Before AI takes off, give it a flight plan.

A managed service in which DeepNet configures, governs, documents, and trains your organization on Claude — delivered as a fixed-scope project by a single engineer. You receive a configured workspace, a written record of your configuration decisions, sample acceptable use guidance, and a live training session.

A service. The Claude subscription is held by your organization. DeepNet implements it, governs it, documents it, and supports it.

Existing DeepNet managed IT clients adopting Claude across a team or an organization. It is designed to be a standard deployment applied consistently across our client base rather than a bespoke build each time.

You can, and many organizations do. What they usually lack afterward is a record of what was decided and by whom. Without that, nobody can say which connector was approved, who can authorize a new integration, what data Claude can reach, or who is accountable for spend. This service produces that record as a deliverable.

No. It sits alongside your managed IT service, and post-deployment support runs through the same service desk you already use.

Scoping, decisions, configuration, documentation handoff, and training and go-live.

The DeepNet-side work is measured in days. Overall duration depends substantially on how quickly your team returns the intake forms, since configuration cannot begin until those decisions are made. We agree a date for their return at the outset.

A single DeepNet engineer delivers the entire scope, including the training session. Your account manager handles the commercial side and steps in if we need help reaching people internally.

We follow up directly, and if we cannot reach the right people we ask your account manager to help. The project simply waits — configuration depends on decisions only your organization can make.

Yes. Seat count and user assignment are confirmed during scoping, and access can be extended later.

A set of decisions, captured through intake forms: who your users are, who owns the workspace, who holds primary administrative rights, who is permitted additional usage, how you want Claude to behave in your organization's context, and your organizational instructions. An engineer walks you through the forms before you complete them.

No. The forms are structured, and an engineer walks you through them and answers questions before you complete them. The decisions are organizational — access, ownership, approvals — not technical.

Someone who can make decisions about access and data handling, and whoever will administer the platform afterward. Both are covered in the training session.

Through your existing identity provider, using the credentials they already use. Claude supports SAML-based single sign-on with providers including Okta, Microsoft Entra ID, and Google Workspace.

No. There is no separate user list to maintain and no new password for staff to manage.

Access follows your existing joiner and leaver process. Automated provisioning and deprovisioning through your identity provider requires the Enterprise plan; Team plans use invite-based or just-in-time provisioning, which we configure during setup.

Yes. Seats, roles, and connected tools remain visible in the admin console and can be adjusted or revoked at any time.

No. Team and Enterprise plans operate under Anthropic's commercial terms, under which your organization is the data controller and Anthropic is the processor. Model training on your content is off by default and cannot be enabled at member level. This is an area where inaccurate third-party commentary circulates, so we point clients to Anthropic's own privacy documentation during setup.

No. Connectors inherit the permissions of the source system. If someone cannot open a file, channel, or record directly, Claude cannot reach it on their behalf either.

File access to SharePoint, OneDrive, or Google Drive is scoped deliberately during configuration, with particular attention to anyone whose existing access spans confidential material. Broad access is a decision your organization makes explicitly, not a default we apply.

Memory is off unless deliberately enabled, and deleted conversations leave the history immediately.

Your designated primary owner holds administrative control of the workspace and can export organizational data. We document who holds that role as part of the configuration record, so it is a known and deliberate appointment rather than an accident of who signed up first.

No. Requests to connect a tool or data source are routed for approval, and higher-risk requests go to the appropriate owner or administrator. Nothing connects on self-approval.

You do. We document your internal approval chain during configuration — who can request an integration, who authorizes it, and what is permitted automatically — and build the workflow around it so both your team and our service desk are working to the same standard.

We provide a sample acceptable use policy, tailored to the capabilities enabled for your organization. It is explicitly a starting point for your own review, not a document to adopt unchanged — it should go through whoever owns policy in your organization.

Yes. Billing and usage controls are configured during setup, and the intake forms capture who is permitted additional usage.

A live session covering both administration and end use. Administrators are taken through user management, access rights, reporting, and usage oversight. End users are taken through working with Claude productively.

Hands-on. Participants work in your own environment during the session and build something real, rather than watching a walkthrough.

How to write down a process, what to give Claude and what to withhold, how projects and conversations work, how to share work with colleagues, and where the genuine limitations and pitfalls are. Everyone leaves with a reusable runbook they have filled in themselves.

Yes, and concretely rather than abstractly. Participants are taught to remove identifying detail before it goes in — using a reference to an agreed location rather than a person's home address, for example. It is taught as a working habit, not a policy slide.

Yes. Connections to your Microsoft 365 or Google applications are configured during the session, so participants finish with working access.

Yes, and it is yours to keep and reuse for people who join later.

Yes. Additional sessions can be arranged, which is common where you want to train several groups separately or bring on new staff later.

A configured workspace matched to your identity, security, and data-handling requirements. A written record of every configuration decision, including access roles and product enablement. Your internal approval workflow, documented. A sample AI use policy. Stakeholder documentation covering reporting and user management. A runbook each participant completes during training. Working connectors to your Microsoft 365 or Google applications. And the session recording, plus training fundamentals that stay available for people who join later.

It is the artifact that makes the deployment governable. It tells you, in writing, what was decided and by whom — which is what you need when someone asks why a connector was approved, or when you are answering a client, insurer, or auditor asking how you control AI in your business.

A short working document each participant fills in during the training session as they capture a real process of their own. It is not a handout. They leave with something they built, in a format they can reuse the next time they want to hand a process to Claude.

User management, access assistance, and baseline integration support through the DeepNet service desk, on the same footing as the rest of your managed service.

Straightforward, low-risk integrations are handled as standard support. Anything more substantial is scoped separately.

Yes, if you prefer. The handoff documentation covers user management and reporting, and the training session takes your administrators through it. Whether you handle it internally or hand it to our service desk is a decision made during configuration.

No. This service covers implementation, governance, documentation, and training for Claude as a workplace platform. Custom application work is a separate engagement.

Not necessarily. Team plans support single sign-on and administrative control. Enterprise adds automated user provisioning, audit logs, and additional compliance controls. We confirm the right tier during scoping based on your size and compliance requirements.

Yes. An existing deployment can be brought into the same configuration and governance standard.