Claude Feature Gap Audit

To use this audit, copy it with the "Copy page" button above or open the Google Doc version. Read the companion comparison, personal Claude vs Claude Enterprise, for the thinking behind it.

Claude feature gap audit

A free template for CIOs, AI Ops leads and security teams who run Claude Enterprise or Claude Team. It compares what your people can get on personal Claude plans with what your organisation has switched on, finds where staff are going round you, and turns the gap into a short, scored plan. It works on its own; you don't need any platform to use it.

0. Audit details

• ☐ Organisation:
• ☐ Claude plan (Team or Enterprise) and seat count:
• ☐ Audit owner (usually the AI Ops lead):
• ☐ Sponsor (CIO or COO):
• ☐ Security and data protection partner:
• ☐ Date of this audit:
• ☐ Next review (monthly, or after each release review):

1. Feature gap table

One row per capability. Check the "personal plans" column against Anthropic's current plans and pricing page when you run the audit, because it changes often. Fill in your own setting, why it's off, and the condition that would let you switch it on.

2. Shadow-use signals

Tick what you've seen. Each one is evidence that the company's Claude is losing to a personal account.

• ☐ Personal Claude accounts on your company domain (Enterprise admins can list Free, Pro and Max accounts on a verified domain).
• ☐ A separate Claude Team account opened by a department.
• ☐ Expense claims for personal AI subscriptions.
• ☐ Screenshots or exports from personal Claude in team documents.
• ☐ Help desk tickets asking when a feature will be switched on.
• ☐ Champions reporting that people "use it at home instead".
• ☐ Low active use in teams that are known to use AI heavily.

3. Controls before features

Switch these on before you widen access, so every feature you enable is governed from day one. Check each against your plan: some are Enterprise only.

• ☐ Domain verified, and "Restrict organization creation" switched on.
• ☐ SSO enforced, not just configured.
• ☐ SCIM or JIT provisioning live (SCIM is Enterprise only).
• ☐ Groups that mirror your departments or cost centres, where your plan supports them.
• ☐ Admin controls for connectors set: which are allowed, read-only first.
• ☐ Custom data retention set (Enterprise only).
• ☐ Audit logs reviewed by a named person (Enterprise only).
• ☐ Compliance API connected to your archive or SIEM, if you need it (Enterprise only).
• ☐ Inference hooks connected to your data loss prevention or AI security tool, if you use one (Enterprise only; Owners and Primary Owners configure them).
• ☐ Organisation instructions and default model set.
• ☐ Spend limits set at organisation and user level.
• ☐ Feedback (thumbs up and down) setting decided.

4. Data rule

Replace case-by-case approvals with a standing rule. Agree it with your DPO and legal team; this template isn't legal advice.

5. Enablement backlog and ICE scores

List the capabilities you'll switch on, and for which group. Score each with ICE: impact, confidence and ease, each from 1 to 10, multiplied for a score out of 1,000. Start with the teams already using personal accounts.

6. Auto mode decision

Anthropic says auto mode is becoming the default on Pro, Max and Team, and that Enterprise admins will get an email before it becomes the default for their organisations. Decide before the email arrives.

• ☐ Our default for Claude apps:
• ☐ Our default for Claude Code (admins can pin it, or turn auto mode off, in managed settings):
• ☐ Work that always needs a person to approve (for example, production changes, payments, external sends):
• ☐ Who is told, and how:

7. Message to staff

Adapt this when you switch capabilities on.

"From [date], [capabilities] are switched on in our company Claude for [groups]. Please use it for work instead of personal accounts. It keeps your work in one place, under our company agreement, and you won't lose it if you change role. Here's what you can and can't put into it: [link to data rule]. Questions: [contact]."

8. Steering group lines

Copy these into your steering group update.

• Capabilities switched on this month: [list], for [groups].
• Shadow-use signals: [what you saw], [what changed].
• Controls in place: [for example, SSO enforced, audit logs reviewed weekly, inference hooks live].
• Decision needed: [for example, approve the data rule for client documents].

9. Implementation notes

• Run this audit monthly, or after each Claude release review. New defaults reopen the gap.
• Switch on by group, not all at once. The teams already on personal accounts are your pilot.
• Only claim your domain once the company's Claude is the better option.
• If you track your AI estate in Kowalah, the free plan covers recording systems and AI assets and logging opportunities.
• Feature availability changes often. Check Anthropic's pricing page and release notes before you rely on any row.