Choosing Claude (5)
-
Claude Enterprise vs. ChatGPT Enterprise vs. Microsoft 365 Copilot: which fits your company?
Pick by where the work happens. Microsoft 365 Copilot fits companies whose work lives in Outlook, Teams, Word and SharePoint. ChatGPT Enterprise fits companies that want OpenAI models with data residency in ten regions. Claude, powered by Anthropic, fits teams doing long-document analysis, coding and agent work who want usage billed at API rates with admin spend limits. As of October 2026, only Copilot and Claude publish an enterprise price.
Read the full guide →Follow-up questions
Is Claude Enterprise cheaper than Microsoft 365 Copilot?The seat is cheaper: $20 per user per month for Claude Enterprise against $30 for the Copilot enterprise add-on, as of October 2026. But the Claude seat includes no usage, so total cost depends on how much your people use. Light users usually add little; engineers using Claude Code add much more.
Does ChatGPT Enterprise publish its price?No. As of October 2026, OpenAI lists ChatGPT Enterprise as custom pricing through sales, with credit-based and token-based options. ChatGPT Business is published at $20 per standard seat per month billed annually.
Which of the three can sign a HIPAA BAA?Anthropic documents a self-serve BAA for Claude Enterprise when the Primary Owner enables the HIPAA-ready offering. We did not see BAA terms on the ChatGPT or Copilot pricing pages, so ask OpenAI and Microsoft directly and get the answer in your contract.
-
Claude Team vs. Enterprise: which plan does your company need?
For Claude, powered by Anthropic, choose Team if you have 2 to 150 users, want a flat bill with usage included up to per-seat limits, and single sign-on is enough identity control. Choose Enterprise if you need SCIM, audit logs, the Compliance API, custom data retention, role-based access, IP allowlisting or a HIPAA BAA, or more than 150 users. Enterprise seats are $20 a month, but every token is billed on top.
Read the full guide →Follow-up questions
Can Claude Team have more than 150 users?No. Anthropic lists Team for 2 to 150 seats. Above that, upgrade the same organization to Enterprise so you keep chats, projects and roles.
Is Claude Code included on Team and Enterprise?Yes. Both plans include Claude Code and Cowork. On Team it draws from each seat's usage limits; on Enterprise every session is billed as usage at API rates.
Can we downgrade from Enterprise back to Team?The self-serve Team-to-Enterprise upgrade is described as irreversible. If you think you may need to go back, talk to Anthropic sales before upgrading. Note that a retention period set on Enterprise still applies to artifacts after a move to Team.
-
What does Claude Enterprise cost? Seats, usage, API rates and rollout services
The Enterprise plan of Claude, powered by Anthropic, costs $20 per seat per month, billed annually with a 20-seat minimum, plus every token your people use billed at standard API rates, as of October 2026. A 200-person company pays $48,000 a year in seats before any usage. Usage depends on which models and surfaces people use, and admins cap it with organization, group and user spend limits. Implementation services are a separate, one-time cost.
Read the full guide →Follow-up questions
Is Claude Enterprise $20 per user?The seat is $20 per user per month, billed annually, as of October 2026. Usage is extra: every token in chat, Claude Code and Cowork is billed at standard API rates, so your real per-user cost depends on how people use it.
Can we cap Claude Enterprise spending?Yes. Owners and billing admins can set spend limits for the whole organization, for groups and for individual users, and can give a group a pooled monthly budget in beta. When a limit is reached, usage stops until the next period or until an admin raises it.
Does Claude Enterprise have a minimum?Self-serve Enterprise has a 20-seat minimum. The Claude Help Center lists a 50-seat minimum for sales-assisted Enterprise, which adds invoicing, multi-currency billing and monthly usage billing in arrears.
-
Does Claude train on your data? Training, retention, zero data retention and where data lives
No, not on business plans by default. For Claude, powered by Anthropic, the Commercial Terms state that Anthropic may not train models on customer content from Team, Enterprise or API use. There is one practical exception: if a user submits feedback with the thumbs up or down button, that conversation can be used for training, and admins can switch that button off. Consumer accounts (Free, Pro, Max) work differently: training there is an opt-out setting.
Read the full guide →Follow-up questions
Does Claude Pro train on my conversations?It can. Pro is a consumer plan, and Anthropic lists model training as opt-out there, controlled by each user's model improvement setting. For company work, use Team or Enterprise, where content is not used for training by default.
Can Anthropic employees read our chats?Anthropic says that for Covered Model retention, no personnel can read retained conversations by default. Human review happens only through a controlled path, for example when automated systems flag content for potential harm. Each access is recorded in a tamper-proof log.
Does deleting a chat remove it from Anthropic systems?On Team and Enterprise, a deleted chat leaves the user's history immediately and Anthropic back-end storage within 30 days. Exceptions are chats flagged for Usage Policy violations, data under legal hold, and anything already submitted as feedback.
-
Is Claude HIPAA compliant? What the HIPAA-ready offering and BAA actually cover
Claude, powered by Anthropic, can be used with protected health information only through a HIPAA-ready configuration covered by a Business Associate Agreement: a Claude Enterprise plan with HIPAA enabled by the Primary Owner, or a HIPAA-ready API organization. No tool is HIPAA compliant on its own; compliance depends on how you configure and use it. Team, Pro, Max and Free plans cannot enable HIPAA, and some Enterprise features stay outside the BAA.
Read the full guide →Follow-up questions
Does Anthropic sign a BAA?Yes, for its HIPAA-ready services: Claude Enterprise with HIPAA enabled, the first-party Claude API through a HIPAA-ready organization, and Claude in Microsoft Foundry with the HIPAA configuration. On Enterprise, the Primary Owner accepts it in organization settings without a separate signature cycle.
Is Claude Team HIPAA compliant?No. Anthropic states that Team and individual plans cannot enable HIPAA, so there is no BAA on Team. Do not put PHI into Team, Pro, Max or Free accounts.
Can we use connectors like Google Drive with PHI?Connectors are available on HIPAA-ready Enterprise plans, but data sent to third parties through them is not covered by the Anthropic BAA. If you use one, you need your own BAA with that third party and a documented risk decision.
Rolling it out (5)
-
How do you roll out Claude company-wide? A phased plan with owners, settings and checkpoints
Roll out Claude in five phases: choose the plan and data rules, set up identity (verified domain, SSO, then SCIM), configure connectors, skills and spend limits before anyone logs in, pilot with 50 to 100 users for two to four weeks, then expand one department at a time with champions and training. Identity, data access and enablement are most of the work. The software is the easy part.
Read the full guide →Follow-up questions
How long does a company-wide Claude rollout take?It depends mostly on identity and data readiness, not on Claude. Anthropic's administrator guide suggests a pilot of 50 to 100 users monitored for 2 to 4 weeks before department-by-department expansion. Companies that already run SSO and SCIM and have clean document permissions move fastest.
Do we need Claude Enterprise, or is Team enough?Team works for 2 to 150 seats when SSO and basic admin controls are enough. Choose Enterprise if you need SCIM, domain capture, audit logs, the Compliance API, custom retention, role-based access, spend limits by group or user, IP allowlisting, or the HIPAA-ready offering.
Should we claim our domain in Claude?Only after you have enforced SSO and set up SCIM or JIT, and only after telling employees. Domain claiming is permanent and moves personal Claude accounts on your domain into the company organization. Restricted organization creation is worth enabling either way.
-
How do you connect Claude to your company's data? SharePoint, Google Drive, Slack, databases and ERPs
You connect Claude to company data through connectors. Use the ready-made ones in the Connectors Directory for Microsoft 365, Google Drive, Gmail, Slack and many SaaS tools. For databases, ERPs and internal systems, point a custom connector at a remote MCP server you build or buy. An Owner adds each connector for the organization, then each person signs in with their own account, so Claude only sees what that person can already see.
Read the full guide →Follow-up questions
Can Claude see files a user does not have access to?Not through per-user connectors such as Microsoft 365 or Google Drive. Each person signs in with their own account and Claude mirrors that account's permissions. A connector set up with a shared credential is different: every member gets that credential's access.
Do I need a Microsoft Global Administrator to connect SharePoint?Yes. An Entra Global Administrator grants a one-time consent for the tenant, and on Team and Enterprise a Claude Owner also enables the connector. Enabling write tools requires the Global Administrator to approve an updated permission set.
Can Claude connect to a database directly?Not without an MCP server in between. You host a server that connects to the database with a read-only role and exposes specific tools, then add it as a custom connector by URL.
-
What are Claude Skills, and how do companies build their own?
A Claude Skill is a folder with a SKILL.md file, holding a short name, a description of when to use it, and step-by-step instructions, plus optional scripts, templates and reference files. Claude loads it only when a task matches. Companies build their own by capturing a workflow people repeat, writing a precise description, testing it on real tasks, then publishing it to the organization, often bundled with connectors as a plugin.
Read the full guide →Follow-up questions
Do Claude Skills require coding?No. Many useful skills are only a SKILL.md file of written instructions. Scripts help for steps that must be exact, but a process owner can write and test an instruction-only skill without a developer.
Which Claude plans include Skills?Free, Pro, Max, Team and Enterprise, with code execution enabled. Sharing and publishing to an organization are Team and Enterprise features, and group sharing and skill scanning are Enterprise.
Can admins push a skill to everyone?Yes, on Team and Enterprise. An Owner adds the skill in Organization settings > Plugins & skills and sets it to Installed by default or Required. Enterprise can also set availability per user group.
-
How do you roll out Claude Code to an engineering team safely?
Roll out Claude Code by choosing the provider first, deploying managed settings before developers get access, piloting on a few repositories with permission rules, sandboxing and an MCP allowlist in place, then expanding while analytics, spend limits and OpenTelemetry track usage. Managed settings override anything a developer configures, so your policy belongs there, not in a wiki page people may not read.
Read the full guide →Follow-up questions
Is Claude Code included in Claude Team and Enterprise?Yes for Team: every Team seat includes Claude Code, and Premium seats add usage. New and self-serve Enterprise seats include it. Older Enterprise contracts include it on Chat + Claude Code or Premium seats, so check your seat types.
Can developers override our Claude Code policy?Not for keys set in managed settings, which take precedence over user, project and command-line settings. They can add entries to merged lists such as permissions.allow unless you set allowManagedPermissionRulesOnly.
Does Anthropic train on our code?Not on Team, Enterprise, Claude API or cloud provider plans, according to Anthropic's admin documentation. Retention depends on your provider and plan; zero data retention is available to qualified Enterprise accounts.
-
Should you build Claude agents on the Messages API, the Agent SDK, or Managed Agents, or buy one?
Use Managed Agents when you want Anthropic to host a long-running agent with the sandbox, tools and session state already built. Use the Agent SDK when you want Claude Code's agent loop inside software you run. Use the Messages API when you need full control of the loop, strict data handling or the simplest request-response flow. Buy or configure an existing product when the work already lives in Claude, Slack, GitHub or Salesforce.
Read the full guide →Follow-up questions
What is the difference between the Claude Agent SDK and Managed Agents?The Agent SDK is a library you run in your own process, giving you Claude Code's tools and loop. Managed Agents is a hosted harness: Anthropic runs the loop and the sandbox, or you run only the tool-execution sandbox, and you drive it through the API.
Is Claude Managed Agents eligible for zero data retention or HIPAA?Not as of October 2026. Anthropic states that Managed Agents is stateful by design and not eligible for ZDR or HIPAA BAA coverage. You can delete sessions and uploaded files through the API.
How much does Claude Managed Agents cost?The Claude pricing page lists $0.08 per session-hour of active runtime plus standard token rates. Web search and code execution have their own charges.
Running it well (5)
-
What does a Claude implementation partner do, and do you need one?
A Claude implementation partner is a consulting firm that helps a company plan, configure, connect, secure and adopt Claude, powered by Anthropic, so licenses turn into changed work instead of idle seats. Typical scope: choosing the plan and deployment route, identity and admin controls, connecting company data, building Skills or agents, training and measurement. You need one when the work crosses several systems or carries regulatory risk. A small team using Claude for drafting often does not.
Read the full guide →Follow-up questions
Is a Claude partner the same as a Claude reseller?No. Claude Partner Network membership is about services: training, technical support and joint go-to-market with Anthropic. You buy Claude plans from Anthropic, or Claude models through a cloud provider, and pay a partner separately for implementation work. Abstrakt Solutions sells services only.
How do I check a firm's Claude Partner Network status?Look the firm up in the Claude Partner Hub directory, which shows its tier and lets you filter by services, industry and deployment surface. Then ask for the Credly badges of the individuals assigned to your project. Anthropic publishes the requirements for Select, Preferred and Global Premier, so you can compare claims against them.
Can a partner help if we already bought Claude Enterprise?Yes, and this is common. Companies often buy seats first and then find that connectors, custom roles, retention and training need setup before usage spreads. A partner can start with an audit of the current admin configuration and a usage review from the analytics dashboard.
-
How do you measure the ROI of Claude in your company?
Measure the ROI of Claude, powered by Anthropic, by comparing a baseline for specific tasks with the same tasks after rollout, across four layers: adoption, task outcomes (time, throughput and quality), cost per task, and the business result the task feeds. Claude Team and Enterprise analytics cover adoption and spend. Time, quality and business impact need your own baselines, time studies and review samples, collected before people start using Claude.
Read the full guide →Follow-up questions
Does Claude show how much time it saves?The analytics dashboard includes an estimated time saved metric, and the Claude Code Value tab shows estimated productivity lift and time recovered with adjustable formulas. These are estimates. Use them to pick workflows to study, then confirm with your own timestamps or time logs before reporting them as results.
Who can see Claude usage analytics?On Team plans, Owners and Primary Owners. On Enterprise, Owners, Primary Owners and Admins, with Admins unable to see spend. Enterprise custom roles can also be granted access to the Analytics area. Only the Primary Owner can create keys for the Enterprise Analytics API.
How long should we wait before measuring results?Collect the baseline before access starts, then measure outcomes after four to six weeks of use. Earlier numbers mostly reflect learning time. Keep tracking rework for at least a month after that, since some errors only show up downstream.
-
What should an AI acceptable-use policy say, and how do you enforce it in Claude?
An AI acceptable-use policy should state which AI tools and accounts are approved, what data may go into them, which uses are banned or need human review, who approves new connectors and agents, how output is checked and disclosed, and what is logged. In Claude, powered by Anthropic, many of those rules can be enforced by settings on Team and especially Enterprise plans. The rest need a process with a named owner.
Read the full guide →Follow-up questions
Do we need a separate AI policy, or can we update our existing acceptable-use policy?Either works if the content is there. Many companies add an AI section to the existing acceptable-use policy and a separate standard for connectors and agents, because those change often and need a technical owner. What matters is that data rules, approval steps and review duties are written down and tied to settings.
Can Claude block employees from pasting sensitive data?On Enterprise, inference hooks let your own service inspect prompts, tool responses and uploaded file text before Claude processes them and return allow or deny, which supports data loss prevention. You build and run that service. On Team, you rely on training, policy and your existing endpoint or network controls.
Can we stop employees from using personal Claude accounts at work?On the corporate network, yes, with Tenant Restrictions on Enterprise: your proxy adds a header listing approved organization IDs and other organizations are rejected. Off the corporate network, it is a policy and device-management question. Enterprise also includes domain capture; confirm with Anthropic how it applies to accounts people already created with your company domain.
-
Is MCP safe for business use? The real risks and the controls that work
MCP is safe enough for business use when you control which servers are connected, what scopes they hold and which actions need a person's approval. The Model Context Protocol is a standard for passing tool calls and results, not a security boundary. The real risks are prompt injection through tool results, poisoned tool descriptions, untrusted or changed third-party servers, over-broad OAuth scopes and data leaving through one tool after being read through another. Each has known controls.
Read the full guide →Follow-up questions
Can a malicious MCP server steal data from my other connectors?It can try. A malicious server can return results or tool descriptions with hidden instructions that push the model to read data through another connector and send it somewhere. Anthropic advises connecting only to servers built and hosted by parties you trust and watching tool inputs and outputs. Requiring approval for outbound tools is the most reliable backstop.
Is a remote MCP server safer than a local one?They carry different risks. A local server runs code on the user's machine with the user's privileges, so a malicious one can do anything that user can. A remote server cannot touch the machine but holds tokens to your systems. Control local servers with managed allowlists and remote ones with scopes, approvals and vendor review.
Should we build our own MCP servers instead of using third-party ones?Build when no trusted vendor server exists, when you need to expose only a narrow set of operations, or when the system holds sensitive data. Building means owning the security duties in the MCP specification, including token validation and per-client consent for proxy servers. For mainstream SaaS tools, a vendor-hosted server is usually the better starting point.
-
How can finance and accounting teams use Claude without weakening controls?
Finance and accounting teams get the most from Claude, powered by Anthropic, on the work around the numbers: drafting variance commentary, explaining reconciliation differences, answering accounting policy questions, reviewing contracts and invoices, and building or debugging Excel models through Claude for Excel. The ledger, approvals and sign-offs stay where they are. For a SOX-minded team the rule is simple: Claude prepares, a named person reviews, and the workpapers show both.
Read the full guide →Follow-up questions
Is Claude for Excel available on Team and Enterprise plans?Yes. As of October 2026 Anthropic lists Claude for Excel as generally available on Pro, Max, Team and Enterprise plans. Admins deploy it through the Microsoft 365 Admin Center or with a manifest file, and it runs on Excel on the web, current Microsoft 365 Excel on Windows and Excel on Mac.
Can Claude for Excel be used in SOX-relevant work?It can support the work, but it does not replace the control. Keep the reviewer's documented sign-off as the control, store evidence in your workpapers rather than in the add-in's local chat history, and discuss the process change with your auditors. Anthropic itself does not recommend the add-in for audit-critical calculations without verification.
Does Claude for Excel activity show up in audit logs?Not in Enterprise audit logs. For Enterprise organizations with the Compliance API enabled, Claude for Excel sessions are included, and Anthropic describes that coverage as public beta. The add-in also does not inherit custom data retention settings.
How we help with Claude (8)
Each service page explains what the work involves, the Claude features it relies on, how a rollout comes together and the mistakes to avoid.
Claude Enterprise Rollout
Plan, identity, roles, spend and retention set up in the right order, so Claude reaches every team without lockouts or surprise bills.
Should we buy Claude Team or Claude Enterprise? Choose Enterprise if you need SCIM, custom roles, audit logs, the Compliance API, custom retention, domain capture or a HIPAA-ready configuration, or if you will pass 150 seats. Team suits smaller groups that need SSO, JIT provisioning, shared projects and central billing. Enterprise has a 20-seat minimum and bills usage at API rates on top of the seat price, so estimate usage by group before deciding.
What decides how long a Claude rollout takes? Mostly your identity team and change windows, not Claude. DNS verification can take up to 48 hours to propagate, SCIM depends on when your identity provider can be changed, and account claiming adds a fixed 30-day migration window. We give you a dated plan after discovery, once we know your provider, group structure and approval steps.
Claude Connectors and MCP
Connect Claude to Microsoft 365, Google Workspace, Slack and your internal systems with scoped access, approvals and logs.
Can Claude see files a user is not allowed to open? Not through the Microsoft 365 or Google Drive connectors. Both use the signed-in user's own account, so Claude reaches only what that account can already open. The exception is a custom connector configured with a shared API key, which gives everyone that key's access.
Do we need to build anything to connect SharePoint and Outlook? Usually no. The Microsoft 365 connector is already in the directory. Your Entra Global Administrator grants tenant consent once, a Claude Owner adds the connector, and members connect their own accounts. The work is deciding which write actions to allow and checking SharePoint permissions first.
Claude Skills and Plugins
Turn the way your teams write proposals, reports and reviews into Skills and plugins that Claude follows the same way every time.
What is the difference between a Skill and a custom prompt? A prompt is typed into one conversation. A Skill is saved once, loads automatically whenever a request matches its description, and can carry templates, reference files and scripts that a prompt cannot. Skills also have versions, owners and organization-wide distribution.
Can admins push a Skill to everyone in the company? Yes, on Team and Enterprise plans. An Owner adds it in Organization settings > Plugins & skills and sets it to Installed by default, which members can turn off, or Required, which they cannot. Enterprise organizations can also set different availability per user group.
Claude Code for Engineering Teams
Roll Claude Code out to every developer with managed policy, sandboxing, CI review and spend controls set before the first session.
Should we buy Claude Team seats or use API billing for Claude Code? Most companies with up to a few hundred developers start with Team seats, because Claude Code, server-managed settings and the analytics dashboard come with them. Console API billing suits teams that want purely metered spend, per-user monthly spend limits in the Claude Code workspace, or API keys for automation. Heavy Team users can move to Premium seats, which carry five times the Standard usage.
Can we run Claude Code through AWS Bedrock or Google Cloud instead of Anthropic? Yes. Claude Code supports Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry through environment variables such as CLAUDE_CODE_USE_BEDROCK, and usage bills to your cloud account. That usage does not appear in the Anthropic analytics dashboards and server-managed settings do not apply, so plan on MDM-delivered policy and OpenTelemetry. A self-hosted Claude apps gateway can add SSO sign-in, managed settings and per-user spend limits on those providers.
Custom Claude Agents
Design, build and run Claude agents on the API, the Agent SDK or Managed Agents, with evals, guardrails and cost controls from day one.
Should we build a custom Claude agent or use Agentforce? If the work starts and ends inside Salesforce, Agentforce usually wins on build effort, and Salesforce offers Claude models hosted on Amazon Bedrock inside Agentforce. Build a custom agent when inputs live outside Salesforce, when the job needs a sandbox with files and shell access, or when outputs go to several systems. Many companies run both, connected through MCP or APIs, and our Agentforce vs. custom agents guide walks through the choice.
What is the difference between the Agent SDK and Managed Agents? Both provide the agent loop so you do not write it. The Agent SDK is a library you run in your own process and infrastructure. Managed Agents is hosted: Anthropic runs the loop and stores sessions server-side, in its cloud sandbox or a self-hosted sandbox you operate. Managed Agents is in beta and is not eligible for ZDR or HIPAA BAA coverage.
Claude Security and Governance
Security review, admin configuration and a written AI policy for Claude, finished before sensitive data goes in.
Does Anthropic train its models on our company's Claude data? Not under commercial terms. Anthropic's Commercial Terms say it may not train models on Customer Content from the services, and claude.com/pricing lists model training as 'None by default' for Team and Enterprise. The exception is if your organization opts in to a program such as the Development Partner Program.
How long does Anthropic keep our Claude Enterprise chats? Until you set a period. Enterprise data is retained indefinitely unless a Primary Owner or Owner sets custom retention, with a 30-day minimum. When a user deletes a conversation, it leaves their history immediately and backend storage within 30 days. Content flagged for Usage Policy violations can be kept up to 2 years.
Claude Training and Adoption
Role-based Claude training, a champions network and shared projects and Skills, measured with real usage data instead of survey optimism.
Does Anthropic offer free Claude training? Yes. Claude Academy has free, self-paced courses with completion certificates, including Claude 101, Introduction to Claude Cowork, AI Fluency: Framework and Foundations, and courses for developers. They are a good baseline; what they cannot do is train people on your files, systems and approval rules, which is where paid sessions earn their cost.
How many champions do we need? One per team that shares a workflow, not one per fixed number of employees. Two departments with the same job title but different systems need two champions. Pick people who do the work every day and give their managers a clear time expectation, or the role fades after the launch month.
ChatGPT or Copilot to Claude
Move teams from ChatGPT or Microsoft 365 Copilot to Claude, or run both side by side, without losing the work people built.
Can we import our ChatGPT chat history into Claude? Not as a company-wide transfer that we could verify. Individuals can use Anthropic's memory import at claude.com/import-memory, which moves summarized preferences rather than transcripts. Export old conversations for your records instead, and put documents that still matter into the relevant Claude Project.
Will our custom GPTs work in Claude? Not directly; they are rebuilt. A GPT's instructions and knowledge files map well to a Claude Project, a repeatable procedure maps to a Skill, and GPT actions become connectors to remote MCP servers. Test each rebuild against real examples from the old GPT before users switch.
Have a question that is not here? Ask a consultant, or start with the Claude consulting overview.