GenAI Prompting 101: How to Design Better Conversations in Mainstay

Picture this: You’ve just opened Mainstay’s new GenAI in Scripts tool, eager to harness the power of Generative AI (GenAI) for your learners. Your cursor blinks on a blank prompt field. Your mind goes equally blank. “What do I even type?” you wonder.

If this scenario feels familiar, you’re in good company. Many of our partners feel a mix of excitement and “prompting paralysis” when they first start writing prompts for GenAI. It’s a bit like learning to ride a bike – wobbly and awkward at first, but easier (and even fun) once you get the hang of it.

GenAI – shorthand for Generative AI – is essentially an AI system that can create original responses on the fly. In Mainstay’s platform, the GenAI in Scripts feature allows you to drop a GenAI “brain” into your student messaging flows. It’s powerful stuff: your chatbot can go from scripted Q&A to fluid, open-ended conversations that feel more human. But unlocking that power starts with one key skill: writing an effective prompt.

This guide will walk you through the process, and by the end, you’ll be crafting prompts with confidence. Let’s turn that blank screen into your new favorite tool.

This Isn’t ChatGPT. It’s Smarter.

When you prompt GenAI inside Mainstay, you’re not starting from scratch. You can design prompts that:

• Pull from your knowledge base, trusted URLs, or internal documentation

• Incorporate contact record data, like enrollment status or declared major

• Leverage past conversations to keep context alive

• Classify open-ended responses and update fields in real time

That’s a big leap from generic chat to intelligent, institutional support.

New to Prompting? You’re Not Alone

Stepping into GenAI prompting can feel weird at first. We’ve heard common questions from partners like:

• “I’m not technical. What if I write the prompt wrong?”

• “Can I really just… ask it to do things? That feels like sci-fi!”

• “How much detail is too much?”

If you can write an email or explain something to a student, you can write a great prompt. Prompting is simply giving clear instructions to your AI assistant.

The 4 Building Blocks of a Great GenAI Prompt

Writing a prompt is easier when you include these four elements: Persona, Learner, Tone, Task.

1. Persona (Who is the AI?)

“You are a warm, helpful academic advisor…” Give your GenAI a character. This makes responses feel more consistent and believable.

2. Learner (Who are they helping?)

“You’re talking to a high school senior who’s nervous about orientation.” Include a short description of the student’s level, mindset, or challenge.

3. Tone (How should it sound?)

“Use an encouraging and upbeat tone.” This ensures your AI mirrors your institutional voice and feels human.

4. Task (What should the AI do… or not do?)

“Explain the next two steps and offer encouragement if they’re feeling stuck.” This is where you define the purpose of the message and any boundaries.

Put together:

You are a warm, knowledgeable financial aid advisor. You are talking to a first-generation freshman who is anxious about filling out the FAFSA. Use a reassuring and upbeat tone to guide them through the next three steps of the FAFSA process and encourage them that they’re on the right track.

Advanced Use: Prompting for Categorization

One of Mainstay’s most powerful features is the ability to categorize open-ended student responses. For example:

Prompt: You are a thoughtful student success coach. You are speaking to a returning adult learner. Ask what their biggest challenge is right now when thinking about coming back to school. Based on their response, categorize it as one of the following: Financial, Academic, Personal, or Other.

This kind of prompt doesn’t just spark a conversation—it writes data back to the student’s record, enabling smarter follow-up and targeting.

Better Prompting, Better Conversations

Mainstay’s GenAI isn’t meant to generate essays—it’s built to hold intelligent, student-centered conversations. Here’s what a great GenAI interaction might look like:

Prompt: You are a calm and supportive advisor. The student has completed their application but hasn’t signed up for orientation. Respond conversationally and ask if they’re planning to attend orientation. If they’re unsure, offer one benefit and ask what’s holding them back. Be brief and leave space for them to share their concerns.

AI Message: “Hey Jaya, Pounce here I saw you wrapped up your application—nice work! Orientation’s coming up soon. Are you thinking about signing up? If you’re unsure, I can share why students usually find it helpful. What’s on your mind?”

This tone invites trust, nudges action, and gives students the chance to share in their own words.

Steer Clear of Common Prompting Pitfalls

Now that you have the basics, let’s cover a few “gotchas” that new prompters often encounter. Avoiding these common pitfalls will save you time and ensure your GenAI chats are effective and ethical:

Being Vague or Too General
If your prompt is too thin on details, the AI might respond with generic or off-target info.

“Help the student with financial aid.”
That’s a well-intentioned prompt, but it’s wide open to interpretation. Always aim to be specific about the context and goal — for example:
“Help the student understand two next steps for their financial aid application.”
In short, more clarity in = more clarity out.

Writing a Novel (Too Much Info)
On the flip side, overloading the prompt with unnecessary details or long-winded background can confuse the AI. You don’t need to paste the student’s entire life story or every nuance of federal loan policy. Stick to the key points the AI needs to know.
Think of it like giving directions — clear and concise beats rambling every time.

Forgetting to Specify Tone
Tone might seem like a “nice to have,” but it’s actually crucial. If you skip telling the AI how to sound, it may default to a neutral or generic tone that doesn’t match your audience.
For example, a prompt that lacks tone guidance might yield a response that’s technically correct but feels cold or robotic. Always include a simple tone cue — friendly, enthusiastic, calm — to humanize the AI’s answer.

Using Jargon or Internal Language
Remember, you’re writing instructions for an AI that will talk to students. If your prompt includes a lot of institutional jargon or complex language, the AI might echo that and confuse your learners.
Write your prompts in plain, conversational language — the way you’d explain something to a student in person.

Instead of: “Articulate the matriculation requirements.”
Try: “Explain what steps the student needs to take to enroll.”
The AI will mirror your language style.

Not Reviewing the AI’s Output
This isn’t exactly a prompt-writing mistake, but it’s a critical step: always test and review the AI’s response.
After you write a prompt, try it out! Ask yourself:

• Does it align with the guidance I provided?

• Is the tone right?

• Are the facts accurate and appropriate?

If something’s off, tweak your prompt and try again. Prompting is an iterative process — even pros revise multiple times. A quick review also helps you catch any unexpected or off-brand responses. (Mainstay’s GenAI has guardrails in place, but it’s always smart to give it a once-over.)

Ignoring Privacy and Boundaries
As you get more comfortable with GenAI, it’s easy to get excited about all the things you can ask it to do. But remember to respect student privacy and ethical boundaries.
Don’t include sensitive personal data about a student in the prompt — you can describe a situation generally without using names or IDs. Keep the AI focused on relevant, supportive dialogue — academic and career questions are great; personal opinions or medical advice are not.

If you stay within appropriate bounds, GenAI becomes a trustworthy assistant that augments your work — not a loose cannon.

Your GenAI Prompt Builder Cheat Sheet

We’ve covered a lot, so we put together a one-page visual summary of everything you need to write strong, effective GenAI prompts. This downloadable checklist is perfect for quick reference or sharing with teammates during a brainstorming session.