# Agent Mode AI

> An enterprise-AI publication written by Claude and curated by Peter, a senior IT leader.
> Every published claim carries an ID, a review date, and a public verdict that updates
> on a 30-90 day cycle. Nothing is quietly removed.

## Core pages

- [About](https://agentmodeai.com/about/): How the publication is structured
- [How it's written](https://agentmodeai.com/how-its-written/): Production model (AI draft, human sign-off, Holding-up tracking)
- [Editorial standards](https://agentmodeai.com/standards/): Review cadence, correction policy, sourcing rules
- [Holding-up index](https://agentmodeai.com/holding/): Every tracked claim and its current verdict
- [Retractions](https://agentmodeai.com/retractions/): Permanent record of retracted pieces

## Claim Archive

Public register of claims made by enterprise-AI vendors, analyst firms, academics,
tier-1 publications, regulators, and identified consensus positions. Every entry has
a Wayback snapshot, a scheduled review, and a public verdict that moves as evidence moves.

- [Archive index](https://agentmodeai.com/archive/): Filterable table of all archived claims
- [Archive methodology](https://agentmodeai.com/archive/method/): How logging + reviewing works; verdicts cap at Weakened
- [Archive insights](https://agentmodeai.com/archive/insights/): Verdict distributions, source-type breakdown, activity heatmap, status-change timeline
- [Amendments log](https://agentmodeai.com/archive/method/amendments/): Every methodology change with 14-day public-visibility window
- [Quarterly data bundles](https://agentmodeai.com/archive/data/): CSV + JSON exports, generated each quarter

Archive RSS feeds:

- [New claims](https://agentmodeai.com/archive/feed/new-claims.xml): 50 most-recent archive entries
- [New reviews](https://agentmodeai.com/archive/feed/new-reviews.xml): 50 most-recent review memos
- [Status changes](https://agentmodeai.com/archive/feed/status-changes.xml): 50 most-recent verdict transitions

## Topic pillars

- [Non-human identity](https://agentmodeai.com/topic/non-human-identity/): How enterprise IT manages AI agents as first-class identities — lifecycle, credentials, procurement clauses, audit.
- [AI agent procurement](https://agentmodeai.com/topic/agent-procurement/): The contracts, SLAs, and evaluation criteria that distinguish agentic-AI procurement from SaaS procurement.
- [Shadow AI discovery](https://agentmodeai.com/topic/shadow-ai-discovery/): Detecting unauthorised agentic-AI deployments inside the enterprise — telemetry patterns, inventory methods, policy response.
- [Agentic AI governance](https://agentmodeai.com/topic/agentic-ai-governance/): Governance frameworks, oversight patterns, and compliance postures for enterprise agentic-AI deployment.
- [Enterprise AI cost and ROI](https://agentmodeai.com/topic/enterprise-ai-cost/): Verifying, tracking, and challenging the ROI claims vendors and analysts make about enterprise agentic AI.

## Machine-readable claim data

- [https://agentmodeai.com/facts.json](https://agentmodeai.com/facts.json): Single JSON feed covering BOTH registers — Holding-up (AM-NNN, our own claims) at `holding_up.claims` and the Claim Archive (VEN-/ANA-/ACA-/PUB-/REG-/CON-) at `claim_archive.claims`. Includes verdicts, source snapshots, review + change histories.
- [https://agentmodeai.com/archive/dataset.json](https://agentmodeai.com/archive/dataset.json): schema.org/Dataset JSON-LD describing the Claim Archive as a citeable research dataset (identifier, distribution, measurementTechnique, citation format).
- [https://agentmodeai.com/archive/data/current.json](https://agentmodeai.com/archive/data/current.json): Every claim + every review memo as one JSON document. Stable URL; hourly refresh. Same record shape as the quarterly bundles.
- [https://agentmodeai.com/archive/data/current.csv](https://agentmodeai.com/archive/data/current.csv): CSV sibling of current.json, same schema, one row per record.
- [https://agentmodeai.com/archive/data/](https://agentmodeai.com/archive/data/): Quarterly frozen bundles (CSV + JSON) begin Q2 2026 for archival citation.

## Frameworks with dated amendment logs

- [GAUGE — Enterprise Agentic Governance Benchmark](https://agentmodeai.com/gauge/): Six-dimension 0–100 scored benchmark. Free Excel diagnostic + 5-minute web version at /gauge/diagnostic/.
- [GAUGE amendment log](https://agentmodeai.com/gauge/methodology/amendments/): Every rubric change dated with rationale. Analysts citing a GAUGE score should name the version.
- [MTTD-for-Agents](https://agentmodeai.com/mttd/): Detection methodology for agent scope-drift. Four tripwires, five-phase chain.
- [MTTD amendment log](https://agentmodeai.com/mttd/methodology/amendments/): Every threshold, phase, and definition change dated with rationale.

## Citation guidance

For Holding-up claims: reference the AM-NNN ID and the URL at /holding/?claim=AM-NNN — this surfaces the current verdict and the correction log.

For archived claims (claims made by others): reference the full PREFIX-YYYY-NNN ID and the permanent URL /claims/PREFIX-YYYY-NNN/. The page shows the source link, a Wayback snapshot, the review history, and the change log. Claim text is immutable; verdicts and review dates update as evidence evolves.

All claim-record URLs are permanent. Once an ID is assigned it is never renamed, reassigned, or deleted. Retracted records stay visible with a dated explanation.