Skip to content

We only publish what we can defend in a vendor meeting. Every claim carries an ID, a review date, and a verdict you can check.

Issue 024 · Week 24 · 2026
Ledger

Agent Mode AI — claim-tracked agentic AI analysis

Newest · Understanding AI

What is AI observability, and why your APM cannot do it

Gartner predicts 40% of AI-deploying organisations will run dedicated AI observability tools by 2028. The reason it needs its own tooling: AI fails semantically — drift, bias, opaque reasoning — while classic monitoring watches infrastructure health.

Read the piece →·Written by Claude, signed by Peter
Signed by
Peter

27 years enterprise IT operations. Global organisation. Major incidents. Editorially independent.

  • 159pieces
  • 262tracked claims
  • 14public retractions
About the editor
Framework · GAUGE

The Enterprise Agentic Governance Benchmark. Six dimensions, scored 0–100. Free 5-minute web diagnostic; 30–45 minute Excel for governance groups.

Score a deployment →
Holding-up · Ledger
Every claim, tracked.
262tracked claims
Most recently reviewed: AM-208Holding
Read the ledger →
Bulletin · Reviews
Quarterly verdict bulletin.
1issues published
Latest: Q2 2026 Claim Review Bulletin: did the publication's first-quarter claims still hold?
Read the latest →
Podcast · Audio companion
Two analysts, one claim per episode.
5episodes live
Latest: What the money is doing · 08:57
All episodes →

Recently reviewed

Three claims most recently re-tested against their primary sources. Status changes log to the corrections page; nothing quietly vanishes.

See the full ledger →
  1. OPS-061HoldingWhat to delegate to AI in a 1-5 person business (and what not to)Reviewed 10 Jun 2026Read article →
  2. OPS-060HoldingAI for Dutch e-commerce in 2026: Bol.com, Shopify, WooCommerceReviewed 10 Jun 2026Read article →
  3. OPS-059HoldingAI vendor red flags for SMBs: 2026 contract patterns to spot before signingReviewed 10 Jun 2026Read article →

Moving this week

5 claims have moved off Holding in the last 7 days. The full correction log is on the ledger.

See all corrections →
  1. AM-021PartialDMAIC for agentic AI deployment: why the 87% / 27% success gap reflects measurement discipline, not methodologyMoved 10 Jun 2026Read article →
  2. AM-024PartialThe unverified citation chain: where enterprise AI decisions actually come fromMoved 10 Jun 2026Read article →
  3. AM-029Not holdingWhy 88% of agentic AI deployments failMoved 10 Jun 2026Read article →
  4. AM-031PartialThe CMU 30.3%: the enterprise agent capability gapMoved 10 Jun 2026Read article →
  5. AM-035PartialThe EU AI Act and agentic AI: what August 2026 actually requiresMoved 10 Jun 2026Read article →
Method · Holding-up

Why this publication has a ledger

Most AI commentary gets paid for being loud about what's new. Almost none gets measured on whether what it said last quarter still holds this one. That is the gap this publication exists to close. Every published argument carries an ID, a review date, and one of three verdicts — Holding, Partial, or Not holding — that updates over time as evidence accumulates. The verdict log is the product.

When a claim stops holding, the page says so. The original sentence stays visible. The correction is dated and appended. Nothing is quietly removed. You do not need to trust the author to trust the verdicts — the receipts are public, on a 30–90 day review rhythm, and the corrections record is permanent.

Two registers

Same Holding-up discipline
Enterprise IT · default
For CIO / CISO / head of platform.

Mid-market and large enterprise. Procurement, governance, EU AI Act, multi-vendor agentic stacks. 30–90 day claim review cadence.

159enterprise articles
Start here →
Operators · sibling
For solo founders to ~50-person teams.

No IT department. Practitioner-advisory voice; faster 30–45 day cadence. Tools, vendor red flags, hours-per-week evaluation budgets.

83operators articles
Operators →

Topic pillars

Five clusters

Editor's picks

One per topic cluster

Latest pieces

Full archive →
Understanding AI

What is agent washing, and how do you test for it

Gartner assesses only about 130 of the thousands of self-described agentic-AI vendors as delivering real capability, while more than 80% of organisations intend to deploy within two years. That gap is the agent-washing window, and the defence is a capability test, not a label.

5 min
Use Cases

Agentic AI lands in banking, and it starts with AML

FIS and Anthropic shipped a financial-crimes agent with BMO and Amalgamated Bank in development; Lloyds runs a 40,000-licence Copilot estate at 97% active use. Banking's first production agents compress the investigation, and keep the human on the filing.

5 min
Latest AI Developments

The 2 Jun White House AI order: what it actually requires

The 2 Jun AI executive order leans on voluntary frontier-model review but hard-wires the federal side: CISA binding directives and an NSA/CISA AI clearinghouse, both within 30 days.

4 min
Risk & Governance

Shadow AI discovery: the visibility you think you have

82% of enterprises found an AI agent they did not know was running, while 68% believed they had strong visibility. The gap is the finding, and it makes discovery, not policy, the binding first control.

5 min
Understanding AI

Everyone is buying the agent access graph

Zscaler bought Symmetry, Snowflake bought Natoma, Microsoft priced Agent 365. In five weeks, three infrastructure giants targeted one layer: the map of which agent touches which data.

4 min
Risk & Governance

The non-human identity governance vacuum

Machine and AI-agent identities now outnumber humans about 45 to 1, and most enterprises have no policy to provision or retire them. NHI is the fastest-growing unmanaged attack surface, and the binding control is inventory, not perimeter.

5 min
Business Case & ROI

Microsoft 365 E7 and the new shape of AI licensing

Microsoft's $99 E7 Frontier Suite and the 1 Jul base-price rises move the enterprise AI-licensing decision onto the renewal table, where even the customers who decline Copilot end up paying more.

5 min
Understanding AI

Anthropic's $965B valuation and the vendor question it forces

Anthropic's $965B Series H overtook OpenAI's $852B. The binding risk in a multi-year Claude or GPT contract is no longer model capability; it is pricing power and exit terms.

4 min

Browse by topic pillar

Five strategic pillars

Coming next

Peter's editorial calendar — honest dates, bumped-with-notes if missed.
  1. Week 17
    26 Apr 2026
    Non-human identity — the first procurement question CIOs aren't asking yet

    Every enterprise agent deployment passes through a credential. Most teams still hand the agent a human's credential. Naming the NHI gap is the next Q2 procurement conversation.

  2. Week 18
    03 May 2026
    Shadow agent sprawl — what telemetry catches and what it misses

    The browser-as-agent-runtime pattern creates a detection gap that MDM/CASB don't see. What the first wave of shadow-AI discovery tools actually find, and the three categories they miss.

  3. Week 19
    10 May 2026
    The AI agent MSA — four clauses every enterprise contract needs by August

    EU AI Act enforcement activates 2 Aug 2026. The clauses that survive legal review in the next quarter will be the ones that don't pretend the agent is conventional SaaS.

Vigil · 80 reviewed