relevant sponsored recommendations.
Traditional advertising networks rely on cookie tracking, public page URLs, or keyword queries. AIAdsExchange introduces a cleaner abstraction: **Structured Workflow Intent**.
When a user acts—submitting a travel form, requesting a software comparison, clicking a booking trigger, or asking a conversational assistant—your application sends a structured intent payload. We scrub PII context, execute policy filters, search inventory, and return a compliant recommendation.
Integrate the SDK directly in your backend or client-side context to request sponsored cards.
{
"intent": "buy_payroll",
"category": "business_software",
"source": "workflow_event",
"country": "US",
"context": {
"company_size": 18,
"budget": "under_300",
"current_tool": "QuickBooks"
},
"placement": {
"type": "sponsored_recommendation",
"max_results": 3
},
"privacy_mode": "strict_scrub"
}
{
"status": "success",
"results": [
{
"title": "Gusto Payroll & HR",
"description": "Automate payroll, tax filings, and compliance for teams under 50.",
"label": "Sponsored",
"bid_type": "CPL",
"reason": "Matched B2B software intent for a U.S. company with under 25 employees.",
"tracking_url": "https://aiadsexchange.com/click/tr_a8f9b2"
}
]
}
| Parameter | Type | Required | Description |
|---|---|---|---|
| intent | string | Yes | The core commercial goal (e.g. buy_payroll, book_flight, compare_saas). |
| category | string | Yes | Vertical alignment (e.g. Travel, B2B SaaS, Shopping & E-commerce). |
| source | string | No | The UI source: chat_prompt, workflow_event, button_click, form_submission, agent_action. |
| country | string | No | ISO 2-letter country code for geographical targeting. |
| context | object | No | Key-value pairs representing anonymized filters (e.g. company_size, budget, duration). |
| privacy_mode | string | No | Encryption or scrubbing parameters: strict_scrub or coarse_anonymize. |