Refining Your Search Strategy With AI

 Introduction

In this chapter, we explore the concept of research as inquiry, a dynamic process that encompasses scholarship as conversation, strategic exploration, scaffolding, and transference. These elements work together to refine the development of an effective search strategy. In this section, we focus on practical approaches to crafting strong prompts to leveraging AI tools to generate strategic keywords for database searches, identifying relevant theoretical frameworks enhance the depth and complexity of your research.

AI does not replace the core concepts we’ve explored in this chapter. Instead, we encourage students to approach AI as a collaborative companion throughout the research process. As noted in the chapter’s introduction, this book is designed with intentional scaffolding to support skill development and transference, the ability to apply what you’ve learned across different contexts. While AI cannot substitute the importance of these foundational practices, it can enhance your research experience when used thoughtfully and with clear intent.

Let’s discuss how this works in practice.

Generating a Search Strategy

There are many guides offering useful tips for leveraging AI to create successful search strategies. With so much information available, it’s easy to feel overwhelmed, so we’ve synthesized some of the best resources into a clear and focused approach. Throughout this course, we treat AI as a creative and critical companion to your research, not a substitute for your intelligence, innovation, and drive to do meaningful research.

Further, we advise against using AI to locate sources for your paper. It has been well-documented that AI systems are prone to “hallucinating” citations that appear credible at first glance but don’t actually exist. Instead, we focus on using AI strategically to refine your research strategy. You will still need to use subject databases, academic journals, books, Google Scholar, and other vetted sources to locate high-quality scholarly materials. The strategy we discuss here is designed to help you navigate those sources effectively by developing strategic keyword searches, participating in the scholarly conversation, identifying gaps in the existing research, and other issues we discuss throughout this course.

Crafting Effective Prompts

One of the initial challenges in using AI effectively is learning how to generate strong, targeted prompts. Harvard University has created an outstanding guide on prompt design that offers practical strategies for creating a search strategy. Their approach emphasizes specificity, clearly articulating what you do and don’t want in a response and employing “Act as if…” scenarios in which you assign the AI a defined role to help generate more focused and relevant outputs. This approach helps create more tailored, relevant outputs by simulating expertise, perspective, or situational awareness. They also highlight techniques such as adjusting for tone, audience awareness, and building on previous prompts: all concepts we discuss throughout this course. For a deeper understanding of how these components function in practice, we encourage you to consult the guide directly.

Similarly, the University of Michigan has developed a resource titled Prompt Literacy in Academics, which extends beyond search strategies to explore the broader role of prompt design in scholarly work. While its scope is wide-ranging, its emphasis on prompt literacy is relevant to our task. One particularly useful feature is the Prompt Library, in which users can browse real chat transcripts. Those examples illustrate how well-crafted prompts can sharpen keyword selection and enhance your overall search strategy.

Both the Michigan and Harvard guides emphasize the importance of the active role of the researcher in shaping AI interactions. This perspective reinforces the idea that effective prompting is a skill rooted in intentionality, clarity, and strategic thinking. Now, let’s put some of these principles into practice by developing a sample search strategy that demonstrates how thoughtful prompt design can enhance your engagement with scholarly sources.

Before we begin, it’s important to remember that this course already includes a range of proven research strategies that do not rely on AI. These are all essential tools, and we encourage you to incorporate them into your personal repertoire. We are primarily interested in how AI can enhance—not replace—these foundational methods by serving as a creative and critical companion. For instance, in later chapters, we will explore how AI can help refine a research question or sharpen a thesis statement. However, before reaching that stage, we need to create a search strategy, so let’s explore how CoPilot can help create strategic keywords for effective database searches and identify a relevant theoretical component for our analysis.

Locating Resources

As noted in the introduction, it’s important to exercise caution when asking AI to locate sources. Many AI tools are known to generate citations for sources that don’t exist or blend elements from multiple sources into misleading composites. For this reason, we discourage asking AI to retrieve a set number of sources and relying on whatever it produces. Instead, approach AI as a brainstorming partner to help generate effective keywords and refine your search strategy. These keywords can then be applied to trusted academic databases, journals, books, and other vetted resources to locate high-quality scholarly materials. To illustrate this approach, let’s walk through a sample search strategy using techniques from the Harvard prompt guide discussed earlier.

Be Specific: I am writing a 4-to-six-page research paper for an upper-level undergraduate literature course on the James Baldwin short story “Sonny’s Blues.” I don’t have a thesis statement yet, but I’d like to explore Baldwin’s treatment of jazz music and racial identity in the story. Please provide a list of ten possible keywords or short phrases that I can use to search an academic database for scholarly resources.

 “Act as if…”: According to the Harvard guide, asking the AI to behave as if it were a type of person, process, or object can be an easy way to start generating better prompts. The AI will attempt to emulate that role and tailor its answers accordingly. So, let’s try adding this component to our prompt:

“I’m writing a literary analysis on James Baldwin’s short story ‘Sonny’s Blues’ for an upper-level undergraduate literature course. While I haven’t finalized a thesis statement yet, I’m interested in exploring Baldwin’s treatment of jazz music and racial identity within the story. Please generate a list of ten keywords or short search phrases I can use to locate scholarly sources in academic databases. Act as if you are my professor, encouraging me to develop a unique and insightful interpretation. The keywords should reflect both thematic depth and critical angles that could help me refine my research focus.”

This prompt yields a useful list of keywords as well as literature databases:

Some Notes About This Response:

Keywords: The list of keywords is relevant and would be good starting points. However, be open to changing them once you get into the databases. This is part of the idea of Searching as Strategic Exploration that we discuss in chapter three. There are several short videos and a discussion of the role of strategic keywords in inquiry that would be useful to review here.

Databases: The list of suggested databases is also good, but we would add Academic Search Premier to the mix.

Don’t Rush! We talk throughout this course about how research is an iterative process that involves a lot of back and forth as you deepen your understanding of your topic. However, AI tools tend to present things in a linear fashion that may be counterproductive. For example, in the response above, CoPilot is offering to convert those keywords into a search strategy and start crafting a thesis before you’ve read the literature.

While it might be tempting to follow the suggested next steps, this would be a good time to slow down, explore the databases, gather relevant sources, read the literature, get a sense of the scholarly conversation, and look for something that sparks your interest. Conducting a literature review can’t be rushed.

This would be a good place to review Chapter 3: Research as Inquiry and Scholarship as Conversation and Chapter 8: Evaluating Scholarly Resources. Chapter 8 includes useful information about reading like a researcher, active reading, and note-taking. Don’t rush through these important steps.

The Theoretical Component

Once you’ve spent some time exploring the relevant databases, journals, and other resources, and you have a sense of where your interests fit within the larger scholarly conversation, the next step would be to add a theoretical framework to your analysis. This should be organically derived from your research, but this is an area where CoPilot can be of use. Let’s begin by creating another prompt:

“I’m writing a literary analysis on James Baldwin’s short story ‘Sonny’s Blues’ for an upper-level undergraduate literature course. While I haven’t finalized a thesis statement yet, I’m interested in exploring Baldwin’s treatment of jazz music and racial identity within the story. I’ll need to include a theoretical component in my analysis but I’m unsure which ones are relevant. Can you provide a list of possible literary theories that I can explore? Act as if you are my professor, encouraging me to develop a unique and insightful interpretation. The keywords should reflect both thematic depth and critical angles that could help me refine my research focus.”

Copilot’s Response:

CoPilot provides a list with a brief explanation of how each theory applies to your topic. From there, it makes further suggestions about how you can select one or two of the options provided and “riff” on how they would work in your paper:

Exercises

Keywords are the tools researchers use to find relevant research materials. Our objective in this exercise is to learn the importance of reflecting on strategic keywords before searching databases and journals. This reflection is an important step that many people skip, but taking time to reflect will yield higher quality results. This exercise will demonstrate how AI can help us through this process.

Activity Overview: We will be doing an A/B comparison between human-generated and AI-generated keywords with the understanding that we can combine the best terms from both and continue refining our terms later.

First Step: Start by brainstorming a robust list of keywords to guide your database and journal searches. Don’t worry about filtering or refining at this stage; just jot down any terms that come to mind. The goal is to capture a wide range of ideas that you can narrow and refine later.

Next Step: Design a prompt that instructs CoPilot to generate a list of keywords. Apply the principles of effective prompt engineering outlined in Chapter 1. For example, ensure your inputs are specific, contextually rich, and precisely worded. Be clear about what you want the AI to include or exclude and consider using “Act as if…” scenarios to assign CoPilot a defined role. This technique helps CoPilot create focused responses and increase the relevance of the output.

Here are Jada’s keywords:

• Keyword A: James Baldwin
• Keyword B: Sonny’s Blues
• Keyword C: African American literature/writers
• Keyword D: African American short stories
• Keyword E: Marxism
• Keyword F: Race/Racism
• Keyword G: Ethnic/black identity
• Keyword H: Masculinity/African American Men
• Keyword I: Autobiography
• Keyword J: Civil Rights
• Keyword K: Ethnomusicology
• Keyword L: Jazz/Harlem Renaissance
• Keyword M: Race relations
• Keyword N: Gender studies/sexuality
• Keyword O: Activism/social justice

Here is the AI keyword prompt we wrote:

I am preparing a literary analysis of James Baldwin’s short story “Sonny’s Blues” for an advanced undergraduate literature seminar. My goal is to move beyond surface-level interpretations and explore the story’s deeper thematic, cultural, and structural complexities. To support my research, I need a list of ten nuanced and academically effective keywords that will help me locate scholarly articles, critical essays, and theoretical frameworks in databases and journals. Please generate keywords that reflect Baldwin’s engagement with race, trauma, music, familial relationships, urban life, and narrative technique, and that will guide me toward sophisticated and interdisciplinary scholarship.

CoPilot-generated keywords:

1. Black existentialism in postwar American literature
   → Explores philosophical themes of identity, suffering, and meaning in Baldwin’s work.
2. Narrative temporality and memory in African American fiction
   → Useful for analyzing Baldwin’s use of nonlinear storytelling and flashbacks.
3. Musical aesthetics and cultural resistance in literature
   → Connects Sonny’s jazz/blues to broader themes of survival and protest.
4. Urban space and racialized subjectivity in Harlem literature
   → Helps frame Harlem as both a setting and a psychological landscape.
5. Familial estrangement and reconciliation in Black narratives
   → Focuses on the emotional and symbolic dynamics between the brothers.
6. Silence, voice, and trauma in James Baldwin’s fiction
   → Investigates how Baldwin uses silence and speech to convey pain and healing.
7. Addiction, incarceration, and systemic violence in literary representation
   → Intersects with sociopolitical readings of Sonny’s struggles.
8. Double consciousness and identity formation in African American literature
   → Applies Du Boisian theory to Baldwin’s characters and themes.
9. Ethics of witnessing and narrative empathy in Baldwin’s storytelling
   → Examines the narrator’s role as a moral and emotional witness.
10. James Baldwin and the aesthetics of resistance
    → Broadens your scope to include Baldwin’s political and artistic strategies.

Notice that the CoPilot results tend to be more detailed and that they include an additional line about the relevance of the keywords. While there is significant overlap with Jada’s terms, CoPilot also missed certain of Jada’s keywords such as Civil Rights, Masculinity/African American Men, and Gender Studies/Sexuality. At this point, we may select and combine the most promising keywords from each list as we reflect on the directions we want our inquiry to go.

Remember that crafting thoughtful prompts is a reflective process. It’s not just about getting better results from CoPilot; it’s also a way to clarify your own thinking. The more intentional and nuanced your prompt, the more insightful and actionable the AI’s response will be.

As your research evolves, you will revisit and adjust your keywords to reflect new insights and directions. This iterative process is key to deepening your understanding and improving the quality of your results.

Reflection activity:

1. How much overlap was there between your keyword list and CoPilot’s? Were there any exact matches or closely related terms? What does this tell you about your initial approach?
2. Did CoPilot suggest any keywords that hadn’t occurred to you? If so, what made those terms stand out? Were they broader, more specific, or from a different disciplinary perspective?
3. How relevant and useful were the keywords generated by CoPilot? Would you consider incorporating any of them into your search strategy? Why or why not?
4. What did this comparison reveal about your own keyword selection process? Did it highlight any gaps, assumptions, or areas for improvement?
5. How might you revise your original list based on this exercise? Consider how your research focus or search results might shift with these new terms.

Final Thoughts

You can continue refining your list until you feel confident in your selection of keywords, recommended databases, and relevant theoretical frameworks.

Once you’ve built a solid foundation, CoPilot will suggest next steps such as writing your introduction or crafting a thesis statement. As noted above, it can be tempting to jump ahead, but we encourage you to wait until you’ve engaged with your sources and gained a clearer understanding of how your ideas fit within the broader scholarly conversation. Committing to a thesis too early can limit your intellectual flexibility and prevent you from discovering more compelling or nuanced arguments. For further guidance, refer to our chapters on developing thesis statements and formulating research questions.

We have stressed that research is an iterative process that benefits from flexibility, reflection, and ongoing engagement with scholarly conversations. Remain open to revisiting and refining your approach as your research evolves. Avoid locking into fixed ideas or interpretations too early. Allow your understanding to deepen as you critically engage with the literature and perspectives surrounding your topic. This openness will strengthen your final work while reinforcing the transferable research skills this book is designed to cultivate.