The editor.

Peter is the named human who signs every piece published by Agent Mode AI. He is the editor, the curator, and the accountable party for what ships. He is not the writer — Claude drafts the prose, end-to-end — and the publication does not pretend otherwise. The arrangement is disclosed on every article header, in the byline, in the footer, and in detail at the accountability architecture for AI-written publications (AM-CANON-001).

Who Peter is

Peter is a Dutch IT operations leader based in Friesland, working at enterprise-scale in industries with long-cycle change management and high regulatory exposure. He's been doing this for more than two decades. He is bilingual EN/NL; this publication is English-primary.

He is editorially independent — his employer is not named on this site, does not commission pieces here, and does not appear in the structured-data author signal. The publication is operated under his own name and reflects his analysis of the enterprise-AI market, not his employer's posture. This is the constraint that makes the publication credible to senior IT readers; it is also the constraint that rules out paid speaking, advisor placements, and other personal distribution surfaces.

You can reach him at peter@agentmodeai.com for corrections, dissent, or substantive feedback on a tracked claim. He's on LinkedIn.

What "signing off" means in practice

Every article on this site is drafted by Claude before Peter sees a word of it. Peter sets the brief — what the piece argues, who the reader is, what claim it asserts — then reads the draft, verifies the cited primary sources, runs the credibility scanner, and either signs off or sends the draft back for a rewrite. Nothing publishes that he hasn't read end-to-end and approved on the specific claim it stakes.

The publication does not pretend the writing is Peter's. The byline reads "Written by Claude · Curated and signed by Peter" because that is the honest description of the production model. The credibility comes from what happens after publication, not before it: the claim is tracked in the public ledger at /holding/, reviewed on a 30–90 day cycle, and visibly held to account as the evidence moves. A claim that turns out to be wrong is not quietly removed. The original text stays; the retraction is dated; the corrected reality is named.

This is the trust mechanism. Peter is the signatory, not the writer, and the byline names him because he is the human the reader can hold accountable for what stands.

Why this works

The argument is canonised at AM-CANON-001 — the foundational reference piece on why an AI-written publication can be more verifiable, not less, than a human-written one, when it runs inside an explicit accountability architecture. Read it at /canon/v1.0/accountability-architecture/.

Corrections

Spot a factual error, a broken citation, or a mis-cited framework? Email corrections@agentmodeai.com. A dated correction note will append to the article, the claim verdict may shift, and the change will appear in the public ledger. The publication's full corrections policy is on the standards page.