Prompt engineering remains a top-of-mind topic for nearly everyone who creates content. As LLMs and proposal management platforms continue to develop, folks continually ask how they can best leverage technology to assist in the content and idea creation process. Just as we should apply different types of technology for various jobs and outcomes, prompts are certainly not one-size-fits-all.
In this article, we explore prompting strategies to help you collaborate more openly with your Gen AI platforms. You may see some of these methods as good for more creative and imaginative use cases, but you are unsure how they would apply to the RFP process. This is where we can think outside the box of content generation and use Gen AI to brainstorm, collaborate, evaluate, and explore new ideas.
1. Open-Ended Prompts
An open-ended prompt is an input or question that encourages broad, thoughtful, and unrestricted responses. Instead of asking for specific or factual answers, open-ended prompts invite exploration, imagination, and elaboration.
The goal is not to box the AI into a narrow task. Unlike more directive prompts, you want to leave room for the tech to take the lead. This type of prompt can be very valuable in creative projects or for brainstorming, positioning, speculative thinking, etc.
Characteristics:
- Non-restrictive: There is no right or wrong answer – best to explore possibilities.
- Exploratory: Encourages diving into themes, ideas, or scenarios without a fixed path.
- Flexible: Adaptable across industries and disciplines, from scenarios to future forecasting.
- Emotionally or Conceptually Rich: Often touches on abstract, nuanced, or imaginative ideas.
Example:
“Imagine we had unlimited time and resources to invest in our platform over the next year. Understanding our key functionalities, customers, and market, what could we build to improve our offering significantly?”
This is a very open-ended, hypothetical question. It is a very “show me the possibilities” line of questioning. It can help to incorporate AI into these “what if” scenarios to leverage their databases and ability to scour a massive amount of information quickly. These types of prompts can be very conversational and should allow the AI to stretch its wings while creating guardrails to make the output relevant. However, with this type of prompt, resist the urge to shape things too much, and you may be surprised by what you uncover.
Practical Application:
- Brainstorming: Exploring new ideas and approaches
- User Research: Crafting survey/assessment questions that reveal qualitative insights
- Strategic Thinking: Considering “what if” scenarios for long-term planning and delivery
- Coaching and Feedback: Assistance with prompting reflection, growth, or awareness
Best Practices:
- Start Broad: Use phrases like “Describe…”, “Imagine…”, or “What would happen if…”
- Avoid Binary Questions: Steer clear of yes/no or multiple-choice formats.
- Encourage Depth: Invite elaboration by asking why, how, or what else.
- Set a Tone or Context (Optional): Light framing can guide the direction without limiting the outcome.
- Leave Room for Interpretation: Don’t over-specify—let the responder explore creatively.
2. Constraint-Based Prompts
A constraint-based prompt sets clear rules or limitations on how the response should be generated. These constraints may involve length, format, tone, period, style, vocabulary, or character count. By narrowing the creative field, constraints push the AI to produce focused, efficient, and often more innovative content.
Think about starting a game of chess with some of your pieces missing. It is more challenging, but it forces you to think smarter about your approach.
Characteristics:
- Defined Boundaries: Includes specific limits such as word count, tone, format, or vocabulary.
- Focus-Driven Output: Reduces ambiguity and helps avoid generic or off-topic content.
- Creativity Within Limits: Encourages clever solutions and innovative thinking within set parameters.
- Efficient Responses: Forces AI to get to the point, often improving clarity and impact.
Example:
“Summarize our most important competitive advantages over XYZ company in 5 bullet points using less than 20 words per bullet. Be specific and focus on outcomes.”
This prompt clearly instructs the AI on the exact task that needs to be completed. You are giving the AI a narrow focus by asking for a very specific concept or outcome (identifying our most important competitive advantage) and providing constraints on length and structure. If this is for brainstorming early in an opportunity cycle, you could use this workflow to uncover key differentiators. If you are later in a cycle, you can use this workflow as a sanity check and quality control to ensure you effectively highlight those differentiators in your offering.
Practical Application:
- Legal and Compliance: Summarizing policies or addressing requirements using approved language or phrasing
- Education and Training: Writing questions and developing assessments with a defined format
- Technical Writing: Creating or communicating product specs with fixed length and terminology
Best Practices:
- Start Broad: Use phrases like “Describe…”, “Imagine…”, or “What would happen if…”
- Avoid Binary Questions: Steer clear of yes/no or multiple-choice formats.
- Encourage Depth: Invite elaboration by asking why, how, or what else.
- Set a Tone or Context (Optional): Light framing can guide the direction without limiting the outcome.
- Leave Room for Interpretation: Don’t over-specify—let the responder explore creatively.
3. Roleplay Prompts
A role-play prompt instructs an AI to respond as if it were a specific character, persona, or professional role. This style of prompt sets a clear identity or context for the AI to adopt, such as a CEO, an evaluator, a developer, or a regulator. It transforms a basic interaction into a more immersive, scenario-driven conversation.
Think of it as handing the AI a suit and script—it knows who it is, how to behave, and what perspective to speak from.
Characteristics:
- Persona-driven: The AI “acts” as a defined identity (e.g., technical writer, proposal evaluator, or solutions engineer).
- Contextual framing: Background or scenario is typically included (e.g., “You are a physician looking to create a care plan for a patient with limited mobility”).
- Consistent voice/tone: Maintains behavior, language, and style suitable for the role throughout the response.
- Dynamic interaction: Encourages back-and-forth conversations in-character.
Example:
“You are a proposal evaluator. You are looking for a solution requiring minimal training and long-term maintenance. Based on the supplied proposal, what feedback would you give the offeror on how their proposed solution meets those goals?”
This prompt gives the AI a role to play. In this situation, we are looking for mock feedback from the “evaluator” on how our proposal supports this need/ask. The goal here is to use AI like another set of eyes. When running these types of prompts, you will likely receive responses that can allow the team to look at your content in a new way. Using AI in the pre-submission evaluation process can help to sharpen the offering and better prepare for real-world mock evaluations.
Practical Application:
- Evaluation and Training: Simulating assessment scenarios or running a mock implementation
- Customized Content: Creating immersive and relevant content from the reader or evaluator’s point of view
- Objection/Feedback handling: Work mock scenarios in which a customer needs more information or documentation on your proposed approach
Best Practices:
- Define the role clearly: Be specific about the persona and their background.
- Set the context or situation: Provide a scenario or goal.
g., “You’re negotiating a contract renewal with a hesitant client.” - Specify tone or attitude (optional): Emotions and style can enhance realism.
- Watch for drift: If the AI drifts out of character, gently guide it back by reminding it of the role.
In summary, it is important to experiment and try new use cases in order to get the most out of your AI tools and platforms. Leveraging different methods can enable teams to enhance all parts of the content creation and editing process, but remember that the fun doesn’t end there. Working together on development, exploring new themes, evaluating your work, and challenging your ideas are all great ways to elevate your use of AI.
What’s next?
We would love to talk to you. Whether you want to learn more about how VT is harnessing AI, discuss best practices, or learn more about how to incorporate Generative AI into your process, we would love to chat. You can connect with our company socials on the left side of this page, or connect with me on LinkedIn for a casual chat.
You can also check out our resource center to download Guides and Whitepapers, as well as free recordings from our past Webinars and Events, for more Gen AI and prompt engineering content.
And finally, be on the lookout for content on AI implementation and change management later this summer.
