When using AI systems, what are some best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals?

Great question, Sarah! As PMs, we’re responsible for ensuring that decisions and results are accurate, relevant, and aligned with the company's goals. That requires a deliberate and structured approach—especially when using AI systems. A few best practices I follow:

#1 Start with a Clear Objective — Vague inputs yield vague outputs
I begin by framing my prompt with context and intent—just like a strong project charter. Then, I let the AI know the why behind the request.

#2 Refine Iteratively — Treat prompting as a dialogue, not a one-and-done command
I often use a “draft–review–revise” cycle to narrow the output until it’s on target because I've found that iteration leads to precision.

#3 Verify with Trusted Sources — AI can hallucinate, especially with niche or time-sensitive topics
If the output contains factual data or citations, I cross-reference the information with primary or verified sources.

#4 Use AI for What It Does Best — It’s a strategic assistant, not a substitute for human judgment
I find AI most valuable for ideation, synthesis, and summarization—not final decision-making.

#5 Document Prompt Structures That Work — Build consistency, scale for effort
Just like project templates, I keep a repository of effective prompt formats for everything from stakeholder emails to workshop planning guides.

Recently, I used ChatGPT to co-develop a stakeholder workshop at a commercial insurance firm. I started with a general objective—analyzing current-state process flows—and used prompt engineering to clarify scope, co-create agenda formats, and generate facilitation guides. What made the output valuable wasn’t just the AI’s speed—it was the structure I brought to the process: specific constraints, organizational context, and iterative refinement. That’s what ensured the deliverables were aligned, high-quality, and immediately usable.

I'm curious to hear how others are building reliability into their AI workflows too—what’s been working for you? 😊

Keep it simple:

1-Clean your data.

2-Ask specific questions.

3-Review the results.

The exact approach will vary with the AI tools you choose and the goals you want to achieve. Once those elements are clear, you’re all set.

Be specific about what you want the AI to do.

Include context, constraints, and desired output format.

Use precise language and structured instructions.

Provide examples or data when possible to guide the AI.

Cross-check AI output against trusted sources or human experts.

Use critical thinking to assess factual accuracy and logic.

Treat AI use as an iterative process. Refine prompts based on earlier results.

Ask follow-up questions or request clarifications.
 

Combine your expertise with AI insights to filter out irrelevant or incorrect outputs.

Don’t rely solely on AI—use it as a support tool.

For data analysis or tasks like summarization, structured inputs (e.g., bullet points, tables) improve clarity and accuracy.

Understand the AI’s training scope and possible biases.

Know that AI might generate confident-sounding but incorrect answers (hallucinations).

Avoid sharing personal or confidential data unless you're using a secure, trusted environment.

2. Craft Effective Prompts