AI Regulation Tracker / Enforcement and investigations
Italy's AGCM Opens Microsoft 365 Copilot AI Bundling Probe
On June 26, 2026, Italy's competition authority opened a consumer-protection investigation, case PS13129, into whether Microsoft made it clear enough that its Microsoft 365 subscription had been integrated with the Copilot and Designer AI services alongside a price increase. This is an opened probe, not a finding. No violation has been established.
On June 26, 2026, the AGCM announced in Rome that it had opened a consumer-protection investigation, case PS13129, into a price increase on the Microsoft 365 subscription service. The named parties are Microsoft Ireland Operations Ltd. and Microsoft S.r.l. What triggered the case is not the price rise on its own. It is how that price rise was tied to a change in what the subscription now includes, and whether Microsoft told its customers about that change clearly enough.
What the authority is actually looking at
According to the AGCM, the concern is that Microsoft may have failed to make it sufficiently clear that its subscription service had been integrated with the Copilot and Designer artificial-intelligence services. In the authority's own words, the concern is that Microsoft "may have failed to make it sufficiently clear that its subscription service had been integrated with the 'Copilot' and 'Designer' artificial intelligence services." The AI bundling sits at the center of the case. New AI capabilities were folded into the product, and the price went up to match, and the question the regulator is asking is whether the average consumer understood what was happening.
The authority goes further on the manner of the communication. It says the way the information was communicated "may also constitute an aggressive practice, as it appears to have unduly restricted consumers' freedom of choice." Under Italy's Consumer Code, an aggressive practice is a specific category, and it turns on whether a trader's conduct limited a consumer's ability to make a free, informed decision. The default-opt-in structure is the reason that framing comes up. Secondary reporting describes a setup where the higher price would apply on renewal unless the customer took action to avoid it, which is exactly the kind of default that draws this sort of scrutiny.
The price figures themselves have circulated in Italian press coverage rather than in the AGCM's own text. According to secondary reporting, the Personal plan moved from about 69 euros to about 99 euros a year, and the Family plan from about 99 euros to about 129 euros. Those numbers are worth citing carefully. They come from reporting, not from an AGCM finding, and the authority's case is about disclosure and the manner of communication, not about whether a particular euro figure was too high.
What this is, and what it is not
I want to be precise, because the word "investigation" gets read as "guilty" far too often.
This is an opened probe. The AGCM has decided the facts are worth examining and has started a formal process. It has not concluded that Microsoft broke any law. It has not found an aggressive practice. It has not imposed a fine. Microsoft has not been found liable of anything. Everything the authority has said so far is framed in the conditional: the company "may have" failed to be clear, the communication "may" constitute an aggressive practice, it "appears to have" restricted choice. That language is doing real work. It marks the difference between a suspicion the regulator is testing and a conclusion the regulator has reached.
It is also worth being clear about which body of law is in play. This is a consumer-protection case under Italy's Consumer Code. It is not an enforcement action under the EU AI Act. The AI element is the trigger and the subject matter, the thing that changed inside the product, but the legal hook is old-fashioned consumer law about transparency and fair commercial practices. That distinction matters if you are trying to understand what signal this sends. It is a consumer-disclosure signal that happens to involve AI, not an AI-specific regulatory ruling.
Why a US SaaS vendor should care
The direct exposure here belongs to Microsoft entities in Italy. But the relevance for US software companies is direct, and it does not depend on the outcome.
The pattern the AGCM is examining is common. A company has an existing paid consumer subscription. It adds generative-AI features, Copilot-style assistants, image tools, and the rest, and it raises the price to reflect the new value. Then it lets the higher price take effect on renewal by default. That is a reasonable-sounding commercial move, and a lot of vendors are making some version of it right now. What this case establishes is that at least one national regulator will look hard at whether the customer was told, plainly, that the product changed and the price changed with it, and whether the default structure gave the customer a real choice.
If you run a SaaS product with EU consumer subscribers and you are bundling AI features into existing plans, the practical read is simple. Disclosure is the exposure. It is not enough that the new features are good or that the price is fair. The question a regulator will ask is whether an ordinary customer understood the change before the money moved, and whether saying no was as easy as saying nothing. That is the live regulatory-scrutiny signal this opening creates, regardless of where PS13129 eventually lands.
What to do now
Read the case for what it is, an opened investigation into disclosure, not a ruling against AI upselling. Look at your own bundling: if you have added generative-AI features to an existing paid subscription and moved the price, check how that change was communicated and whether the higher price applies by default on renewal. Make the disclosure explicit and make it early, so a customer sees what changed and what it costs before a renewal charge lands. Give a real opt-out, since the aggressive-practice concern here turns on whether declining was as easy as accepting. And watch case PS13129 for its outcome rather than treating today's opening as a verdict, because how the AGCM resolves it is what will actually tell you where the line falls.
Questions professionals are asking
Has Microsoft been found to have broken the law?
No. On June 26, 2026, the AGCM opened an investigation, case PS13129. Opening a case means the authority is examining the facts. It has not found a violation, has not determined liability, and has not imposed a fine. Microsoft has not been found liable of anything.
What exactly is the AGCM investigating?
Whether Microsoft made it sufficiently clear that its Microsoft 365 subscription had been integrated with the Copilot and Designer AI services, alongside a price increase, and whether the way that information was communicated unduly restricted consumers' freedom of choice. The authority frames the latter as a possible aggressive practice under Italy's Consumer Code.
Is this an AI Act enforcement action?
No. It is a consumer-protection case under Italy's Consumer Code (Codice del Consumo). The AI bundling is the trigger and the subject matter, but the legal basis is consumer law about transparency and fair commercial practices, not the EU AI Act.
Are the price figures official AGCM findings?
No. The figures reported in Italian press, Personal from about 69 to about 99 euros and Family from about 99 to about 129 euros, come from secondary reporting, not from the AGCM. The authority's case is about disclosure and the manner of communication, not about a specific euro amount.
Why should a US SaaS company care about an Italian case?
The relevance is direct. Any US vendor bundling generative-AI features into an existing paid EU consumer subscription with a default-renewal price increase now faces a live regulatory-scrutiny signal for consumer-protection review of AI upselling and disclosure. The exposure is whether customers were told plainly and given a real choice, not whether the AI features or the price were reasonable.
RELATED BRIEFINGS
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. An opened investigation is not a finding of wrongdoing. Confirm how any law or investigation applies to your situation with qualified professionals in the relevant jurisdiction.