How to read the traffic light

    Your industry: Wealth managers and financial service providers

    Tailored to wealth managers and financial service providers: FINMA principles (notice 08/2024), the outsourcing rules (Art. 17 FinIO for wealth managers, Circular 2018/3 for banks and insurers) and professional secrecy are factored into every row.

    Green: proven and recommended

    Works with client data: the setup is proven, contract and governance are in place. This is exactly how you should do it.

    Yellow: unclear, depends on the setup

    Depends on the contract, configuration and industry: feasible with a documented risk assessment, not without it.

    Red: not possible

    Not defensible for client data in our risk assessment: no contract, no control or no documented safeguards. Often still the best choice for research, marketing and anonymised content.

    For wealth managers, green only applies when three conditions are met: the outsourcing is documented under Art. 17 FinIO, audit rights for the supervisory organisation and FINMA are contractually secured, and professional secrecy is technically protected. If any of these is missing, every green option turns yellow.

    All AI options at a glance

    Filter the table by traffic-light colour. On a smartphone it scrolls sideways.

    SolutionClient dataModelsData locationModel qualityEffortCostWhat to watch
    Azure OpenAI Switzerland (Standard)
    Microsoft Azure, Switzerland North region
    GreenGPT-4o (until April 2027), Embeddings
    Processing
    Switzerland
    Storage
    Switzerland
    Solid, several generations behind the frontierMedium, integration by IT or partnerLow, per token, no base feeThe reference setup under FINMA logic: Swiss data location, choose the Standard deployment type (not Global or Data Zone), outsourcing contract under Art. 17 FinIO or Circular 2018/3 with confidentiality and audit rights. For new customers, GPT-4o (until April 2027) is effectively the only option. Apply for Modified Abuse Monitoring: without this approval Microsoft stores flagged samples, and authorised staff can review them even in the Swiss region.
    Azure OpenAI Switzerland (Provisioned)
    Microsoft Azure, Switzerland North, reserved capacity
    GreenGPT-5.1 (until May 2027)
    Processing
    Switzerland
    Storage
    Switzerland
    Very good, strongest model with CH guaranteeMedium, integration by IT or partnerFrom approx. USD 13,000 per month (minimum 50 PTU)Like the Standard variant, but with regional Provisioned capacity in Switzerland North (not Global or Data Zone Provisioned) and GPT-5.1 (until May 2027); include the capacity contract in the risk assessment per supervisory notice 08/2024. From approx. USD 13,000 per month (minimum 50 PTU): unrealistic for most wealth managers, and a successor in Switzerland is not guaranteed.
    Swiss AI platforms
    e.g. AIgent, Swisscom, Phoenix Technologies (kvant), Infomaniak
    GreenOpen models such as DeepSeek, Qwen, GLM, Apertus
    Processing
    Switzerland
    Storage
    Switzerland
    Good to very good, a step behind the frontierLow, ready-made interface, some with agentsMedium, licence per user, sometimes plus usageThe simplest FINMA-compliant surface without US ties: check ISO 27001, references and confidentiality, disable optional US models.
    AI in your industry software
    e.g. practice, law firm, CRM or portfolio systems
    Depends on providerUsually GPT or Claude in the background
    Processing
    Depends on provider
    Storage
    Depends on provider
    Good, tailored to the use caseLow, the system is already in useLow to medium, often as an add-on moduleClarify in writing which model runs where. The outsourcing remains your responsibility even with an industry vendor (Art. 17 FinIO or Circular 2018/3).
    Microsoft 365 Copilot
    Microsoft, with EU Data Boundary
    YellowGPT models from OpenAI, operated by Microsoft
    Processing
    EU/EFTA
    Storage
    Switzerland (Data Location or ADR)
    Very good, strong in everyday office workLow, licence and configurationMedium, licence per user and monthPermissible with safeguards: Flex Routing off, Anthropic models off, permissions cleaned up. Document the contract and client information.
    Azure OpenAI Data Zone EU
    Microsoft Azure, resource in Switzerland North
    YellowLatest GPT generation, e.g. GPT-5.4 to GPT-6.1
    Processing
    EU Data Boundary, confirm countries with Microsoft before use
    Storage
    Switzerland
    Frontier, latest GPT modelsMedium, integration by IT or partnerLow, per tokenFINMA permits processing abroad with safeguards: risk assessment, contract, client information and an explicit assessment of the residual CLOUD Act risk. Feasible with our reference architecture: client data stays in Switzerland, only pseudonyms are processed in the EU.
    Claude via AWS Bedrock (EU)
    Amazon Web Services, EU regions
    YellowOpus 5 and Sonnet 5 in Ireland or Stockholm, Opus 5.5 and Sonnet 5.5 via EU inference
    Processing
    EU
    Storage
    No prompt storage by Bedrock
    Frontier, especially strong for text and analysisMedium to high, AWS account and development neededLow to medium, per tokenLike Azure Data Zone EU; also cover AWS as a sub-processor in the contract. AWS evidences the FINMA requirements with a dedicated ISAE 3000 Type 2 report.
    Gemini via Google Vertex AI (EU)
    Google Cloud, EU regions
    YellowGemini models
    Processing
    EU
    Storage
    EU or Switzerland
    Frontier, strong with long documentsMedium to high, Google Cloud project and developmentLow to medium, per tokenLike Azure Data Zone EU; have the processing location verified in writing before use. Google evidences the FINMA requirements with an ISAE 3000 report.
    Claude directly from Anthropic
    Claude API, Team and Enterprise
    RedClaude Fable 5.1, Opus 5.5, Sonnet 5.5
    Processing
    US or global
    Storage
    US
    Frontier, newest Claude models firstLow, ready to useMedium, licence or per tokenUS processing without documented safeguards: unsuitable for client data in this configuration. Breaching professional secrecy (Art. 69 FinIA, for banks Art. 47 Banking Act) is a criminal offence, even through negligence with a fine of up to CHF 250,000.
    Claude in Microsoft Foundry or Copilot
    Microsoft, operated on Anthropic infrastructure
    RedClaude models
    Processing
    Outside the EU Data Boundary
    Storage
    Not guaranteed
    Frontier, same as Claude directLow, switchable at a toggleMedium, in licence or per tokenProcessing outside the EU Data Boundary: unsuitable for client data, keep the switch off. Here too, breaching professional secrecy is a criminal offence.
    Copilot with Flex Routing enabled
    Microsoft 365, setting available since April 2026
    RedGPT models
    Processing
    At peak load also US, Canada, Australia
    Storage
    Switzerland
    Very good, same as CopilotLow, a single settingMedium, licence per user and monthPeak-load processing in the US, Canada, Australia: disable for FINMA-supervised client data; one setting suffices. Without it you risk a criminal breach of professional secrecy.
    Private AI accounts
    ChatGPT, Claude.ai, Gemini with a private or free account
    RedAll of them
    Processing
    Unclear, mostly US
    Storage
    US
    Frontier, which makes it so temptingNone, and that is exactly the problemLow, free or privately paidA clear breach of professional secrecy (Art. 69 FinIA): a criminal offence, even through negligence with a fine of up to CHF 250,000. Prohibit by policy, offer an approved alternative.

    Note on the US CLOUD Act

    Microsoft, AWS and Google are US corporations and are subject to the US CLOUD Act, even with data centres in Switzerland or the EU. This residual risk belongs in every risk assessment.

    The best solutions: three proven setups for wealth managers and financial service providers

    Hardly any organisation needs just one tool. What has proven itself is the combination of an AI employee for daily work and a secure interface for automations.

    For FINMA-supervised organisations we recommend the pragmatic setup: Copilot for daily work, sensitive client data only through the Swiss pipeline, all under an outsourcing contract under the FINMA rules (Art. 17 FinIO, Circular 2018/3). For new customers, GPT-4o (until April 2027) is effectively the only option there; Provisioned with GPT-5.1 costs from approx. USD 13,000 per month and is unrealistic for most. If you need frontier models, our reference architecture is the answer: client data stays in Switzerland, only pseudonyms are processed in the EU.

    Recommended for most SMEs

    Pragmatic: Copilot plus Azure Switzerland

    • M365 Copilot for emails, documents and meetings, with storage in Switzerland and Flex Routing disabled
    • Azure OpenAI Switzerland North for automations with sensitive client data, such as document review or reports: for new customers, GPT-4o (until April 2027) is effectively the only option there
    • A clear policy: what may go into Copilot, what only goes through the Swiss pipeline
    When policy demands «Switzerland only»

    Strict: Swiss platform plus Azure Switzerland

    • A Swiss AI platform as the working surface for all employees
    • Azure OpenAI Switzerland North (GPT-4o) or Provisioned (GPT-5.1, from approx. USD 13,000 per month) for automations
    • External models in the platform selectively disabled per use case
    When quality matters above all

    Reference architecture: frontier models with client data in Switzerland

    • Azure OpenAI Data Zone EU or Claude via AWS Bedrock EU
    • Pseudonymise where possible before data reaches the model
    • Risk assessment, contract and client information with extra care
    How the reference architecture works

    What data protection requires

    For Swiss organisations, the revised Data Protection Act (DPA) applies: cloud AI is usually commissioned processing, and disclosures of data abroad require, depending on the country, an adequate level of protection, safeguards or an exception (Art. 16 and 17 DPA). Responsibility stays with you, even when outsourcing to a cloud provider. Where sectoral supervision applies in addition, the industry selector by the cloud options shows it.

    • AI inventory: which applications, which data, which model, which data location
    • Risk assessment per application, including cross-border processing and the US CLOUD Act
    • Policy for employees: approved tools, banned shadow AI, human oversight
    • Commissioned processing contract under Art. 9 DPA: confidentiality, audit rights, data location
    • Review disclosure of data abroad: adequate level of protection, safeguards or exceptions under Art. 16 and 17 DPA
    • Privacy policy and information for data subjects kept up to date
    • Records of processing activities maintained under Art. 12 DPA

    What the supervisor additionally requires

    On top of data protection, supervisory rules apply to FINMA-supervised organisations: technology does not release you from responsibility, not even when outsourcing to a cloud provider.

    • Outsourcing documented under Art. 17 FinIO: outsourced function, provider and data location; significant outsourcing notified to your supervisory organisation
    • Professional secrecy under Art. 69 FinIA secured technically and contractually: a breach is a criminal offence, even through negligence with a fine of up to CHF 250,000
    • FINMA supervisory notice 08/2024: AI inventory, risk assessment and policies under the principle «same business, same risks, same rules»
    • FINMA Circular 2018/3 applied analogously: contract with confidentiality and audit rights for your supervisory organisation and FINMA

    This overview does not replace legal advice. Model availability and regions change constantly; October 2026 edition. Verify details with the provider before use.

    Sources

    • FDPIC: Guide to checking the admissibility of transferring personal data to countries without an adequate level of data protection
    • FDPIC: Factsheet on commissioned processing of personal data under the revised DPA
    • Microsoft Learn: Model availability by region in Azure Foundry
    • Microsoft Learn: Deployment types (Global, Data Zone, Standard, Provisioned)
    • AWS: Model availability by region in Amazon Bedrock
    • Google Cloud: Data residency for generative AI
    • FINMA: Supervisory notice 08/2024 on governance and risk management when using AI
    • FINMA: Circular 2018/3 Outsourcing Banks and Insurers
    • Federal Act on Banks and Savings Banks, Art. 47 (banking secrecy)
    • FinIO Art. 17 (outsourcing) and FinIA Art. 69 (professional secrecy)

    We implement the recommended setup for you

    Taskstreams builds and operates the pipeline that fits your policy: Switzerland only with Azure OpenAI Switzerland North, or the frontier model in the EU while your client data stays in Switzerland. Contract and governance are included; you receive finished AI employees for your processes, from document review to client reporting, instead of your own infrastructure project.

    • Processing and storage in Switzerland (Azure Switzerland North)
    • Contract, policies and documentation for data protection and supervision included
    • Fixed-price pilot, then ongoing operations by us