There are dozens of AI tools and everyone has an opinion. Here’s a practical framework for choosing the right one for your work—based on actual use, not marketing.
The AI tool landscape is overwhelming. A new product launches every week, every vendor claims theirs is best, and the advice you find online is usually written by people who are paid to recommend something.
I’ve personally used ChatGPT, Claude, Microsoft Copilot, Cursor, and PMI Infinity. I use Copilot and Cursor daily at work. Here’s my honest take on what each tool is good for, who it’s actually for, and how to cut through the noise and make a decision.
Start Here: What Does Your Organization Approve?
Before evaluating any tool, answer this question first: what has your organization approved for work use?
This isn’t bureaucratic advice — it’s practical. Using a consumer AI tool for work on an organization that has an enterprise agreement with a different vendor creates data governance problems I’ve written about in Data Privacy in AI Projects and Your AI Usage Is Being Logged. Your organization’s approved tool is almost always your starting point, not your fallback.
If your organization has Microsoft 365, there’s a good chance Copilot is either already available to you or coming soon. If you’re in a smaller organization without an enterprise AI agreement, you’re likely choosing between consumer tools — which has different implications for how you handle work data.
Once you know what’s approved, then evaluate.
The Tools, Honestly
Microsoft Copilot for M365
Best for: PMs working in Microsoft 365 environments
Copilot is deeply integrated into the tools most PMs already live in — Teams, Outlook, Word, PowerPoint, Excel. That integration is its biggest advantage. You can ask it to summarize a Teams meeting you missed, draft a status report from your notes, or pull action items from an email thread without switching applications.
At Kwik Trip, Copilot is our enterprise AI tool. It operates within our Microsoft 365 tenant, which means our data governance and compliance framework applies. For PMs working in regulated industries or with sensitive data, that enterprise wrapper matters.
Limitation: Copilot is most powerful when you’re working within M365. If your organization uses Google Workspace or other tools, the integration advantage disappears.
Who it’s for: PMs in Microsoft-heavy organizations who want AI embedded in their existing workflow.
ChatGPT (OpenAI)
Best for: General-purpose AI assistance, brainstorming, drafting
ChatGPT is where most people start, and for good reason — it’s capable, accessible, and the free tier is genuinely useful for low-stakes tasks. The paid tier (GPT-4o) is significantly better for complex reasoning and longer documents.
For PM work, ChatGPT handles first drafts, brainstorming, risk identification, and communication polish well. It doesn’t integrate with your work tools, so everything is copy-paste — but for a lot of tasks, that’s fine.
Important caveat: The free and consumer paid tiers are not appropriate for work data that’s sensitive. If you’re using ChatGPT for anything beyond generic, non-confidential work, you need ChatGPT Enterprise — which has different data handling terms — or you should be using your organization’s approved tool instead.
Who it’s for: PMs who want a capable general-purpose assistant for non-sensitive tasks, or organizations with ChatGPT Enterprise agreements.
Claude (Anthropic)
Best for: Long documents, nuanced drafting, complex analysis
Claude handles long-context work exceptionally well — if you need to analyze a lengthy contract, summarize a 50-page requirements document, or work through a complex stakeholder situation in writing, Claude often outperforms the alternatives. It tends to be more careful about acknowledging uncertainty, which matters when you’re using AI output to inform decisions.
Like ChatGPT, the consumer version isn’t for sensitive work data. Claude for Enterprise has appropriate data handling for organizational use.
Who it’s for: PMs who do a lot of document-heavy work or need AI assistance with complex, nuanced communication.
PMI Infinity
Best for: PMP holders and PMI members wanting PM-specific AI assistance
PMI Infinity is PMI’s AI tool, built specifically for project management. It’s integrated with PMI’s body of knowledge, standards, and resources — which means it speaks PM fluently. Ask it about earned value, risk response strategies, or stakeholder engagement approaches and you get answers grounded in PMI frameworks rather than generic advice.
If you’re a PMP or active PMI member, it’s worth exploring. It won’t replace a general-purpose AI for broad tasks, but for PM-specific questions and professional development, it’s purpose-built in a way the general tools aren’t.
Who it’s for: PMPs and PMI members who want AI assistance grounded in PM methodology.
Cursor
Best for: Software development — not general PM work
Cursor is an AI-powered code editor. It’s genuinely impressive if you’re writing code — it understands your codebase, suggests completions, and can make changes across multiple files based on natural language instructions.
I use Cursor daily at Kwik Trip, but I want to be direct: Cursor is a developer tool. Unless you have software development experience or a lot of patience for a steep learning curve, it will frustrate you more than it helps you. It assumes you understand what it’s doing with your code. If you don’t, you can’t catch its mistakes — and it does make mistakes.
The exception for PMs: Cursor can be genuinely useful for assembling complex documents from multiple sources — pulling together meeting notes, Excel files, emails, and other inputs into a structured requirements document, for example. This doesn’t require development experience, but it does require patience with the tool’s setup and learning curve. If that’s your primary use case, it’s worth the investment. Just know that’s not what Cursor was designed for, and the learning curve reflects that.
Who it’s for: PMs with a software development background, or PMs who are willing to invest significant time learning the tool — particularly for complex document synthesis across multiple sources.
Claude Code
Best for: The same use case as Cursor, with a slightly gentler learning curve
Claude Code is Anthropic’s developer-focused AI tool, similar to Cursor in purpose. I find it slightly easier to get started with than Cursor — particularly if you take the time to set up the Claude environment properly (CLAUDE.md files, project context, memory). The setup investment pays off in more useful, consistent assistance.
That said, the same caveat applies: this is a tool for people comfortable with code. The learning curve is real, and “slightly gentler than Cursor” is a relative statement. If you’re a non-technical PM, neither Cursor nor Claude Code is your starting point.
Who it’s for: PMs with development experience who want an alternative to Cursor, or technically curious PMs willing to invest time in proper setup.
How to Choose: A Simple Framework
Step 1: What does your organization approve? Start here. Use the approved tool. If nothing is approved, flag it to your IT team and get clarity before proceeding.
Step 2: What’s your primary use case?
- Embedded in M365 workflows → Copilot
- General drafting and brainstorming → ChatGPT or Claude
- PM methodology and PMI resources → PMI Infinity
- Code and technical work → Cursor or Claude Code (if you have the background)
Step 3: What’s your technical comfort level? Consumer AI tools (ChatGPT, Claude consumer) have almost no learning curve. Enterprise tools (Copilot) require understanding your M365 environment. Developer tools (Cursor, Claude Code) require development knowledge. Be honest about where you are.
Step 4: What data will you be working with? If the answer is anything sensitive — employee data, financials, PII, proprietary information — consumer tools are off the table regardless of capability. Enterprise tools with appropriate data governance are your only option.
The One Thing Most PMs Get Wrong
They pick a tool based on what they read online, use it for two weeks at a surface level, decide it’s not that useful, and stop.
Every AI tool has a learning curve — not a technical one, but a prompting one. The PMs who get the most out of these tools are the ones who’ve learned to ask better questions: giving context, being specific about format and audience, iterating on outputs rather than accepting the first draft.
The tool matters less than the habit. Pick the one that fits your context, commit to using it for a real project, and learn as you go. The productivity gains are real, but they take more than two weeks to show up.
Links & References
Related Monday Business Posts:
- I’m a PM Who Uses AI Daily — what AI actually helps with in day-to-day PM work
- 5 PM Tasks AI Does Surprisingly Well — where to start if you’re building the habit
- Data Privacy in AI Projects — what to keep out of AI tools
- Your AI Usage Is Being Logged — what your organization can see
Tools Referenced:
The best AI tool for a PM is the one your organization approves, fits your actual use case, and that you’ll use consistently enough to get good at. Everything else is noise.