How AI Reads Press Releases: 75.8% of Activity Occurs Within the First 24 Hours
- Melissa Strle
- 26 minutes ago
- 4 min read
Disclosure Intelligence Series
How disclosure works in modern capital markets
When a press release is distributed, most issuer teams treat the process as complete. In reality, this marks the point at which AI systems begin processing, interpreting, and positioning the announcement for the market.
To better understand how press releases are processed in modern markets, TMX Newsfile analyzed AI crawler activity across a 72-hour window following press release distribution.
The study tracked approximately 120 releases distributed across TMX Newsfile’s global distribution network during a peak news cycle. This allowed us to observe how quickly press releases are processed, how widely they are distributed, and which AI systems are interpreting them.
The findings point to a structural shift in how information is consumed. Disclosure is no longer only consumed by people. It is also processed by AI systems.
This study answers three operational questions for issuer teams:
How quickly are press releases processed by AI systems?
How do AI systems use news releases once they are captured?
Which AI-driven platforms process press releases first?
The results show that AI ingestion is immediate, global, and highly concentrated within the first 24 hours of release.
AI Processing Speed and the 24-Hour Visibility Window
Speed is one defining factor in how press releases are interpreted in modern markets. Once a press release is distributed, AI systems begin processing it almost immediately, establishing an initial understanding of the announcement that can influence how it is surfaced, summarized, and referenced.
To quantify this timing, we analyzed AI crawler activity over a 72-hour period following the distribution of approximately 120 press releases across TMX Newsfile’s network of more than 500 channels. The results show a highly concentrated window of activity within the first 24 hours.
Figure 1: AI crawler activity over a 72-hour post-release window
A three-day view of AI crawler activity shows that 75.8% of all observed AI activity occurs within the first 24 hours after publication. The data also reveals a rapid decline in activity after the initial visibility window.
The majority of activity occurs on Day 1
Activity declines significantly after the initial window
By Day 3, AI visibility is materially reduced
This defines a clear AI visibility window, where the first 24 hours are driven by active AI processing, interpretation, and positioning of press releases in the market.
How AI Systems Process Press Releases
AI systems process press releases for two fundamentally different purposes.
Some systems retrieve press release information in real time to answer user questions, while others continuously ingest press releases to train future AI models and improve long-term understanding of companies, sectors, and market events.
To better understand this behavior, TMX Newsfile analyzed the balance between grounding activity and training activity during the first 72 hours after a press release is issued.
Figure 2: Grounding vs. training AI activity over a 72-hour window
The results show a clear distinction between real-time AI retrieval and longer-term model training.
On Day 1:
92.7% of AI activity is training based, where large-scale crawlers collect press release data to build and improve future AI models
7.3% is grounding based, where systems retrieve information in response to user questions
Grounding activity declines rapidly after the initial visibility window, falling by approximately 92% from 571 hits on Thursday to just 45 by Saturday.
Training activity also declines over time, but remains substantially higher than grounding activity, decreasing from 7,293 hits on Thursday to 808 by Saturday.
This reveals an important shift in how press releases are processed by AI systems.
Both grounding and training activity decline after Day 1, but training activity remains substantially higher throughout the 72-hour window.
As a result, the first 24 hours play a critical role in shaping how AI systems understand and retain a press release.
Which AI Platforms Process Press Releases First?
Figure 3: AI platform activity during the first 24 hours
AI processing activity on Day 1 comes from a globally distributed mix of platforms, including systems linked to:
ByteDance / Doubao (China)
OpenAI (United States)
Meta (United States)
Amazon (United States)
Baidu / Ernie Bot (China)
While this chart reflects the most active AI platforms observed during the study window, major platforms such as Google, Microsoft, Anthropic, and X also actively process press release content across TMX Newsfile’s network.
The results show that press releases are rapidly processed across major AI ecosystems in both North America and Asia within hours of publication.
This suggests that corporate announcements are no longer confined to traditional investor audiences or domestic distribution channels. Instead, press releases are increasingly being interpreted, categorized, and retained by globally distributed AI systems almost immediately after release.
What This Means for Disclosure Strategy
AI systems begin processing press releases immediately after publication, with the highest concentration of activity occurring within the first 24 hours.
As a result, visibility is no longer determined solely by where a press release is distributed. It is also shaped by how AI systems interpret and categorize information during the first hours after release.
Timing AI activity is most concentrated immediately after publication, making the first release window critical for visibility.
Structure Information must be clearly organized to support both human interpretation and machine processing.
Consistency When press releases, filings, and supporting information are managed separately, inconsistencies can emerge during the period when AI systems are actively processing and categorizing information.
In modern markets, the first 24 hours after a press release is published play a major role in shaping AI visibility.
For issuer teams, this increases the importance of clarity, consistency, and execution at the moment of release, when AI systems across multiple global platforms are actively processing and categorizing information.
Once released, a press release becomes part of a broader AI-driven ecosystem that interprets, categorizes, and stores information across digital platforms.
