Edcafe AI Chatbot for Reflective Writing: A Drafting Scaffold
Published on 23 March 2026
by Ashlee Tai, Academic Literacies Centre, Global Cultures and Languages Hub, Xi’an Jiaotong-Liverpool University
Edcafe is an education-focused AI tool that helps teachers efficiently produce learning resources and teaching materials (Image 1). Its AI Chatbot enables teachers to build a customised chatbot tailored to specific teaching goals through a guided setup, requiring no technical skills or complex prompt engineering. This makes it ideal for embedding in classroom activities to support student learning. Key features of the Edcafe chatbot include:
- Instructional control: Teachers can set learning goals, proficiency levels, scope, tone, and more. They can also upload course materials, such as workbooks, slides, and rubrics, to the chatbot’s knowledge base, enabling it to generate class-relevant responses. Unlike a generic LLM chatbot that gives variable answers based on students’ prompts, Edcafe’s chatbot provides more targeted and relevant support.
- Fast setup: Edcafe automatically generates prompts based on teachers’ settings, reducing the need for trial-and-error prompt writing.
- Teacher visibility: Edcafe offers session summaries, chat history reports, and engagement data, helping teachers review how students interact with the chatbot and monitor the learning process for potential issues. In contrast, many typical chatbots do not provide teachers with access to students’ interactions.
- Guided practice: The chatbot scaffolds the learning process by providing students with specific prompts and feedback, rather than defaulting to generic responses.
In one of XJTLU’s Year 2 science modules, students at the CEFR B2–C1 level are required to write a reflective essay as part of their assessment. Since most students are unfamiliar with reflective writing, a common challenge is their tendency to merely describe events without progressing to authentic reflection, which involves analysing what they have learned, why it matters, and what they will do differently next time. Additionally, students often struggle to structure their ideas using Rolfe’s1 What? / So what? / Now what? framework, and to use appropriate reflective language.
To address these challenges, particularly in large classes where individual support is limited, an AI chatbot has been developed to assist students during the drafting process. Integrated into class activities as a teaching assistant, the chatbot augments teacher guidance by providing timely learning support. While drafting, students can interact with the chatbot to clarify how to apply the framework, receive prompts that encourage critical reflection on their experiences, learning outcomes, and future actions. They can also obtain language support to help them revise their work.
1. Rolfe, G., Freshwater, D., & Jasper, M. (2001). Critical reflection for nursing and the helping professions: A user’s guide. Palgrave Macmillan.
Image 1 edcafe
Set-ups/Technology required
You can register for a free account on Edcafe. After creating your chatbot, you can share the access link or QR code with students for in-class use (see Image 2). Students can then access the chatbot on their own devices, such as mobile phones, tablets, or laptops, to practise reflective writing.
Procedure
Steps to create the Edcafe AI chatbot:
Step 1:
- Click ‘Create new’ (Image 3) and select ‘Chatbot’ (Image 4).
Step 2:
Write your teaching prompt in the field labelled “What do you want the chatbot to do?” (an example is provided below), then click “Create Chatbot” (Image 5).
Example prompt: “to help students develop reflective writing skills using Rolfe’s framework (What, so what, now what), including how to reflect with guiding questions, how to structure reflective thoughts, and how to write reflectively with correct tenses and vocabulary.”
This prompt enables the chatbot to scaffold the writing process by guiding students to reflect on experiences, organise ideas using Rolfe’s framework, and improve language accuracy and academic reflective vocabulary.
Step 3:
Edcafe will auto-generate ‘Chatbot instructions’ (Image 6). You can refine these using the ‘Prompt Assistant’ (Image 7). Additionally, upload relevant materials (e.g., task sheets, rubrics, worked examples) to the ‘Knowledge’ section to ensure the chatbot’s responses align with your lesson objectives (see Image 8).
Step 4:
Test the chatbot via ‘Test me out’ (see Images 9-10). You can role-play as a student to check that the chatbot provides appropriate scaffolding (Images 11-12). After publishing, you can monitor students’ interactions and track their learning progress (Image 13).
Image 2 Chatbot QR code
Image 3 Create a new tool
Image 4 Create a chatbot
Image 5 Describe your chatbot
Image 6 Chatbot instructions
Image 7 Refine chatbot instructions with your comments
Image 8 Upload knowledge
Image 9 Chatbot Test
Image 10 Chatbot Test_User view
Image 11 Preview
Image 12 Preview_response example
Image 13 View chats
Procedure
Teaching steps:
Task:
After the teacher’s instruction on reflective writing and Rolfe’s framework (What?/So what?/Now what?), students will write a short reflective piece (150-250 words) on the topic:
Reflect on an AI tool you used to help you complete a study task in your science module.
Step 1: Teacher instruction
Provide instructions on reflective writing for 20 to 30 minutes, including the definition, an introduction to Rolfe’s framework with worked examples, guidance on appropriate tense usage, and useful reflective language
Step 2: Student writing preparation with the chatbot
Share the chatbot link or QR code with students and guide them to use it to review the framework, respond to guiding questions, and develop a brief writing plan for 10-20 minutes. During this time, you can monitor students’ interactions with the chatbot through the pathway shown in (Image 13) and provide instant guidance as students need it.
Step 3: Independent writing
Students write their short reflective essay on the given topic using their devices for 30 minutes during class.
Step 4: Peer review
Students work in pairs to review each other’s writing using a rubric aligned with the assessment criteria for 10-15 minutes. Teacher’s general or individual feedback can be provided in the following class.
Optional extension
Provide instruction
Caveats and Alternatives
On the Edcafe free plan, teachers can create up to 3 custom chatbots and host up to 40 participants per chatbot using standard LLM models, with chatbot history reports included. For larger cohorts or access to advanced LLM models, teachers may consider updating their plan, rotating student groups, or reusing the same chatbot across classes.
Additional Information
Here is the reflective writing AI chatbot link:
https://app.edcafe.ai/chatbots/68e262bb20a4a8c1d8019eb7
Editor: Kun Li ; Sub-editors: Shuhan Li, Liu Cao (AI Collection); Copy editors: Helen Beech, Kathryn Paterson, Garth Elzerman; Graphic designer: Chenyan Zhang
Published by Xi’an-Jiaotong Liverpool University. Ready to TELL online magazine Copyright © 2022 XJTLU