About Agentry

Agentry is an English-language news outlet covering what AI agents do — fraud committed, court rulings handed down, deployment failures, regulatory action. We write only when an autonomous agent has taken a verifiable action with real-world consequence.

Editorial Line

Most AI coverage is about what models could do, what labs announced, what benchmarks they hit. We cover something different: what AI agents actually did — in courtrooms, in production systems, in financial transactions, in the real world.

We cover:

• Verified fraud, scams, and crime where AI agents are the instrument or the actor
• Court rulings, sentencing, and regulator action involving AI systems
• Autonomous trading actions and deployment failures with documented impact
• AI Act enforcement and security incidents tied to specific agents

We do not cover:

• Model releases, capability benchmarks, or lab announcements
• Funding rounds and acquisitions (unless tied to a verified incident)
• Hypothetical risks, opinion pieces, or generic AI commentary
• Stories without a primary source — court filing, regulator press release, established newsroom byline, or security firm advisory

The Four Beats

Every story belongs to one of four beats. The beat is a function of what the agent did, not where the story originated.

• Crime — deepfake fraud, voice-clone scams, autonomous misconduct, AI-enabled crime
• Economy — autonomous trading actions, agent-driven companies, AI labour displacement events
• Policy — court rulings, regulator action, EU AI Act enforcement, sentencing
• News — verified incidents, deployment failures, real-world agent actions

EU AI Act Article 50 Disclosure

Articles on Agentry are written by AI agents under human editorial supervision. We disclose this on every story machine-readably (in JSON-LD and YAML frontmatter on /agent/ routes) and human-readably here.

Creator: Agentry Newsroom (autonomous AI editorial agent). The agent reads the brief, researches the story via primary sources, drafts the article in English, and selects an appropriate scene-only image.

Editor: Susanne Sperling, Stratechmedia ApS — human in the loop. Every article is reviewed for factual accuracy, source quality, and editorial fit before publication. No article publishes without explicit human approval.

Tools: Stories are sourced via PublisherAgent (proprietary news-gap scanner), researched via Perplexity (primary-source verification), drafted by Anthropic Claude, and illustrated via Replicate Flux. All are off-the-shelf, well-documented systems.

Date reviewed: Each article carries the date the human editor reviewed and approved it. This appears in the footer of every story and in the article schema.

Verification

Every claim in an Agentry article is grounded in primary sources — court filings, regulator press releases, established newsroom reporting, or security firm advisories. Sources are linked at the bottom of each article. If a claim cannot be sourced to a primary record, the article is killed before publication.

When an autonomous agent acts (e.g. an AI coding agent that deletes a database, a deepfake that defrauds a finance officer), we name the system, the operator, the date, the loss, and the source — never abstractions, never composite cases.

Corrections

If you spot an error, contact us. We correct factual errors in place and add a correction note at the bottom of the article. Major corrections (changing the thrust of a story) are surfaced in the article header. We do not silently rewrite.

Independence

Agentry is independently published by Stratechmedia ApS, a Danish company. We accept no payment for coverage. AI labs, vendors, and tooling companies have no influence on what we write. If we cover an action by a tool we use, we disclose it explicitly in the article.

Editor & Publisher

Editor: Susanne Sperling, susanne@stratechmedia.com

Fifteen years in Danish digital media. Former CCO/CFO of bold.dk (2008–2023). Founder of Stratechmedia ApS. Editor of KrimiNyt.dk.

Publisher: Stratechmedia ApS, Copenhagen, Denmark.

Contact

For tips, corrections, press inquiries, or feedback:

Email: susanne@stratechmedia.com

We aim to respond to all queries within five business days.

Machine-readable view

Agent and crawler clients can access this site at /agent (markdown index) and /agent/<slug> (per-story markdown). Article-level provenance metadata (creator, editor, tools, dateReviewed) is exposed in YAML frontmatter and JSON-LD.