How do banks check a Claude Partner Network partner? Tiers, certification and what to ask

What are the Claude Partner Network tiers?

Anthropic publishes the Services Track requirements on its public page "Services Track and Claude Partner Hub" (3 June 2026, anthropic.com/news/services-track-partner-hub); the figures below are quoted from that page only. Anthropic states that every firm is measured against the same requirements regardless of size and that standing is verified quarterly. Requirements as published on that page on 26 October 2026 (Anthropic may change them):

Swipe to compare →

How Anthropic defines the measures: a certified practitioner is a staff member with a current Anthropic certification who has used Claude in the past 90 days; a customer in production is one the firm has taken live with Claude; a public endorsement is a customer willing to vouch for the firm in a published customer story.

Dates Anthropic publishes: promotions on 1 January and 1 July (plus an extra review on 1 October 2026 in the program's first year); downgrades only at the annual review on 31 December, with 90 days' notice. Anthropic also says referral credit and deal protection are tracked separately from tier standing, and that getting started is free.

What does each tier mean for a bank buyer?

These readings are ours, drawn from the criteria above, not statements by Anthropic.

• Registered: the firm has joined and committed to certifying staff. The criteria do not include customer deployments, so a Registered tier says little about delivery in production.
• Select: at least two joint customers in production in the last 12 months and one public story, with at least 10 active certified people. Ask whether any of those deployments are in regulated financial services; the tier does not say.
• Preferred and Global Premier: larger certified teams and more deployments, across regions at Global Premier. Scale is not the same as fit: a large firm may staff your project differently from its headline numbers, so ask who is named on your team.
• No tier tells you about industry specialisation (Anthropic lists industry and use-case specializations as upcoming), information security, data handling or regulatory experience. Those you must check yourself.

What does Claude certification actually certify?

Anthropic says certifications belong to individuals, not firms, and are earned through exams at the Anthropic Partner Academy. A firm being "certified" therefore means specific people hold certifications; the practical question is whether those people will work on your project. In its June 2026 announcement Anthropic reported more than 10,000 consultants holding a Claude certification.

How can a bank verify a partner's claim?

Anthropic says the Claude Partner Hub shows partners their standing against the published requirements, refreshed daily, and that a public directory shows each partner's tier, certified team, customer deployments and public references. On 26 October 2026 the directory at partnerhub.claude.com/directory offered a Partner Tier filter (Global Premier, Preferred, Select) and filters including Industry Expertise (with a "Financial Services & Insurance" option), Service Delivery Location and Deployment Surface; the listings we saw were large global firms, all marked Global Premier, and the page offers a "View more" option. We could not see certified-people counts on the cards. The directory is dynamic and may change.

So: look up the firm in the directory, compare the claimed tier with the listed tier, and ask for a reference you can call. Absence from the directory is a question to ask, not proof of anything; a firm may be new or not yet listed (the Partner Tier filter we saw offers Global Premier, Preferred and Select, not Registered).

What should a bank or wealth manager ask a Claude partner?

A checklist for CIO, COO and procurement teams in Switzerland, Germany and the US. Not legal advice.

1. Tier and standing. Which tier, and what is the listing in Anthropic's directory? When was it last reviewed (quarterly verification)?
2. Certified individuals. How many hold a current certification and have used Claude in the last 90 days, and which named people (with certificates) are on our team?
3. Production references. How many customers are live in production, and in regulated financial services? Can we speak to one? Public stories are only a subset; some banks cannot be named.
4. Model route and data residency. Which route does the partner propose (Anthropic API, Amazon Bedrock, Google Cloud (Gemini Enterprise Agent Platform, formerly Vertex AI)) and which region or endpoint setting? Ask for the vendor documentation behind the answer. Residency differs by route and model and changes; for a Swiss view see Claude for Swiss banks: data residency, FINMA and client data.
5. Security and privacy management. Is the partner's own ISMS certified (for example ISO/IEC 27001 and 27701), what is the scope, and can you see the certificate? A CPN tier is not a substitute.
6. Regulatory fit, per jurisdiction. Switzerland: FINMA Guidance 08/2024 on AI governance, FINMA Circular 2018/3 on outsourcing, banking secrecy (Art. 47 BankA), nFADP. Germany and EU: EU AI Act, DORA third-party risk (Regulation (EU) 2022/2554) and your national supervisor's requirements. US: the Interagency Guidance on Third-Party Relationships: Risk Management (Federal Reserve SR 23-4, FDIC and OCC, 7 June 2023). Ask how the partner supports your assessment; the bank stays responsible. See EU AI Act compliant AI agents for banks.
7. Governance and delivery. Observability, audit trail, human approval steps, testing, exit and portability if the model changes. See AI agent development for banks and AI pilot to production for banks.
8. Contract and subcontractors. Who is the contracting entity, which subcontractors and sub-providers touch data, and what audit rights does the bank get?

Where does S-PRO stand?

S-PRO is a Certified partner in the Claude Partner Network. Our team includes Claude-certified engineers and architects. S-PRO is a software development company headquartered at Löwenstrasse 45, 8001 Zürich, founded in 2014, with 250+ engineers, R&D centers in Zurich and Łódź, and ISO 27001 and ISO 27701 certification. We build AI agents for banks, wealth managers and fintechs with human approval and audit logging, and agree the model route and data-residency options per project. On this page S-PRO makes no claim about a specific CPN tier or directory listing; ask us and check Anthropic's directory the same way you would for any partner. To discuss a Claude project, see Claude implementation partner for financial services.

Book a discovery call (45 min)