AI agent governance: analysis and sources

Every article starts from a text you can open: a regulation, an opinion, a study. Sources are dated and listed at the end of each article, so anyone can check before quoting.

Latest article, governance

95% of organisations have already had an AI agent incident, says Gartner

The survey Gartner released on 16 November puts a number on what many teams see without saying: agents already cause incidents, while most leaders believe they are under control.

, 2 min read

95%

of organisations surveyed faced at least one agent-related incident in 12 months

All articles

25

organisations interviewed; none lets its agents act with unrestricted autonomy

Governance

AI agents at work: what the OECD learned from 25 organisations

The paper the OECD published on 16 September 2026 does not rest on vendor promises but on interviews with teams that have agents in production. What they describe amounts, almost point by point, to the list of problems agent governance has to solve.

, 2 min read

4 levers

transaction caps, budgets, approved counterparties and escalation rules, per Circle

Agentic economy

Agentic economy: agents that spend within delegated permissions

Circle, the issuer of USDC, defines the agentic economy in terms that speak directly to finance teams: agents acting within delegated permissions, often without a person approving each step.

, 1 min read

5 degrees

of autonomy in the CIANum scale the CNIL refers to

Personal data

Agentic AI: what France's data protection authority says about agents and personal data

No definition of agentic AI commands agreement, the CNIL notes. It nevertheless agrees with other authorities on one criterion, autonomy of action, and describes risks that stem less from the technology than from the delegation of power.

, 3 min read

14 Jul 2026

the x402 Foundation starts at the Linux Foundation to standardise agent payments

Agentic commerce

Payments made by an AI agent: who answers when an order goes wrong?

In eighteen months, card networks, Google, OpenAI, Stripe and Coinbase published what an agent needs to pay. These tools prove what was authorised. They do not say what the company was entitled to let its agent decide, nor who answers for a purchase nobody really wanted.

, 3 min read

2 Aug 2026

the transparency obligations of Article 50 apply

Regulation

AI Act Article 50: what your agents must disclose since 2 August

This is the AI Act obligation the Omnibus did not postpone. It looks simple; it still requires knowing where, when and to whom each agent speaks.

, 2 min read

84%

of the global AI agent sector held by OpenAI, Google and Anthropic in May 2026, per Sensor Tower as cited in the opinion

Agentic economy

AI agents: what France's competition authority found in its opinion of 17 July 2026

A little over 100 pages, a definition of agentic commerce and 6 recommendations: opinion 26-A-05 is the most complete public French document on the place AI agents are taking between companies and their customers.

, 3 min read

4

dimensions to describe an agent: autonomy, efficacy, goal complexity, generality

Governance

Agentic profiles: what the journal Nature proposes for governing agents

Rather than deciding whether a system is an agent or not, two researchers propose to describe it along 4 graded dimensions. The idea is useful; it leaves untouched the question of the power entrusted to the agent.

, 2 min read

2 Dec 2027

high-risk rules start to apply to the sensitive areas listed in Annex III

Regulation

The AI Act after the Omnibus: what applies to AI agents, date by date

Published in the Official Journal on 24 July 2026, the Digital Omnibus on AI moves some AI Act deadlines and leaves the rest untouched. For a company running agents, the question is no longer whether the Act applies but which of its decisions it targets, and from when.

, 4 min read

14 Feb 2024

a Canadian tribunal holds Air Canada responsible for what its chatbot told a customer

Liability

Who is responsible when an AI agent decides?

In February 2024 Air Canada argued that its chatbot alone answered for what it told customers. The tribunal refused. The rule that follows is old and simple: a company answers for what the agent it puts into service does, which forces it to know what it has entrusted to that agent.

, 3 min read

2018

the GDPR requires an independent DPO reporting to the highest management level

Organisation

Data needed a DPO. Who will answer for what agents decide?

Before 2018, few companies had a person in charge of data protection. The GDPR made one mandatory in some cases and gave the role a status. AI agents now raise a question of the same kind.

, 2 min read

7 Dec 2023

the CJEU holds that a credit score can be an automated decision

Personal data

When does an agent decide under the GDPR? What the SCHUFA ruling says

Many companies think they escape Article 22 of the GDPR because a human signs at the end. The SCHUFA ruling shows the question turns not on the signature but on the real weight of what the machine produced.

, 2 min read

50+

definitions of the agentic economy published between 2021 and March 2026

Agentic economy

Agentic economy: more than 50 definitions, one dividing line

The term "agentic economy" is everywhere and does not mean the same thing everywhere. Two pieces of work help sort it out: Microsoft Research in 2025 and a taxonomy published at the end of March 2026.

, 2 min read

9 Dec 2025

MCP joins the Linux Foundation’s Agentic AI Foundation

Standards

MCP and A2A under open governance: what it changes for companies

The two protocols that connect agents to tools and to each other no longer depend on a single vendor. That is good news for interoperability. It says nothing about what an agent is entitled to do once connected.

, 2 min read

22 Jan 2026

Singapore presents its agentic AI governance framework in Davos

Governance

Singapore publishes a governance framework for agentic AI

Singapore presents its text as the first comprehensive guide to deploying agents responsibly. It sets a clear rule: humans remain ultimately accountable.

, 2 min read

40%

of agentic AI projects canceled by the end of 2027, per Gartner

Market

40% of agentic AI projects canceled by 2027: what Gartner actually says

The prediction dates from June 2025 and still circulates. It deserves more than a passing quote: the 3 causes Gartner gives show where the fate of agent projects will be decided.

, 2 min read

RSS feed