Resources / Governance Toolkit / Checklists
General AI Assistant Checklist
This is your most important tool decision. The General AI Assistant becomes your organization's default -- the tool everyone reaches for first. Getting this right eliminates 60-70% of shadow AI usage on its own.
The Options
Claude Team
Anthropic. Strong reasoning, long context, excellent writing. SOC 2 Type II.
Gemini Business
Google. Deep Google Workspace integration. Included in some Workspace plans.
ChatGPT Team
OpenAI. Largest ecosystem, GPT-4o + image generation built in. SOC 2 Type II.
M365 Copilot
Microsoft. Embedded in Office apps. Requires Microsoft 365 E3/E5.
Local Open-Source
Llama, Mistral, etc. via Ollama or vLLM. No data leaves your infrastructure. Requires GPU hardware and in-house ML expertise.
Evaluation Checklist
Work through each section. Your answers will narrow the field quickly.
1. Industry & Compliance Requirements
Are you in a regulated industry? (Healthcare, finance, legal, government)
Yes, highly regulated: Prioritize SOC 2 Type II + BAA (HIPAA) or equivalent. Claude Team, ChatGPT Team, and M365 Copilot all offer BAAs. Local open-source gives maximum control. Gemini Business has SOC 2 but verify BAA availability for your specific plan.
Somewhat regulated: SOC 2 Type II on Team/Business plans is sufficient. All commercial options qualify.
No: Compliance doesn't narrow your options. Move to the next question.
Do you have data residency requirements? (EU, specific country)
Yes, strict residency: M365 Copilot offers the most granular data residency controls via Azure regions. Local open-source is the only option for air-gapped environments. Check current regional availability for Claude Team and ChatGPT Team -- both are expanding but not universally available in all regions.
No: All options work. Move on.
Does your legal team require zero data retention by the AI provider?
Yes: Claude Team, ChatGPT Team, and Gemini Business all offer zero-training guarantees on business plans. Local open-source has zero external retention by default. M365 Copilot retains data within your Microsoft tenant. Verify the specific data retention policy for whichever tool you're evaluating -- "no training" and "no retention" are different things.
No: All business-tier plans already commit to not training on your data.
2. Existing Tech Stack
What's your primary productivity suite?
Google Workspace: Gemini Business has the strongest native integration -- it lives inside Docs, Sheets, Gmail, and Slides. This is a significant advantage if your team lives in Google apps all day.
Microsoft 365: M365 Copilot integrates directly into Word, Excel, Outlook, Teams, and PowerPoint. If you're already on E3/E5 licensing, the per-user add-on cost may be lower than a separate tool.
Mixed / Other: Claude Team or ChatGPT Team work as standalone tools regardless of your productivity suite. No lock-in to a specific ecosystem.
Do you need API access for custom integrations?
Yes: Claude (Anthropic API), ChatGPT (OpenAI API), and Gemini (Google AI API) all have robust APIs. Pricing is usage-based and varies. M365 Copilot has more limited API extensibility. Local open-source gives full API control via frameworks like Ollama or vLLM.
No: Stick to evaluating the web/app interface and team management features.
Does your team use SSO (Okta, Azure AD, Google Identity)?
Yes: All commercial options support SSO on their business/enterprise plans. Check that your specific identity provider is supported. M365 Copilot inherits your existing Azure AD setup automatically.
No: Not a differentiator. All options handle standalone accounts.
3. Team Size & Budget
How many users need access?
Under 50: Per-seat pricing on any Team plan is straightforward. At this scale, the per-user cost difference between tools is minimal relative to the productivity gain.
50-500: Negotiate. All vendors offer volume discounts at this tier. Get quotes from at least 2-3 providers. Mention that you're consolidating from shadow AI -- vendors respond well to this.
500+: Enterprise agreements unlock significant discounts (40-60% off list price). At this scale, M365 Copilot may have a cost advantage if you're already on E3/E5. Local open-source becomes cost-competitive when amortized across this many users -- but factor in infrastructure and ML engineering headcount.
What's your per-user monthly budget?
Under $20/user/month: Claude Team and ChatGPT Team are in this range at list price. Gemini Business may be included in your existing Workspace license. M365 Copilot is typically $30/user/month on top of existing licensing.
$20-40/user/month: All commercial options are viable. Focus on capability fit rather than cost.
Minimize ongoing cost: Local open-source has no per-user fees after initial infrastructure investment. Break-even typically happens at 100+ users over 12 months, but requires in-house expertise to maintain.
Pricing changes frequently. Always get a current quote from the vendor rather than relying on published list prices.
4. Primary Use Case Fit
What are the top 3 tasks your team will use AI for?
Writing & communication: Claude Team is strongest here -- long-form writing, nuance, and tone control are standout capabilities. ChatGPT Team is also strong.
Data analysis & spreadsheets: M365 Copilot (Excel integration) and Gemini Business (Sheets integration) have the edge when working directly in spreadsheet applications. For ad-hoc analysis, any tool works well with copy-paste workflows.
Research & synthesis: All options handle this well. If source citations matter, consider pairing your General AI Assistant with Perplexity Pro rather than switching your primary tool.
Image generation: ChatGPT Team includes DALL-E. Gemini includes Imagen. Claude does not have built-in image generation. If image generation is a frequent need and you want to minimize tool count, this matters.
Code assistance: If significant, evaluate separately using the Code Assistant Checklist.
How important is long-context processing? (Uploading large documents, lengthy conversations)
Critical: Claude Team leads here with a 200K token context window. Gemini offers up to 1M tokens on some models. ChatGPT Team and M365 Copilot have smaller context windows but are sufficient for most business documents.
Not important: Not a differentiator for your use case. Move on.
5. Admin & Rollout Considerations
How important is centralized admin control?
Very important: M365 Copilot has the most mature admin console (inherits Microsoft 365 admin center). ChatGPT Team and Claude Team both have team management dashboards. Gemini Business is managed through Google Workspace Admin.
Minimal: All business plans include basic team management.
How technically sophisticated is your average user?
Low technical literacy: M365 Copilot and Gemini Business have the lowest learning curve because they're embedded in tools people already use. No separate app to open.
Moderate: ChatGPT Team has the largest public knowledge base -- most employees have at least heard of ChatGPT, which reduces training friction.
High: Claude Team tends to be preferred by power users who value reasoning depth and instruction-following precision.
Quick Decision Matrix
| If your priority is... | Start with |
|---|---|
| Best writing & reasoning quality | Claude Team |
| Google Workspace integration | Gemini Business |
| Largest ecosystem + built-in image generation | ChatGPT Team |
| Microsoft 365 integration + admin control | M365 Copilot |
| Zero external data transfer | Local open-source |
| Lowest training friction | ChatGPT Team or M365 Copilot |
Still not sure?
Run a 2-week pilot with your top 2 choices. Give 10-15 users access to each tool for real work tasks, then compare satisfaction scores and actual usage. A pilot costs less than choosing wrong and re-migrating 6 months later.