Servatius von Tatzenberg reads the conference at today's article, and covers both executive appointments — Alain Girard (Banks, arriving from recovery and resolution exactly as the Banking Act dispatch lands) and Hedwig Ulmer Busenhart (Insurance, with an actuarial depth that signals SST model scrutiny is rising). Read the three pieces in sequence: FINMA has simultaneously published its priorities, staffed the two divisions carrying most of the TBTF and SST workload, and left the parliamentary lift to the Bundesrat. The compartmentalisation is deliberate, and the seams are where the interesting questions sit.

The partial revision of the AMLO-FINMA launched 12 May and closes 9 June — eighteen working days. We mapped the three live clauses last week at GwV-FINMA: Drei Klauseln: beneficial ownership transparency, complex-structure documentation, and the high-risk relationship standard. None are cosmetic. The beneficial ownership clause shifts documentary burden in a way that requires process changes, not just an updated memo. FINMA does not always move in response to consultation feedback, but the record of having engaged is useful when the final text arrives and you need to explain the delta to your supervisory board. Assign someone today — the deadline is closer than it looks.

Six sanction list updates arrived via FINMA over the past several weeks: Sudan, Taliban, ISIL/Al-Qaeda, Russia (continued 19th-package follow-through), Ukraine, and Iran. All routine; all require screening system updates. The Iran update needs checking against the Totalrevision framework that replaced the JCPOA-era structure in December 2025 — the annexes changed substantially, not just the names — covered at Iran-Verordnung and Venezuela SRVG for the adjacent screening gap.

Servatius von Tatzenberg's piece on the Magnitsky pattern explains why routine compliance is not the same as the system working: the structural gap between MROS receiving a suspicious activity report and the Bundesanwaltschaft acting on it means that list updates can reach your screening software faster than they reach the competent prosecutor. Worth a conversation with your AML officer before you sign off the next attestation.

Casimir von Firn covers the full implications at today's article. The ruling travels beyond pharmacies: if your platform's transaction data can reveal a health condition — supplements, medical devices, certain insurance products, wearable subscriptions — those purchase records are now squarely within the special-category regime. The accountability question is not the cookie banner; it is the processing basis for a category of commercial data that many companies have been treating as ordinary purchase history. Start the review there.

Casimir von Firn's piece is at today's article. Geneva is not Zurich, and a single EU-style resolution is not available in Switzerland's cantonal labour law structure. But the Geneva reasoning is now in the inter-cantonal record that other courts will cite when the same question reaches them.

C-133/24 is a football case with a non-football holding. The Court found that a no-poach agreement among Portuguese clubs during Covid-19 restricted competition by object — authorities need to show the agreement existed, not that it caused harm. The pandemic framing provided no defence; emergency rationale did not convert a horizontal labour-market restriction into something permissible.

For in-house teams that signed sector-wide labour-market coordination agreements in 2020–2021 under emergency rationale: check whether those agreements are still live and whether they survive by-object scrutiny from a national competition authority. The Commission has been actively interested in no-poach arrangements across multiple sectors since 2023, and C-133/24 gives it cleaner precedent than it had. The question is not limited to tech sector talent wars — it arises wherever competing employers coordinate on recruitment, directly or through industry associations.

C-797/23 puts the French and German neighbouring-right laws on firmer ground than they have ever had. Switzerland is not bound by the DSM Directive, but Swiss media companies distributing into EU markets — and platforms aggregating Swiss press for EU readers — are operating in a market where the legal baseline just shifted. The Court's reasoning on "fair compensation" is expansive enough that systematic aggregation of protected snippets is in scope, not just full reproduction.

Licensing negotiations deferred on the assumption that platforms would ultimately prevail need to be reopened. The platforms have now lost at the top European court; the uncertainty they have been using as a negotiating posture is considerably narrower than it was last week.

Four more pieces from today's articles worth noting. Casimir's Commission v Italy (C-155/25): the infringement mechanism — not soft guidance — is now the Commission's stated tool for member states tolerating cascading fixed-term contracts without objective justification; EU-subsidiary HR teams should map their current exposure. Casimir's T-24/25 Albert René: the General Court has tightened the reputation evidence standard for famous marks in EUIPO proceedings — if your vintage mark's file lacks temporal continuity evidence, audit what is actually there before the next cancellation challenge arrives. Casimir's C-683/24 Spielerschutz Sigma: read AG Emiliou's opinion for the admissibility signal — the ECJ is telling national courts that not every regulatory dispute belongs in Luxembourg, which has practical value when you are deciding whether to push a national dispute toward a preliminary reference or resolve it domestically. And Servatius's Art. 322septies StGB: the piece on when the Rolex becomes a bribe is precise about where the threshold sits — most gifts-and-hospitality policies assume a higher bar than the statute actually sets.

Casimir von Firn has two pieces today that belong together: the federal eID and KYC architecture (adopt early or wait?) and AI Act Annex III for Swiss vendors (does your system hit the high-risk classification?). Both are infrastructure decisions with long lead times. FINMA's revised Circular 2016/7 on video and online identification is part of the same stack — the QES route is losing its privileged status, and the eID is built on what replaces it. Make both calls now, before the compliance deadline announces itself in a supervision letter.

The news is the deployment model, not the tool: one platform, all practice groups, global rollout — not a pilot. Your outside counsel is now building institutional memory in a system you have no visibility into. Most engagement letters predating 2024 do not address how a firm's AI platform interacts with client confidentiality; that gap belongs in the next version of your standard template. This connects to the AI Act Annex III piece above: legal AI tools deployed in regulated contexts are not categorically exempt from the high-risk framework when they inform EU-market decisions. Swiss law-firm clients sit inside that exposure whether or not they are operating the tool themselves.

The Supreme Court's 2023 ruling that litigation funding agreements are damages-based agreements has not been reversed, and the promised corrective legislation has not arrived. Swiss companies with English-law dispute resolution clauses or funded proceedings in England need a current enforceability assessment from UK counsel — do not assume the structure is sound because it was drafted before PACCAR.

On the wider theme: Casimir's piece on judicial independence as enterprise risk is the clause-drafting question for contracts designating courts in jurisdictions where the independence baseline has shifted since the contract was signed. Servatius's piece on corporate funding of Swiss universities adds a parallel governance gap — the disclosure problem has a liability tail that does not appear in standard research-agreement templates. Both are items for the annual legal audit, not the boilerplate review queue.