Google Gemini Prompt Generator for Task-Driven Chatbots

Leon Nicholls
10 min readJul 1, 2024


Are you excited about building your own AI chatbot but overwhelmed by the idea of writing the perfect prompts? Don’t worry — we’ve got your back!

Google Gemini is here to make your chatbot dreams come true. This cutting-edge AI is designed to create chatbots that can engage, inform, and even entertain. But here’s the secret sauce: the Google Gemini Prompt Generator. This handy tool streamlines the whole prompt-writing process, giving you a head start on crafting prompts that get the most out of Gemini. Whether a beginner or an AI pro, the Prompt Generator is your ticket to building task-driven chatbots that genuinely impress.

Note: This article spotlights techniques for the Google Gemini Advanced chatbot (a paid service). While these concepts also apply to the free version, we’ll focus on the enhanced capabilities offered by the Advanced subscription.

Why Customizing Google Gemini is Your Chatbot’s Secret Weapon

Alright, let’s get down to business. Why bother customizing Google Gemini instead of just using it straight out of the box? It’s simple: customization is the key to making your chatbot truly shine.

Consider it this way: a generic AI model is like a blank canvas. It has potential but needs more personality and a specific skill set. It’s a jack-of-all-trades, master of none. If you want your chatbot to excel at a particular task — whether it’s writing captivating stories, explaining quantum physics like a pro, or giving personalized shopping recommendations — you need to give it a bit of specialized training.

Customization is like tailoring a suit to fit perfectly. Wouldn’t you wear a one-size-fits-all suit to a job interview? The same principle applies to chatbots. A customized chatbot is like a well-trained expert with the knowledge and skills to handle specific user requests with finesse.

Here’s a quick example. Let’s say you want to build a chatbot for a travel agency. A generic chatbot can answer basic questions about destinations and prices, but a customized chatbot can go further. It could create personalized itineraries based on a user’s preferences, recommend hidden gems off the beaten path, and even help book flights and hotels. Talk about a travel companion that goes above and beyond!

Customizing Google Gemini allows you to create chatbots that understand user requests, anticipate their needs, and provide tailored responses. It’s like conversing with a knowledgeable friend who always knows exactly what you’re looking for. And let’s face it, who wouldn’t want that kind of experience from their chatbot?

Your Blueprint for Chatbot Brilliance

Now that you’re convinced that customization is the way to go, let’s explore how. The secret weapon in your prompt engineering arsenal is a well-structured template. Think of it as a blueprint that guides you through crafting prompts that get the most out of Google Gemini.

Here’s the prompt template we’ve found to be super effective:

You are a prompt engineer specializing in crafting instructions for Google Gemini to create AI chatbots. Your goal is to generate clear, concise, and effective instructions for a chatbot that will respond to the following user task:

[[User task]]

Format your response as follows:


* Begin with: “YOU ARE A…”



* Clearly state the primary goal of the task, aligning it with the AI’s defined role.


* **Action Verbs:** Use strong action verbs to guide the AI’s behavior (e.g., “Analyze,” “Generate,” “Provide,” “Explain,” etc.).

* **Contextual Clues:** Give the AI information about the user’s likely intent or the type of response expected.

* **Tone and Style:** Specify the desired tone and style of the response (e.g., “Be friendly and approachable,” “Maintain a professional and informative tone,” or “Use a humorous and lighthearted style”).

* **Specific Instructions:** Provide detailed instructions on the type of content to generate, including any style, tone, format, elements to include or avoid, and desired length.


* **Variables:** If the user task requires specific input, use variables to represent those inputs (e.g., {{userName}}, {{numberOfItems}}, {{topic}}).


* Describe the desired structure of the output, including any sections, elements, or formatting requirements.


* Include other relevant guidelines, constraints, or considerations (e.g., target audience, ethical considerations, specific AI limitations).


* Instruct the AI to generate safe, appropriate, and non-harmful content.


User task: Write a short story about a time traveler who visits ancient Rome.



TASK OBJECTIVE: Write an engaging short story about a time traveler who visits ancient Rome.


* Action Verbs: Craft a narrative that follows the time traveler’s journey.

* Contextual Clues: The story should focus on the time traveler’s experiences and observations in ancient Rome.

* Tone and Style: Use a descriptive and imaginative style to bring the story to life.

* Specific Instructions:

* Describe the time traveler’s initial impressions of Rome.

* Highlight encounters with historical figures or events.

* Explore the cultural differences and challenges the time traveler faces.

* Conclude with the time traveler’s reflections on the experience.


* (Optional) Incorporate the time traveler’s name: {{travelerName}}

* (Optional) Include a specific year in ancient Rome: {{year}}


* The story should be approximately 500–800 words in length.

* Use clear paragraph breaks and vivid language.


* Ensure the content is appropriate for all ages and avoids harmful or discriminatory stereotypes.

Note: If you are wondering how this prompt was designed, techniques discussed in previous blog posts (1,2) were used to generate a starting point, and then lots of testing and tweaking resulted in this version.

Now, let’s break down each section of this template and see why it works so well:

  • AI ROLE: This is where you set the stage for your chatbot’s personality. Is it a helpful assistant, a creative storyteller, a knowledgeable tutor, or something else entirely? Determining its role helps the AI understand how to approach the task.
  • TASK OBJECTIVE: This is the heart of your prompt. Clearly state what you want the chatbot to do. Do you want them to write a poem, solve a math problem, or summarize a news article? The more specific you are, the better the chatbot can deliver what you want.
  • CONTENT GUIDELINES: This section is all about fine-tuning your chatbot’s responses. Use action verbs to tell the AI what to do (e.g., “analyze,” “generate,” “explain”). Provide contextual clues to help the AI understand the user’s intent. Specify the desired tone and style (e.g., friendly, professional, humorous). Give detailed instructions on the type of content you want, including format, length, and any specific elements to include or avoid.
  • USER INPUT HANDLING: If your chatbot needs input from the user (like their name or a topic they’re interested in), use variables to represent that input. This makes your chatbot interactive and dynamic.
  • OUTPUT FORMAT: This is where you tell the chatbot how you want the response to look. Should it be a list, a paragraph, or a poem? Do you want it to use specific formatting like bullet points or headings? The more detailed you are, the more polished your chatbot’s responses will be.
  • ADDITIONAL INSTRUCTIONS: This is the catch-all section for any instructions you want to give the AI. This could be information about the target audience, ethical considerations, or any limitations you wish to place on the chatbot’s behavior.
  • ETHICAL REMINDER: This is a crucial final step. Remind the AI to generate safe, appropriate, and non-harmful content. We want our chatbots to be helpful and friendly, not offensive or inappropriate!

Why the Example Matters: This template’s included example is more than just there for show. It’s a crucial teaching tool, demonstrating the power of single-shot prompting. Think of it this way: LLMs like Google Gemini are incredibly smart but still learn best by example. By providing a single, well-crafted example within your prompt (like the short story prompt above), you’re giving the AI a transparent model to follow. This can significantly improve the quality and relevance of its output, even without needing a massive dataset of examples.

By using this template as your guide, you’ll be well on your way to crafting prompts that unleash the full power of Google Gemini and create intelligent and engaging chatbots.

Chatbot Engineering in Action

Alright, enough theory — let’s get dirty with real-world examples! We’ll use our trusty template to create prompts for three different chatbots: a storytelling chatbot, a tutoring chatbot, and a news summarization chatbot.

Example 1: The Storytelling Chatbot

Imagine you want to build a chatbot that can spin captivating tales on demand. Here’s how we’d use our template to bring this chatbot to life:

  1. Use the prompt template to create a customized power prompt for the user task of sci-fi storytelling.
  2. Take the generated prompt, tweak it if needed, and use it to turn Google Gemini into your own AI storytelling chatbot.
  3. Follow your chatbot instructions to provide any inputs or preferences for generating a story.
  4. Keep chatting with your chatbot to guide how the story plays out.

Example 2: The Tutoring Chatbot

Now, let’s say you want a chatbot that can explain complex concepts in a way that’s easy for anyone to understand:

  1. Use the prompt template to create a customized power prompt for the user task of tutoring a student on algebra.
  2. The generated prompt was tweaked by conversing with Gemini to consider the learner’s proficiency and age.
  3. Using the Gemini “Copy” menu option, the prompt markdown was copied into a new Gemini chat session to turn Gemini into a tutoring chatbot.
  4. Take the generated prompt, tweak it if needed, and use it to turn Google Gemini into your tutoring chatbot.
  5. Provide the student’s age and the question. The tutoring chatbot will then generate a detailed explanation.
  6. Keep the conversation going by asking about specific examples or deeper dives into particular scenarios.

Example 3: The News Summarization Chatbot

Last but not least, imagine you want a chatbot that can quickly summarize news articles, saving you time and effort:

  1. Use the prompt template to create a customized power prompt for the user task of news summarization.
  2. Using the Gemini “Copy” menu option, the prompt markdown was copied into a new Gemini chat session to turn Gemini into a tutoring chatbot.
  3. The content from “The Keyword” blog post was saved as a PDF.
  4. The PDF was then uploaded to the chatbot session, and the chatbot was instructed to summarize the blog post.

Note that using files as a chat context opens up many file processing and analysis cases.

These examples demonstrate the versatility of our prompt template. By simply adjusting the AI role, task objective, and content guidelines, you can create chatbots tailored to specific tasks and user needs.

Tips and Tricks from the Pros

Now that you’ve got the basics down, let’s kick things up a notch. Here are a few advanced tips and tricks that will help you become a true prompt engineering ninja:

1. The Iteration Game: Prompt engineering is more than a one-and-done deal. It’s an ongoing process of experimentation and refinement. Feel free to test different variations of your prompts and see how they impact the chatbot’s responses. Think of it as a conversation with the AI — you’re constantly learning and adjusting to get the best results.

2. Chain Reactions: Sometimes, more than a single prompt is needed to guide a complex conversation. That’s where prompt chaining comes in handy. It’s like giving your chatbot a series of instructions, one after the other, to help it navigate a multi-step task. For example, you could start by asking the chatbot to generate a list of ideas for a short story, then use a second prompt to ask it to write a detailed outline based on one of those ideas. This way, you’re breaking down a complex task into smaller, more manageable chunks.

3. Examples for the Win: As you get more comfortable with prompt engineering, you might want to explore even more advanced techniques like “few-shot learning.” This involves giving the AI a few examples of the responses you wish before you ask it to do the task. It’s like showing students sample essays before asking them to write their own. Another excellent technique is “reinforcement learning from human feedback,” where the AI learns by getting feedback on its responses from real people. These are just a few exciting possibilities as the field of prompt engineering continues to evolve.

The key takeaway is that prompt engineering is all about experimentation and continuous learning. The more you practice, the better you’ll become at crafting prompts that get the most out of Google Gemini. So, feel free to get creative and try out new things. You might stumble upon a prompt that revolutionizes how we interact with AI chatbots!


We’ve discussed why customizing your chatbot is like giving it a superpower, dissected a super-handy prompt template, and even gotten our hands dirty with real-world examples to show you how it’s done. But remember, the journey doesn’t end here.

Prompt engineering is a constantly evolving field; the best way to master it is through practice. So, go ahead and experiment! Take our template, tweak it, play with different ideas, and see what amazing chatbots you can create. Feel free to get creative and push the boundaries.

Check out my reading list of other Google Gemini articles.

This post was created with the help of AI writing tools, carefully reviewed, and polished by the human author.