Updated November 2026. Plans and prices rechecked against Anthropic’s pricing page and help center on November 20, 2026.
Claude Enterprise is the plan most law firms end up pricing once privilege, audit or retention questions arrive. Claude Enterprise pricing also no longer works the way most articles describe it. Anthropic now sells Enterprise as a seat fee plus all usage billed at API rates, so the seat price is only the start of the number.
Short answer first, since it’s the one people came here for. The Claude Team plan is enough for most mid-size firms right up until audit, SCIM or retention requirements force the question, and then you need the Claude Enterprise plan. That’s the whole post in two sentences. Everything below is the reasoning, the current numbers behind Claude Team pricing versus Enterprise, and a checklist so you’re not guessing at which moment the switch happens. If you’re searching for Claude for business, Anthropic’s business plans are these two: Team and Enterprise.
Nobody upgrades to Enterprise to get a smarter Claude. Both plans include Claude’s core features, connectors and Claude Code. Firms upgrade because Team stops short of a few controls that eventually matter: SCIM user provisioning, audit logs, and custom data retention controls instead of Anthropic’s default terms.
Say your firm needs fewer than 150 seats (Team covers 2 to 150 people) and runs a mixed caseload. If no client’s outside counsel guidelines have addressed AI data handling yet, Team will likely carry you further than you’d expect. If a client’s OCG or your cyber insurer is already asking about audit trails, SCIM offboarding, or custom retention, you’re past Team. Claude for lawyers mostly comes down to two decisions: which plan, and which matters need the stronger controls.
For a side-by-side assessment of data access, pricing, and security, read our Claude vs Copilot comparison.
Claude Enterprise pricing vs Claude Team pricing at a glance
| Claude Team | Claude Enterprise | |
|---|---|---|
| Seats | 2 to 150, self serve | 20 seat minimum self serve, 50 seat minimum sales assisted |
| Price (US list, before tax) | Standard seat $20 per seat per month billed annually, $25 billed monthly. Premium seat $100 billed annually, $125 billed monthly | $20 per seat per month billed annually, plus all usage billed at API rates |
| What the seat covers | A fixed seat price. Premium seats carry 5x the usage of Standard seats | Access only. No included token allowance and no per seat usage limits |
| Spend controls | Org and user level spend caps | Admins set spend limits at the org and user level |
| SSO | SSO and domain verification | SSO and domain verification, plus domain capture, IP allowlisting and network level access control |
| Admin console | Central billing and administration, connector controls, usage analytics, desktop app deployment | Everything in Team, plus role based access, a Compliance API and an Analytics API |
| Data retention and training posture | No model training on your content by default; default retention | Same training default, plus custom data retention controls |
| Audit logs | No | Yes |
| Security additions | Not included | Compliance API, Analytics API, customer managed encryption keys, US only inference, inference hooks, IP allowlisting, network level access control, role based access |
| HIPAA | Not offered | HIPAA ready configuration with a BAA for eligible organizations |
| Connectors | Available; admins control remote and local connectors | Same connectors and controls |
| Billing | Credit card; monthly or annual | Annual only. Card or ACH on self serve; card, ACH or invoicing on sales assisted |
| Best fit | Governed seats and connector control. No compliance review is forcing the question yet | Client security reviews in play, OCG language on AI data handling, HIPAA needs, or SCIM driven offboarding at scale |
Claude Team pricing and Claude Enterprise pricing are US list prices from Anthropic’s pricing page and its Enterprise page, read November 20, 2026, before tax. Anthropic says prices and plans are subject to change at its discretion. Enterprise usage is billed on top of the seat fee, so your real cost depends on which models and tasks your people run.
What Claude Enterprise costs once usage is billed
The Claude Enterprise cost question has two parts now. The first is the seat. At $20 per seat per month billed annually, 20 seats comes to $4,800 a year in seat fees before any usage.
The second part is everything the seat doesn’t include. Anthropic’s help center says the Enterprise seat fee covers access only. Every token your people use across chat, Claude Code and Cowork is billed on top at standard API rates, with no included allowance and no per seat usage limits. Usage depends on which models and tasks your people run, so any number I gave you here would be a guess. I won’t give one.
What you can do is cap it. Admins can set spend limits at the organization and individual user level, and a firm should set them before the first rollout rather than after the first invoice. How you pay for usage depends on how you buy. Self serve Enterprise buys usage credits upfront. Sales assisted Enterprise bills usage monthly in arrears.
Anthropic also says Enterprise pricing can be customized through its sales team with volume and committed spend discounts, and it doesn’t publish discount rates. Treat any discounted figure you see quoted elsewhere with suspicion, and read Anthropic’s billing guidance before you sign.
What Claude Team gives a firm, and where it stops
Team gets you more than people expect. Central billing instead of a stack of personal credit cards. SSO through your identity provider instead of forty separate passwords. An admin console where someone can see which connectors are turned on, switch the risky ones off, and check usage analytics. And the commercial training posture: by default your firm’s content isn’t used to train Anthropic’s models, which is the single biggest gap between this plan and an individual using Claude on their own.
Here’s where it stops. No SCIM provisioning, so when an associate leaves, removing their access is a manual step somebody has to remember. No audit logs, so if a client or court ever asks what was asked of Claude and when, there’s no clean record to hand over. And no custom data retention controls, just Anthropic’s default commercial terms.
For a lot of firms, none of that bites for a long while. Know the gap exists before a client’s security questionnaire finds it for you. If the open question is whether anyone should be on a personal account at all, we compare the two in Claude Team plan vs personal Pro accounts.
What the Claude Enterprise plan adds that law firms need
SCIM provisioning is the one I’d flag first. It ties Claude access to your HR and identity system, so when someone leaves or changes roles, access changes with them instead of sitting on a to-do list. For a firm with a few hundred attorneys and real turnover, that closes a genuine gap.
Audit logs matter more than most firms expect until they need one. If a malpractice claim, a bar complaint, or a privilege dispute ever requires reconstructing what an attorney asked an AI tool and when, Team gives you nothing. Enterprise gives you a record.
Then there’s retention. Enterprise adds custom data retention controls, so a firm can set retention to fit a client’s terms instead of living with the default. If a matter calls for a specific arrangement, such as zero data retention, confirm in writing with Anthropic what your agreement covers.
The newest control is inference hooks. When a firm turns them on, Claude sends every prompt to a server the firm hosts, and every tool call too, before the tool runs. That server answers allow or deny, and Claude continues only after it gets the answer. Owners and Primary Owners turn them on. Anthropic’s help center lists data loss prevention and transcript archival among the common uses. Team doesn’t include them.
The Compliance API works after the fact. It gives programmatic access to Claude usage data, including activity logs, chat histories and file content, filtered by user and time range. For a firm answering a client’s audit request, that’s the difference between pulling a record and rebuilding one from memory.
Two more Enterprise only items can matter for firms: a HIPAA ready configuration for health related matters, and IP allowlisting with network level access control. Real differences, just rarely the first reason a firm upgrades.
So does upgrading change what Claude can do? The model stays the same. What changes is governance, retention and audit.
When Claude Pro is Enough
Pro is fine for a solo practitioner or small group doing legal research, drafting, or general writing that never touches client-identifying data. No firm data flowing through it, no connectors into your document management system, no reason to involve IT.
Here’s the part worth knowing before anyone reaches for a personal account anyway. Anthropic’s pricing page lists model training on the consumer plans (Free, Pro, Max and Ultra) as opt-out, so conversations can be used for training unless the individual turns it off. Team and Enterprise list no training on your content by default. That’s a different posture, and on a personal plan the setting sits with the individual, where your firm can’t see or control it. We cover the confidentiality risk of personal accounts in AI confidentiality risk in Claude Pro accounts.
So the real risk with Pro is behavior. A partner pastes a draft complaint into a personal account because it’s faster than an IT ticket. For the decision itself, see Claude Team vs Pro for law firm IT.
Governance Checklist Before Client Data Touches Any Claude Plan
- Confirm which agreement governs your firm’s usage, Team’s default terms or a negotiated Enterprise agreement, before a single attorney logs in. Don’t assume. Read it.
- Decide who controls connectors. Team and Enterprise both include admin controls for remote and local connectors, so set them before the first attorney connects a mail or document system.
- Build a leaver process specifically for Team, since there’s no SCIM behind it. Deprovisioning is a manual step, and manual steps get forgotten during a busy departure.
- Flag which matters need custom data retention before Claude touches them, not after. A client’s OCG with AI data terms your firm hasn’t negotiated means that matter needs Enterprise.
- Set an explicit policy on personal accounts. Nothing stops an attorney from using Claude Pro on their own card for firm work, unless someone checks.
Self serve or sales assisted Claude Enterprise
Both routes carry the same features and the same seat price. The differences are in how you buy and pay.
Self serve means a 20 seat minimum, buying online, US dollars only, card or ACH, and usage credits purchased upfront. Sales assisted means a 50 seat minimum, card, ACH or invoicing, multiple currencies, and usage billed monthly in arrears. It also opens up AWS Marketplace, trials and consultation with Anthropic’s sales team. Firms still on older Enterprise seat types move to the single seat model at their next renewal.
Where This Leaves Your Firm
This question doesn’t live apart from the Microsoft side of your stack. I’ve written before about what Copilot Premium changes for data protection at law firms and about why DLP alone won’t protect agents like Claude and Copilot. I’ve also covered the M365 Copilot licensing shape for 2026. The access and retention questions underneath all these tools rhyme more than vendors like to admit. So do the ethics questions, which we cover in our post on ABA Opinion 512 confidentiality duties.
For daily work, see Claude vs Copilot for legal work, task by task.
On the Claude side, Anthropic’s own plan comparison (https://claude.com/pricing) is the source to check, since features move between plans. If any attorney has considered a personal Claude account for firm work, the Pro section above is the part to share.
Once the plan is chosen, Claude Enterprise governance is the next step: set the pilot boundary before client data goes in.
Quick answers
How much does Claude Enterprise cost?
Claude Enterprise lists at $20 per seat per month billed annually, plus all usage billed at API rates. Self serve needs at least 20 seats and an annual commitment. The seat fee covers access only, so the monthly total depends on which models and tasks your people run. Prices are from Anthropic’s pricing page, read November 20, 2026, before tax.
How much is Claude Team per seat?
On annual billing, a Claude Team Standard seat is $20 per seat per month, or $25 billed monthly. A Premium seat, with five times the usage of a Standard seat, is $100 billed annually or $125 monthly. That’s the Claude Team plan pricing on Anthropic’s pricing page as of November 20, 2026, before tax. Claude Team cost scales with seat count and seat type.
What is the minimum number of seats for Claude Enterprise?
Self serve Claude Enterprise needs at least 20 seats. Sales assisted Enterprise needs at least 50. Both carry the same features and the same seat price, and Enterprise is sold on annual terms only. If your firm is below the Enterprise floor, Team covers firms of 2 to 150 people.
Does Claude have a business plan?
Yes. Anthropic’s business plans are Team and Enterprise. Team suits firms of 2 to 150 people that want central billing, SSO and admin controls. Enterprise adds SCIM, audit logs, custom retention, the Compliance API and inference hooks, with usage billed at API rates on top of the seat fee. If you’re searching for Claude for business, those are the two plans to compare.
Does Anthropic train on Claude Team or Enterprise data?
Not by default. Anthropic’s pricing page lists model training as none by default for Team and Enterprise. Its privacy center, in an article dated August 18, 2026, says it won’t use inputs or outputs from its commercial products to train models unless you report feedback or choose to allow it. Owners can turn off feedback with the Rate chats setting.
Is Claude Team OK for privileged work?
For most matters, yes. Team runs on commercial terms, meaning no training on your content by default, which covers the core confidentiality concern. Where Team falls short is process: no audit trail to reconstruct what was asked, and no custom retention if a matter calls for it. Firms with regulated clients or specific OCG retention language move those matters to Enterprise.
Do we need SSO and SCIM at our size?
SSO comes with Team, so the real question is SCIM, which is Enterprise only. In the hundreds of attorneys with regular turnover, SCIM saves real IT hours and closes a real gap. Under fifty people, manual provisioning is probably fine for a while yet.
Related Cocha guides
Primary sources
- Anthropic, Plans and pricing, https://claude.com/pricing (read November 20, 2026). Supports Team Standard and Premium seat prices, the 2 to 150 Team range, the Enterprise seat price plus usage at API rates, and annual only billing. Also supports the plan comparison rows, model training settings by plan, and the note on customized Enterprise pricing through sales.
- Anthropic, Claude Enterprise plan, https://claude.com/solutions/enterprise (read November 20, 2026). Supports $20 per seat per month billed annually, usage billed at API rates, the annual commitment and the 20 seat minimum.
- Anthropic, Claude Team plan, https://claude.com/get-claude-team-plan (read November 20, 2026). Supports no training on your data by default, SSO, and org and user level spend caps.
- Anthropic Help Center, What is the Enterprise plan?, https://support.claude.com/en/articles/9797531-what-is-the-enterprise-plan (article dated September 1, 2026; read November 20, 2026). Supports the access only seat fee, usage at API rates, spend limits, the Enterprise security features and the Compliance API description. Also supports the self serve and sales assisted differences and the move from older seat types at renewal.
- Anthropic Help Center, Inference hooks overview, https://support.claude.com/en/articles/16059458-inference-hooks-overview (read November 20, 2026). Supports how inference hooks work, who turns them on and the common uses.
- Anthropic Privacy Center, Is my data used for model training?, https://privacy.claude.com/en/articles/7996868-is-my-data-used-for-model-training (article dated August 18, 2026; read November 20, 2026). Supports the commercial training default, the feedback exception and the Rate chats setting.
If you’re not sure which plan your firm needs, or whether your current setup already has gaps like these, our AI Readiness Snapshot is built to find them. It’s a 60-minute working session followed by written findings and a 30-minute review. It’s read only, with no changes and no software to install. It looks at who has access to what, how well your data is protected and monitored, whether shadow AI is already in use, and where your Microsoft Secure Score stands.




