This article demonstrates how a Google Sheets-based prompt generator can support teachers in using GenAI to create intentional, genre-aligned mentor texts for multilingual learners. Drawing on the WIDA 2020 ELD Standards Framework and genre-based pedagogy, the article includes elementary classroom examples that model how teachers can use structured prompts to generate mentor texts aligned to language and content goals.
Keywords: artificial intelligence, generative artificial intelligence, multilingual learners, WIDA, functional approach, language features, language functions, elementary
The transformative potential of generative artificial intelligence (GenAI) in multilingual learner (ML) settings is related to pedagogical decisions of users. GenAI tools can be an effective classroom asset only through the intentional design of teacher-led inputs. Moving away from the idea of prompting as a spontaneous or trial-and-error task, prompt engineering is a disciplined instructional act (dos Santos, 2025). This approach demands that educators bridge the gap between technological capability and student needs by meticulously considering the classroom context, the specific parameters of the request, such as tone and audience, and the targeted learning outcomes. This article puts these ideas into practice to propose intentional and purposeful uses of GenAI in ML education.
Scaffolding prompt design through a planning tool
Effective use of GenAI in ML classrooms depends not on the technology itself, but on the quality and intentionality of the prompts that teachers design (de Oliveira & dos Santos, 2025; dos Santos, 2025). Rather than viewing prompts as ad hoc or improvised inputs, prompt engineering should be an instructional practice that is grounded in the instructional context, the key characteristics of the request (including design, intended audience, and thematic focus) and the purpose of the prompt, or what the teacher aims to achieve through the generated text (dos Santos, 2025; dos Santos et al., 2025).
Central to this work is the Prompt Creation Reference Chart for Mentor Texts, developed by de Oliveira and dos Santos (2025), which outlines a set of guiding questions teachers can use to design prompts that yield instructionally aligned mentor texts for multilingual learners (MLs). The chart is grounded in Systemic Functional Linguistics (SFL), a theory that views language as a functional tool to convey messages in specific social contexts (Halliday & Matthiessen, 2014). By aligning with both SFL and the WIDA (2020) English language development standards framework, the chart helps teachers convert complex language functions, such as explanation or argumentation, into precise GenAI prompt input. This approach reflects an understanding of genre and language to guide teachers in designing prompts that support meaning-making in instructional contexts (Halliday & Matthiessen, 2014). The chart prioritizes key instructional decisions, such as instructional context, genre family, topic, language features, and instructional purpose, before any interaction with GenAI occurs. Because GenAI relies on clear parameters to produce high-quality output, SFL’s emphasis on genre families and register ensures that AI-generated mentor texts mirror the most prominent linguistic features students will need to develop for disciplinary success. In doing so, it aligns teachers’ work with the WIDA ELD standards framework components and sample planning questions, which are grounded in content-based language learning and frame language and content as integrated components of instruction through explicit attention to language functions, language expectations, and proficiency levels in lesson and unit planning (Shafer Willner, 2025; WIDA, 2020). Through the lens of the SFL theoretical framework, the prompt generator introduced in this article acts as a bridge, turning theoretical linguistic principles into automated instructional support for teachers of MLs.
The Prompt Creation Reference Chart for Mentor Texts (de Oliveira & dos Santos, 2025) provides a strong planning framework for classroom teachers. We take the chart to further apply it to daily planning routines, based on our elementary classroom teaching experiences. We developed a Google Sheets-based prompt generator that transforms the chart’s guiding questions into an automated, teacher-friendly workflow. In this tool, each guiding question from the Prompt Creation Reference Chart for Mentor Texts is embedded as a column header in a Google Sheet (see Figure 1). Teachers respond to each question in a single row, entering information such as their instructional role, grade level, instructional setting, content topic, genre family, desired word count, targeted language features, and instructional purposes. A formula then automatically synthesizes these responses into a complete, coherent prompt that can be copied and pasted directly into a GenAI platform.
This prompt structure builds on the Prompt Creation Reference Chart for Mentor Texts described by de Oliveira and dos Santos (2025), which we adapted in a Google Sheet for use as an automated prompt generator. Please save a copy of the Prompt Creation Reference Google Sheet to be able to edit this tool.
I am a [teacher role] teaching [grade level] in the [instructional setting]. My class is learning about [topic]. Generate a text in the genre family [genre family]. The text needs to provide [information type] about [content focus]. The text should be no more than [word count] words and will serve as a mentor text for my students. Its language features associated with the genre family [genre family] include: [language features]. Generate this text bearing in mind these purposes: [instructional purpose]. Before launching the prompt, ask me questions about any additional information you may need to complete this task.
Figure 1. Prompt Creation Reference Chart for Mentor Texts Google Sheet
This automation does not replace teacher expertise; rather, it amplifies instructional decision-making by ensuring that each pedagogical choice is explicitly represented in the prompt. By scaffolding prompt writing into manageable steps, this tool reduces the cognitive load associated with prompt construction, enabling teachers to focus more fully on refining instructional intent while supporting ethical engagement with GenAI and improving the quality of AI-generated responses (dos Santos, 2025).
The design of the prompt generator also aligns closely with Fitzpatrick’s PREPare framework for effective AI prompting, which emphasizes Prompt, Role, Explicit instructions, Parameters, and Evaluation (Fitzpatrick, 2023). This emphasis on structured, principled prompt design is supported by research on large language models demonstrating that more precise tasks and directives lead to more effective and aligned model outputs (Liu et al., 2026). Notably, the Prompt Creation Reference Chart for Mentor Texts already incorporates several core PREPare components – Prompt, Role, Explicit instructions, and Parameters – through its emphasis on the teacher’s role, task characteristics, and instructional purpose in prompt design (de Oliveira & dos Santos, 2025; dos Santos, 2025). The Google Sheets-based prompt generator builds on this foundation by making these elements actionable within a planning tool and by extending the framework to include an explicit Evaluation step.
Prompt is articulated through the instructional task description and text expectations.
Role is specified by positioning the teacher and the instructional context.
Explicit instructions are embedded in the genre family, content focus, and selected language features.
Parameters are established through word count and genre constraints.
Evaluation is incorporated through the addition of an ARE step (Ask questions, Rate, Emotions), implemented by prompting the GenAI system to request additional information before text generation (e.g., “Before launching the prompt, ask me questions about any additional information you may need to complete this task”).
By embedding PREPare principles directly into the structure of the spreadsheet, the tool supports teachers, particularly those new to GenAI, in producing consistent, high-quality prompts while maintaining instructional intentionality without requiring extensive technical knowledge.
Leveraging prompt design prior to GenAI use
The Google Sheets tool is designed to be used before teachers interact with any GenAI platform. This workflow reflects a key insight: prompts function most effectively when they are deliberately planned and structured before interaction with GenAI, rather than generated spontaneously in the moment (dos Santos, 2025). Teachers are encouraged to revisit and revise their responses within the spreadsheet, particularly the genre family, language features, and instructional purpose, prior to copying the generated prompt. The generated prompt remains fully editable, allowing teachers to make final adjustments before using it with a GenAI platform. This step serves as a built-in reflection point. Teachers may ask:
- Are these the language features I want to be made prominent in the mentor text to support the intended Key Language Use?
- Do the selected language features reflect the language I want students to notice, discuss, and practice through the mentor text?
- Does the instructional purpose align with the language function students are expected to practice?
- Is the word count sufficient for a mentor text?
Engaging in this reflection process reinforces the irreplaceable role of the educator as an active reviewer, particularly in verifying content accuracy, ensuring instructional relevance, and continually refining interactions with AI tools (Shafer Willner, 2025). By treating the spreadsheet as a planning space rather than a mere technical tool, teachers engage in intentional prompt engineering that remains grounded in pedagogy.
Unstructured use of AI risks reproducing generic texts that fail to meet multilingual learners’ linguistic and cultural needs (de Oliveira & dos Santos, 2025). The Google Sheets-based prompt generator offers a systematic alternative that centers on teacher planning. When used in conjunction with intentional language feature selection and professional judgment, this approach enables teachers to generate mentor texts that are aligned to the content and also linguistically purposeful and instructionally targeted for multilingual learners. At the same time, it guides educators in prompt engineering, building confidence in using GenAI to create mentor texts and allowing teachers to devote more time to meaningful, collaborative instructional planning (dos Santos, 2025; Shafer Willner, 2025).
Additionally, the Google Sheets-based tool allows teachers to maintain a centralized, organized record of their prompts, which can be edited, reused, and adapted over time to support the creation of new mentor texts. Through this process, teachers have the opportunity to refine their use of GenAI tools while also deepening their instructional expertise by repeatedly engaging with decisions related to genre, language features, and instructional purpose (de Oliveira & dos Santos, 2025).
Decisions related to genre families and language features
As part of the prompt design process, the user must make informed decisions related to the genre family and language features relevant to the language lesson. WIDA (2020) provides that information. From an SFL perspective, the standards include not only the communicative skills students must develop, but also language patterns essential for the interpretation and expression of concepts in the content areas of English language arts, Mathematics, Science, Social Studies, and for social and academic purposes. We recommend that every ML teacher familiarize themselves with how the standards are written. That information will serve to complete the request and purpose sections of the Google Sheets-based tool presented in this article.
To access the standards and the respective Key Language Use (KLU) and language features to fill out in the Google Sheet, we recommend the WIDA Digital Explorers, an interactive online interface in the Satchel Commons platform hosted by Common Good Learning Tools (2025) and operated by the Wisconsin Center for Education Research. Upon entering the Satchel, users select the WIDA ELD Standards Framework, 2020 Edition under the WIDA Language Development Standards drop-down menu, as shown in Figure 2.
Figure 2. WIDA Digital Explorers home page
The right side of Figure 3 provides essential metadata, including citations and official source links, to support the implementation of the WIDA 2020 framework in the classroom. On the left, users will find three primary dropdown menus designed to filter and display specific data:
- Language expectations by WIDA ELD Standard statements and grade-level cluster, which organizes goals by subject area and grade levels.
- Language expectations by Key Language Use and Communication Mode, which focuses on the functional genre families to narrate, inform, explain, and argue.
- Proficiency level descriptors by grade level cluster and communication modes, which define what students can do at various stages of language development.
For prompt creation, the first dropdown item is the most practical: Language Expectations by WIDA ELD Standard Statements and Grade-Level Cluster. Figure 3 shows that step.
Figure 3. WIDA Digital Explorer: standards filter selection
In the sample demonstrated in the next section, we selected the following items shown in Figure 4 to fill out the Google Sheet cells and create a prompt for a kindergarten mentor text:
- ELD Standard 2: Language for Language Arts
- Kindergarten
- ELD-LA.K. Narrate.Expressive
- Write “narrate” in the cell under “What is the genre family (Key Language Use [KLU]).”
- Language functions and associated language features
- Select the appropriate language functions aligned with the topic and genre
- Copy and paste them into the cell under “What are the language features you want?”
Note that the language functions are first listed in bullet points. They will inform the language learning objectives for the lesson. Below that section, each bullet point is broken down into language features. The teacher may choose all the language features associated with a single language function, as shown above in Figure 4, or choose features associated with various functions, as shown below in Figure 5. Selecting the appropriate language features is crucial in prompt creation because these are the language constructions that Gen-AI tools will incorporate into the mentor text. They will support the language lesson as students see those constructions and practice reading, writing, and communicating their understanding of the disciplinary content.
Figure 4. WIDA Digital Explorers: Standards, language functions, and language features selection
Figure 5. WIDA Digital Explorer: Selecting language features of multiple language functions
Sample mentor texts generated through structured prompt design
This section provides two examples of structured prompt design, with step-by-step instructions on how to use the Google Sheets-based prompt generator. The first sample focuses on a mentor text for a 2nd-grade science lesson, and the second sample does the same for a kindergarten language arts lesson.
The first sample is a resource for a 2nd-grade ESOL teacher to teach a lesson about canyons. The language learning objective is aligned with the Explain KLU. For the first step, a prompt was created using a Google Sheets–based generator. That prompt was then entered into Gemini to produce the mentor text. Figures 6 and 7 show the process and product, respectively.
Figure 6. Sample Google Sheets-based prompt generator for a 2nd Grade Science Explain mentor text
Figure 7. Sample Gemini (Google, 2026) output from the 2nd-grade science mentor text prompt
The second sample shows the creation of a resource for a kindergarten ELA lesson aligned with the KLU for narrative genres on the topic of the first day of school. The Google Sheets-based prompt generator produced a prompt (Figure 8) that was entered into the GenAI platform ChatGPT and resulted in the mentor text shown in Figure 9.
Figure 8. Sample Google Sheets-based prompt generator for a kindergarten Language Arts Narrate mentor text
Figure 9. Sample ChatGPT 5.2 (OpenAI, 2026) output from the kindergarten mentor text prompt
Across both the kindergarten narrative and second-grade explanation examples, the mentor texts reflect the parameters specified in the Google Sheets-based prompt generator by aligning to the targeted genre, adhering to the intended instructional context, and intentionally incorporating the identified language features. In each case, these design choices support the stated instructional purposes and result in mentor texts that are developmentally appropriate, linguistically purposeful, and accessible for multilingual learners at different grade levels.
Conclusion
This article provides a systematic way for teachers to generate mentor texts that are just right for their students’ language learning outcomes. We focus on teachers’ pedagogical decisions to show that the success of GenAI in the classroom is a byproduct of human intentionality. As we have explored, prompt engineering is a disciplined instructional act (dos Santos, 2025) that requires a deep understanding of how language, context, and audience intersect.
Beyond individual pedagogical practices, this tool serves for collaborative professional development. Team leaders, Multilingual Learner Program directors, school administrators, and instructional coaches can utilize this systematic prompt-generation approach to facilitate professional learning sessions where grade-level teams collectively analyze the linguistic demands of upcoming units to co-construct prompts. By integrating this method into collaborative planning cycles, educators can move away from individual experimentation with GenAI and toward a shared, evidence-based practice leveraging AI tools effectively. When school leaders prioritize this form of structured professional learning, they take GenAI from its novelty status, which still intimidates many teachers, and turn it into a sustainable, collective asset that ensures instructional equity for multilingual learners across classrooms.
Bridging the gap between technological capability and the unique needs of MLs necessitates a structured, purposeful approach. By integrating the principles of Systemic Functional Linguistics with deliberate prompt design, educators can ensure that GenAI serves as a catalyst for meaningful meaning-making. This framework empowers teachers to make AI a reliable, high-leverage asset that directly supports the diverse linguistic trajectories of their students.
References
Common Good Learning Tools. (2025). WIDA Digital Explorers: Framework Index [Com]. Common Good Learning Tools. Wida.Satchelcommons. https://wida.satchelcommons.com/
de Oliveira, L. C., & dos Santos, A. E. (2025). Using AI-text generated mentor texts for genre-based pedagogy in second language writing. Journal of Second Language Writing, 67, 101184. https://doi.org/10.1016/j.jslw.2025.101184
dos Santos, A. E. (2025). Optimizing AI text generators for multilingual learners: The art of crafting effective prompts. GATESOL Journal, 34(1), 35–43. https://georgiatesoljournal.org/index.php/GATESOL/article/view/193
Fitzpatrick, D. (2023). The PREPare framework for AI prompting. The AI educator. https://aipioneers.org/the-prepare-framework/
Google. (2026). Gemini (Gemini 3 model, Feb 6 version) [Large language model]. Retrieved Feb. 6, 2026, from https://gemini.google.com/.
Halliday, M. A. K., & Matthiessen, C. M. I. M. (2014). An introduction to functional grammar (4th ed.). Routledge.
Liu, Y.-Y., Zheng, Z., Zhang, F., Feng, J.-C., Fu, Y.-Y., Zhai, J.-D., He, B.-S., Zhang, X., & Du, X.-Y. (2026). A comprehensive taxonomy of prompt engineering techniques for large language models. Frontiers of Computer Science, 20(3), 2003601. https://doi.org/10.1007/s11704-025-50058-z
OpenAI. (2026). ChatGPT (GPT-5.2 model, Dec 11 version) [Large language model]. Retrieved Feb. 6, 2026, from: https://chat.openai.com/chat.
Shafer Willner, L. (2025). AI-powered, integrated unit goals and lesson objectives for K–12 English learners. GATESOL Journal, 34(1), 17–34. https://georgiatesoljournal.org/index.php/GATESOL/article/view/199
Shafer Willnear, L. (2025). Aligning AI for Multilingual Learners: Integrating the WIDA ELD Standards Framework Into Prompts [Handout]. WIDA. https://wida.wisc.edu/sites/default/files/resource/video-transcripts/WIDA-Webinar-Packet-AI-Multilingual-Learners.pdf
WIDA (2020). WIDA English language development standards framework, 2020 edition: Kindergarten–grade 12. Board of Regents of the University of Wisconsin System.
