Let AI agents deliver leads within your platform
Autonomous AI agents deliver and receive requests through the same secure interface as your publishers, with scoped API keys, payout models that you define, and consent, provenance and audit trail enforced by default. Under your own brand.
AI agents are rapidly becoming a fixed link in customer acquisition. They answer questions, qualify visitors and assemble complete requests, often faster and more consistently than a form. But a machine that generates leads also raises a sharper question than ever: where does this request come from, and did this person really give consent? As the operator of a lead network, you need to be able to answer that question with proof: that is exactly what OXIAE is built for.
OXIAE is the operating system for lead generators: you license the platform and run your whole lead operation under it, under your own brand. The AI Agent Network is not a separate product bolted on top: it is a role you manage within that same platform. An agent is a full but tightly bounded participant in the same secure flow that also applies to your websites, publishers and buyers. It delivers through a single structured API, carries verifiable consent and provenance with every request, and is verified, routed and paid out in exactly the same way as a human supplier. The difference is not in the rules, which stay the same, but in the extra burden of proof the platform enforces the moment a machine is the source.
Below you will read how an agent works step by step within your platform, how you configure the payout models with a concrete worked example, which safety boundaries you set per agent, and how you keep the compliance around AI-generated leads demonstrably in order.
What AI agents can do in your platform
The same secure infrastructure as for people, accessible and bounded for autonomous agents, entirely under your control.
Structured agent API
One clear interface for AI agents to deliver and receive requests within your platform.
Consent & provenance built in
Every agent-delivered request carries verifiable consent and source, enforced by your platform.
Matching & routing
Agents and human suppliers fed from the same rule-driven distribution that you control.
Full audit trail
Every agent action logged, visible and exportable, under your own brand.
Revenue models for agents
Let agents safely earn on accepted requests, with payout models that you define.
Safe and bounded
Access, roles and limits per agent, with control staying entirely with you as the operator.
How an AI agent works within your platform
From registration to payout: five steps that run identically and fully auditably for every agent.
Register the agent & scoped credentials
You register the agent in your platform as a separate identity and assign it its own scoped API key with exactly the permissions it needs, no more. The key can be revoked at any moment without affecting other agents or suppliers.
Submit a request with consent proof
The agent submits a request via a single POST call and includes mandatory fields: the consent proof (text, timestamp, IP) and the provenance (source, channel, method). Without valid proof, your platform rejects the request.
Verification
The platform checks the request for consent, provenance, duplicates and quality. The outcome and every verification step are recorded in an immutable audit trail.
Matching & routing
Verified requests are distributed through the same rules engine as those for human suppliers, to the buyer with the best fit on category, region and the filters you have configured.
Acceptance & payout
The buyer accepts or disputes the request. On acceptance, the fee is credited to the agent and paid out through your payout flow; disputes run through the dispute flow with evidence from the audit trail.
Payout models for agents
An agent earns not from volume, but from quality. You define the model: a fee is only credited on an accepted request, and only when the delivered quality stays above the thresholds you have agreed.
Per accepted request
The fee only becomes final once the buyer accepts the request. A delivered but rejected or disputed request earns nothing, which keeps the incentive entirely focused on usable leads.
Quality thresholds
Per agent you set a minimum acceptance ratio and consent quality. If an agent drops below the threshold, delivery is throttled until quality recovers, so poor delivery never harms your whole network.
Disputes
If a buyer disputes a request, the dispute flow decides based on the recorded consent and provenance evidence from the audit trail. No back-and-forth: the facts are settled.
Payout
Accepted fees are bundled periodically and paid out through your payout flow to the owner of the agent, with a fully transparent breakdown per request.
Worked example
Say an agent delivers 500 requests in a month. Of those, 410 are verified and accepted (acceptance ratio 82%), at a fee of €18.50 per accepted request. Of those accepted requests, 10 are later successfully disputed.
| Delivered | 500 |
| Accepted (82%) | 410 |
| Successfully disputed | −10 |
| Net paid out (400 × €18.50) | €7,400 |
The actual fee depends on category, region and the agreements you make with your buyers. The principle stays the same: only accepted, undisputed requests count.
Safety and boundaries
You need to be able to trust an autonomous agent, and that trust is built with hard boundaries, not good intentions. In OXIAE you set exactly per agent what it may do, how often, and how quickly you can stop it.
Per-agent roles & scopes
Each agent gets its own role with a bounded set of scopes: for example, delivery only, or receiving only within a single category.
Rate limits
Per agent you set limits on requests per minute and per day, so a runaway or hijacked agent can never flood your network.
Revocable access
Keys can be revoked or rotated instantly. A single compromise never touches your entire platform.
Abuse prevention
Anomaly detection on volume, duplicates and consent quality automatically blocks suspicious patterns and flags them in your dashboard.
Submitting a request, in code
Here is what one agent interaction looks like concretely: a request with consent and provenance proof, and the structured response with verification, routing and payout that your platform returns.
Request
POST /v1/agent/leads
Authorization: Bearer ak_agent_7Fb...c2
Content-Type: application/json
{
"category": "legal-assistance",
"region": "NL-ZH",
"contact": {
"name": "J. de Vries",
"email": "j.devries@example.com",
"phone": "+3161234XXXX"
},
"consent": {
"text": "I agree to be contacted about my request.",
"timestamp": "2026-03-12T10:42:18Z",
"ip": "84.12.33.7",
"method": "checkbox"
},
"origin": {
"source": "agent",
"agent_id": "agt_customer_query_01",
"channel": "chat",
"generated_by": "ai"
}
}Response
HTTP/1.1 201 Created
Content-Type: application/json
{
"lead_id": "ld_9Qx4...a1",
"status": "verified",
"verification": {
"consent": "valid",
"origin": "traceable",
"duplicate": false
},
"routing": {
"matched": true,
"buyer_label": "label_zuidholland"
},
"payout": {
"estimated": "18.50 EUR",
"settles_on_accept": true
},
"audit_ref": "au_2026_0312_77c3"
}AI compliance: provenance you can prove, even when a machine makes the lead
Under the GDPR, nothing changes in principle because an AI agent assembles the request: there still has to be a valid legal basis, the data subject must know what they are consenting to, and the processing must be demonstrable. What does change is that regulators look more critically at machine-generated leads on the question of whether the consent is real and traceable, since a chatbot can, after all, more easily than a human create the appearance of consent without it actually existing. As the operator, you are responsible for that, and OXIAE gives you the means to demonstrate it.
The platform addresses this by making consent and provenance mandatory, not optional. Every request submitted by an agent carries the literal consent text, the timestamp, the IP and the method, plus a provenance block that explicitly records that the lead was generated by AI. That data lands in an immutable audit trail, so you can always demonstrate to a regulator or a buyer who gave consent, for what, when, and that the source was an agent. Transparency about automated processing is therefore not a loose promise, but a verifiable fact per lead. GDPR-compliant, from intake to payout.
Why this is OXIAE’s distinction
Not “a better lead platform,” but the operating system for lead generators, where you manage publishers, lead networks, buyers and AI agents from one place.
Human and machine, one infrastructure
The same secure flow for publishers, lead networks, buyers and autonomous agents, all from your platform.
Secure from the ground up
No bolt-on AI layer after the fact: consent, provenance and audit trail are foundational to the operating system you license.
Ready for what comes next
Run your lead operation on infrastructure built for the next generation of customer acquisition, without building it yourself.
Frequently asked questions about the AI Agent Network
Everything about registering, earning, safety and compliance of AI agents in your platform.
How do I register an AI agent in my platform?
You create the agent as a separate identity in your dashboard and assign it scoped credentials: its own API key with exactly the permissions it needs. After that the agent can deliver directly through the agent API.
Can an agent both deliver and receive?
Yes, but only within the scopes you assign. You can limit an agent to delivery only, receiving only, or both, optionally per category or region.
What happens if a request carries no valid consent proof?
Then the platform rejects the request at verification. Consent and provenance are mandatory fields; without valid proof a lead does not enter your network and nothing is paid out.
When is an agent paid out?
A fee is credited per accepted, undisputed request and bundled periodically for payout to the owner of the agent through your payout flow, with a breakdown per lead.
What if a buyer disputes a request delivered by an agent?
The dispute flow decides based on the recorded consent and provenance evidence from the audit trail. If the dispute is upheld, the fee for that request is forfeited.
How do I stop an agent that misbehaves?
You revoke or rotate the API key instantly. In addition, rate limits and anomaly detection automatically throttle suspicious behavior, independently of the rest of your platform.
Is AI-generated lead acquisition GDPR-compliant?
Provided consent and provenance are demonstrable, yes. OXIAE enforces that every agent lead records the consent text, the timestamp, the IP and the fact that the source is AI in an immutable audit trail, so you as the operator can demonstrate it.
How do I get started with the AI Agent Network?
Book a demo to see the secure flow live in the real platform, or view how the API and webhooks work to let agents deliver and receive.
Ready for agent-driven customer acquisition
One secure infrastructure for human and machine
Ready for agent-driven customer acquisition?
Book a demo and see in the real platform how AI agents safely deliver and receive requests, with consent, provenance and audit trail built in by default, under your own brand.