A useful context
An API, MCP tool, Nicola response or machine-readable service answers a real task. Placement follows the task and surface.
URSY.AD · Machine Attention
Machines research, retrieve, compare and call useful services. Agent Discovery Sponsorship makes disclosed advertiser facts available in those contexts, then asks what happened next.
Machine Attention means software systems reading, calling or using a digital service. It does not mean a machine thinks like a person.
The product
A clearly disclosed block of sponsored information delivered on a relevant machine-facing surface. The receiving system decides whether the information is useful.
An API, MCP tool, Nicola response or machine-readable service answers a real task. Placement follows the task and surface.
The advertiser is named. Claims carry sources. Sponsorship is explicit. There are no hidden instructions to recommend or prefer the brand.
Separate accepted emissions from discovery, deeper use, observed citations, human referrals and conversions. Publish uncertainty alongside results.
What counts
These are the stages we look for, not a claim that every stage occurred. One stage never proves the next.
A qualifying sponsor-bearing application emission was centrally accepted under the published proof policy. It does not prove socket delivery, model reading, recommendation, citation or a sale.
How we prove it →A qualifying machine fetched a campaign-identified advertiser destination after a qualifying exposure within a declared attribution window.
How we prove it →Change in qualifying advertiser exploration after discovery relative to a declared baseline or control.
How we prove it →Measured change in observed AI answers mentioning or citing the advertiser against baseline or control.
How we prove it →A verified or probable human arrived at the advertiser through AI search or an assistant.
How we prove it →A declared conversion is attributed under published rules to an AI-mediated referral chain.
How we prove it →Public study · first-party evidence
What happened after the sponsor was served? Each stage has its own evidence requirement.
384
MeasuredUnit: events
Exact rolling30d accepted POST application emissions, excluding known audit events; partial estate coverage is disclosed.
Source: https://ursy.ad/api/v1/vme/campaigns/teas-ursy-2026-09-v1/report
Window: Rolling 30 days; see partial observation coverage
Read the definition →Not measured
Not MeasuredUnit: events
The TEAS campaign-specific arrival and exposure matching bridge is not connected.
Source: No verified source connected
Window: Rolling 30 days; see partial observation coverage
Read the definition →Not measured
Not MeasuredUnit: seconds
No directly linked exposure/discovery timestamps are connected.
Source: No verified source connected
Window: Rolling 30 days; see partial observation coverage
Read the definition →Not measured
Not MeasuredUnit: reads
Post-discovery advertiser exploration and a declared baseline/control are not connected.
Source: No verified source connected
Window: Rolling 30 days; see partial observation coverage
Read the definition →Not measured
Not MeasuredUnit: arrivals
TEAS verified human AI click-back evidence is not connected.
Source: No verified source connected
Window: Rolling 30 days; see partial observation coverage
Read the definition →Not measured
Not MeasuredUnit: conversions
Verified advertiser conversions and published chain attribution are not connected.
Source: No verified source connected
Window: Rolling 30 days; see partial observation coverage
Read the definition →A configured sponsor is not a verified exposure. Not measured means the required evidence is unavailable; zero means measurement found no qualifying events.
Adverts without tracking people
Human Attention and Machine Attention share the privacy promise. Their delivery and outcome metrics stay separate.
Sponsored information is commercial information, with an identified source and supportable claims. A machine receives no command to rank, recommend, conceal or buy.
Read the sponsorship policy →Human impressions and machine exposures are never added into a combined reach figure. A zero or an inconclusive finding can be a valid study result.
Open the methodology →