6 Best Practices in Using AI

Simple habits that help you get better, more reliable results

Artificial intelligence tools can be powerful, but the quality of the result still depends heavily on the quality of the request. Whether someone is using AI to draft content, summarize information, analyze data, generate ideas, troubleshoot code, or support strategic planning, the same principle applies: better input usually leads to better output.

At BSquare Advisors, we view AI as a support tool, not a substitute for judgment, strategy, or professional expertise. Used well, AI can help individuals and organizations move faster, clarify their thinking, and improve productivity. Used carelessly, it can create confusion, produce incomplete work, or reinforce errors that should have been caught before the final product.

The difference often comes down to how the user engages with the tool. Here are six best practices for using AI more effectively.

1. Provide Complete Context

AI performs best when it understands the full assignment. A vague prompt often produces a generic answer. A detailed prompt gives the tool more information to work with and helps it tailor the response to the actual need.

For example, instead of asking, “Can you help me write this email?” provide the audience, purpose, tone, background, and desired outcome. If the email is going to a senior leader, a student, a client, a tenant, a colleague, or a vendor, that context matters. The same message may need to sound more formal, warmer, more direct, or more diplomatic depending on who will receive it.

For coding tasks, context is even more important. Users should provide the coding environment, relevant error messages, the goal of the project, and the full relevant code snippet. Without that information, AI may troubleshoot the wrong issue or make assumptions that do not apply.

Good context includes:

The purpose of the task.
The intended audience.
The desired tone or style.
Important background facts.
Any constraints or non-negotiables.
Examples of what should be avoided.

The more complete the context, the more useful the response.

2. Be Specific and Concise

A strong AI prompt does not need to be long, but it does need to be clear. The best prompts are specific enough to guide the output without overwhelming the tool with unnecessary information.

For example, “Write a professional LinkedIn post about our new service” is useful, but it can be improved. A stronger prompt would be: “Write a 150-word LinkedIn post announcing BSquare Advisors’ new crisis management service. The tone should be polished, confident, and useful to small businesses, nonprofits, and education leaders. Avoid sounding sales-heavy.”

That prompt gives direction. It identifies the platform, length, subject, audience, tone, and what to avoid.

Being specific also reduces the need for repeated revisions. If the user wants a warmer tone, a shorter response, an executive-level summary, a more conversational draft, or a public-facing version, that should be stated upfront.

Specific prompts save time because they reduce guesswork.

3. Plan the Conversation

Before opening an AI tool, it helps to know what kind of assistance is actually needed. Many users begin with a broad request, then realize halfway through that they needed a different kind of support.

A little planning can improve the entire exchange.

Before starting, consider:

What is the goal of this task?
Do I need writing, editing, research, brainstorming, analysis, or formatting?
What information does the AI need to know?
What should the final product look like?
Will this be used internally, publicly, or professionally?

Planning the conversation is especially helpful when working on complex projects such as presentations, policy documents, marketing materials, grant narratives, interview preparation, business strategy, or communications plans.

AI can be useful in stages. A user might first ask for a structure, then ask for a draft, then ask for edits, then ask for a more polished version. That kind of intentional workflow often produces stronger results than asking for the final product all at once with limited context.

4. Outline Requirements Clearly

For writing assistance, the most effective prompts clearly explain the assignment. AI can produce very different results depending on the audience, format, purpose, and tone.

A request for a “professional bio” could result in a short website bio, a LinkedIn summary, a conference speaker bio, an executive biography, or a resume profile. Each one serves a different purpose. The user should define the intended use at the beginning.

Strong writing prompts include:

The target audience.
The purpose of the piece.
The desired tone.
The preferred length.
The key points to include.
Any language that must be avoided.
Whether the output should be formal, conversational, persuasive, warm, assertive, or concise.

This is particularly important for organizations. Public-facing content should sound aligned with the brand. Internal messages should reflect the organization’s culture. Client communications should be clear, accurate, and appropriately professional.

AI can draft, but the user must still direct.

5. Define the Research Focus

When using AI for research or analysis, the question must be clearly defined. Broad research requests often generate broad answers. A focused question produces a more focused response.

Instead of asking, “Tell me about trends in higher education,” a stronger prompt would be: “Identify current trends affecting student affairs offices at law schools, with a focus on student wellness, disability accommodations, crisis response, and communication expectations.”

That request gives the research a direction.

For research and analysis tasks, users should identify:

The specific research question.
The topic boundaries.
The time period.
The type of sources or data being used.
The desired format of the answer.
Whether the response should be descriptive, comparative, strategic, or analytical.

It is also important to remember that AI-generated research should be reviewed carefully. For current events, legal developments, policy changes, institutional data, pricing, regulations, or market conditions, users should verify information with reliable and up-to-date sources.

AI can help organize research, identify themes, and explain concepts, but verification remains essential.

6. Review Before Sending

AI can improve productivity, but it does not remove the need for review. Every AI-generated draft should be checked for accuracy, tone, completeness, and appropriateness before being used.

This is especially important for professional communications. A message may be grammatically correct but still too cold, too long, too vague, too defensive, or too informal for the situation. A summary may sound polished but omit a critical fact. A recommendation may seem reasonable but fail to account for the organization’s actual constraints.

Before using AI-generated content, review it for:

Accuracy.
Tone.
Completeness.
Audience fit.
Brand alignment.
Legal, ethical, or operational concerns.
Unintended implications.

The final judgment should always remain with the person or organization using the content.

AI should help sharpen the work, not replace accountability for it.

Final Thought

AI is most effective when it is used with intention. The strongest results come from clear direction, complete context, focused questions, and thoughtful review.

For businesses, nonprofits, educational institutions, and individual professionals, the real opportunity is not simply using AI more often. The opportunity is using it more strategically.

When used well, AI can help clarify ideas, strengthen communication, improve efficiency, and support better decision-making. But like any tool, its value depends on the judgment behind it.

At BSquare Advisors, we believe effective communication begins before the message is drafted. It begins with clarity of purpose, understanding of audience, and attention to detail. Those same principles apply when working with AI.

Better prompts produce better outputs. Better review produces stronger work. Better strategy produces better results.