Sluis vs Claude for Work for EU organisations

Verdict: For EU organisations, Sluis is the better choice because it is easier to run and richer to work in: people sign in with Microsoft, Google or Apple and build agents by describing them, with versioned Skills, private and shared Knowledge, spreadsheets, presentations, memory, scheduled agents and Claude plus other providers in one workspace. It is also clearer for admins, with one policy, one audit trail per request and one usage-based bill, and safer for data, with EU hosting and EU-only routing enforced while Anthropic states Claude Team and Enterprise data is stored in the US.

"Claude for Work" is the name Anthropic still uses in its privacy documentation for its commercial plans. In practice you buy Team, or Enterprise, which adds heavier security and administration controls. Both give people Claude chat, Claude Code and Cowork under one organisation account. Sluis Workspace is a different kind of product: an employee AI workspace run from the EU, in front of models from several providers, where every request is inspected, routed by residency policy and sealed in an audit chain.

This page is written by Sluis, so weigh our claims accordingly. Every Anthropic statement comes from Anthropic's own pages, checked on 1 October 2026 and listed under Sources. Plans, prices and features change quickly, so where something depends on the plan or the date, we say so.

At a glance

	Sluis Workspace	Claude Team and Enterprise
Operator	Sluis by SevenLab, legal entity 7Lab B.V., Amsterdam	Anthropic, a US company
Data location	EU-hosted managed service	Anthropic states that data is stored in the US
Routing control	EU-only by default, enforced per request; other jurisdictions are an explicit, recorded opt-in	By default traffic may route to the US, Europe, Asia and Australia; Enterprise accounts on usage-based billing can enable US-only inference
Models	Several providers, allow-list per organisation, your own keys	Claude models
Personal data before the model	Detected and reversibly pseudonymised before dispatch	Not documented as a built-in step in the sources we checked
Audit	Hash-chained ledger you can verify offline	Audit logs and Compliance API on Enterprise, per Anthropic
Identity	Entra ID, Google, Apple SSO; SAML and SCIM on request (enterprise); passkeys, authenticator app, organisation-wide 2FA requirement	SSO, domain capture and role-based permissions on Team; SCIM added on Enterprise
Workspace	Agents built by describing them, versioned Skills, private and shared Knowledge, spreadsheets and presentations, scheduled agents, MCP tools	Claude chat, Claude Code and Cowork
Price model	Provider list price plus 10%, no seat fee, prepaid	Seat plus usage; Enterprise is $20 per seat per month plus usage at API rates
Certifications we can cite	ISO 27001 certified	SOC 2 Type 2, ISO 27001, 27017, 27018, CSA STAR, per Anthropic
Best fit for EU organisations	Sluis	Teams that want Claude's first-party apps

Everyday work in the workspace

Sluis is one workspace for the whole job. Sign in with Microsoft, Google or Apple and start; there are no seat licences to assign. Describe an agent in chat and Sluis sets up its instructions, tools and Knowledge. Turn a useful workflow into a versioned Skill your team can run, and bring in private documents or publish shared Knowledge. Attach .xlsx and .pptx files and get an .xlsx or .pptx back; give an agent a .pptx or .docx template so its files follow your house style. Sluis remembers lasting details such as your role and how you like answers (memory can be switched off), runs agents on a schedule (once, daily or weekly), lets agents hand parts of a task to sub-agents when administrators allow it, and offers a reasoning level per chat, image generation when policy allows, and MCP tools. Claude models and models from other providers are available in the same place. Apps cover the web, macOS, Windows, Linux, iOS and Android; the desktop builds are alpha.

Clarity for admins

One policy decides which models, capabilities and jurisdictions are allowed, and it applies to Workspace and API traffic alike, through a gateway built in Rust for predictable latency and memory safety. The console shows one audit row per request with its checks grouped underneath, and workloads, budgets and spend in one overview. The bill is a single usage-based invoice: provider list price plus 10%, with no seat fees.

Where your data goes

Sluis. The managed service is hosted in the EU. Routing defaults to EU-only, and a request that would reach a provider outside your allowed set is refused with a 403 before dispatch. Allowing other jurisdictions is an explicit decision that needs a transfer-terms acknowledgement and is recorded in the audit trail; EU ownership of the provider is a separate, stricter restriction. Claude models are reachable through Sluis via Google Vertex in an EU region (EU region, US-owned provider), or directly from Anthropic after you allow the US jurisdiction. Limits on our side: Workspace web search queries leave the residency boundary, so keep web search off for an EU-only policy, and Agent Harness model traffic (Claude Code, Codex and Cursor on developer machines) is recorded with an unverified region rather than enforced. See security and compliance and residency and models.

Anthropic. Anthropic's privacy center says it uses several cloud providers listed in its subprocessor list, that by default it may route customer traffic to select countries in the US, Europe, Asia and Australia, and that data is stored in the US. Enterprise customers on usage-based billing can enable US-only inference. We found no option in these sources to choose an EU region for storage or inference in the Claude Team and Enterprise apps.

Anthropic's EU regional offering runs through its cloud partners: its regional compliance page describes Europe availability on AWS Bedrock and Google Vertex, with Microsoft Foundry marked coming soon. That serves teams calling Claude through those clouds, not the Claude apps. On Anthropic's own API the documented inference options are global or US-only, and workspace storage is US-only.

Jurisdiction and the CLOUD Act

The US CLOUD Act reaches data in a provider's possession, custody or control wherever it is stored. For a US provider, who controls it matters more than where the servers are. Our CLOUD Act guide sets out a four-factor assessment.

With Sluis, routing is policy: sensitive categories go to EU-owned providers while lower-risk work uses others, enforced per request. Sluis also pseudonymises personal data before dispatch, so a provider sees placeholders. Detection is not perfect, and using Sluis does not by itself establish GDPR compliance. See personal data in prompts.

With Claude's Team and Enterprise apps the provider is a US company and data is stored in the US. That is not disqualifying: you can rate the use and add contractual measures such as the Data Processing Addendum Anthropic publishes. But as far as these sources say, the answer to "can this stay in the EU" is no for those plans.

Cost model

Sluis charges provider list price plus 10% for managed models, or €0.50 per 1M input and output tokens in Sluis fees with your own key (provider charges are separate). There is no subscription and no seat fee, and no free tier: self-service starts with at least €25 of prepaid credit, and payment-method surcharges apply. See pricing.

Anthropic's help center lists Team Standard seats at $25 per member per month billed monthly or $20 billed annually, and Premium seats at $125 or $100, for US customers, excluding tax; Team supports up to 150 seats. Enterprise is listed at $20 per seat per month billed annually, with a minimum of 20 seats, plus usage billed at API rates. Prices vary by region.

Claude Enterprise is already a hybrid of seat and usage. If both bill the same list price per token, Sluis's 10% markup is below a $20 seat fee when a person's monthly usage is under about $200. Above that, Claude Enterprise is cheaper for that person, so heavy daily users favour it. Below it, and for teams with AI agents that need no seat, Sluis is cheaper.

Models, agents, admin control and audit

Sluis connects several providers, including EU-owned ones such as Mistral, Scaleway and Nebul, and Workspace has its own agents and Skills that you build by describing them in a chat. For developers, Sluis speaks the Anthropic Messages API, so Claude Code can point at Sluis with a base URL change. Support for further Anthropic features such as Skills, managed agents and Message Batches is available on request for enterprise contracts.

Claude gives you Anthropic's models and its own agent products: Claude Code, Cowork and connectors to tools such as Microsoft 365 and Slack.

Sluis keeps administration in one console. Owners choose which models members may use, a monthly euro budget per member, and which capabilities are on. The console shows workloads, budgets and spend, and one audit row per request with its checks grouped underneath. Every call goes to a hash-chained ledger you verify offline. Members sign in with Microsoft, Google or Apple, and owners can require two-factor authentication with an authenticator app or passkey. SAML and SCIM are available for enterprise contracts on request. The desktop apps are alpha builds; Workspace itself runs in the browser.

Anthropic's Enterprise plan documents SSO, SCIM, role-based access, audit logs, a Compliance API, customer-managed encryption keys, IP allowlisting and custom retention.

Who might still pick Claude for Work

Teams whose people specifically want Claude with Claude Code, Cowork and Anthropic's connectors on release day.
Organisations for which customer-managed keys, IP allowlisting or a Compliance API are procurement requirements today.
Heavy daily users, for whom a seat plus API-rate usage can be cheaper than a percentage markup, and buyers whose CLOUD Act assessment accepts a US provider with US storage.

For the wider field, see European ChatGPT alternatives for organisations.

Our verdict

Sluis wins for EU organisations. It is hosted in the EU, enforces EU-only routing on every request, pseudonymises personal data before dispatch, and keeps an audit chain you verify offline, while Claude Team and Enterprise store data in the US according to Anthropic. In Sluis one workspace brings chat, agents, Skills and Knowledge together across several providers, Claude models included, on one usage-based bill with no seats. See Sluis Workspace and pricing.

FAQ

Can Claude Team or Enterprise keep our data in the EU?

In the sources we checked, no: Anthropic states that data is stored in the US, and traffic may route to the US, Europe, Asia and Australia by default. Enterprise accounts on usage-based billing can enable US-only inference. EU regions are documented for Claude through AWS Bedrock and Google Vertex.

Is Claude for Work GDPR compliant?

Anthropic publishes a Data Processing Addendum, lists GDPR among its compliance programmes and says it does not train on commercial customer data by default. Whether your use is compliant depends on your assessment, contract and configuration.

Can I use Claude models through Sluis?

Yes. Claude is available via Google Vertex in an EU region, and directly from Anthropic after an owner allows the US jurisdiction. The Vertex route is US-owned, so it is excluded when the EU-owned-only restriction is on.

Does Sluis replace Claude Code?

No. Claude Code keeps running on the developer's machine. Sluis governs the traffic it sends: inspection, routing, audit and budgets, through the Anthropic Messages API or the Agent Harness, where the region is recorded as unverified.

Which is cheaper, Claude Enterprise or Sluis?

If both bill the same list price per token, Sluis is cheaper per person below about $200 of monthly usage, and Claude Enterprise's $20 seat is cheaper above it. Sluis also needs no seat for AI agents, and gives one usage-based bill with budgets per member and a spend overview. Compare real per-token rates, tax and your usage pattern.