AI and digital tools referenced across all 21 chapters - mapped to stuck points, writing stages, and CEFR levels
| Tool | Primary Function | Stuck Point | Writing Stage | CEFR Min. | Free Tier |
|---|---|---|---|---|---|
| Google Docs | Collaborative drafting, commenting, Suggesting mode peer review | Revision Structure | All stages | A2 | โ Full |
| Gemini | AI writing assistant - brainstorming, vocabulary, CLEAR prompts, Ask Don't Tell feedback | Content Language | Stages 1, 2, 5 | B1 | โ Full |
| Padlet | Collaborative brainstorming, gallery walk, reducing blank-page anxiety | Engagement Content | Stage 1 (pre-writing) | A2 | โ Limited (3 boards) |
| Grammarly | Grammar and spelling checking, vocabulary enhancement suggestions | Language | Stage 4โ5 (editing) | B1 | โ Basic |
| Google Slides | Drag-and-drop clause ordering, visual sorting activities, class presentations | Structure | Stage 2 (planning) | A2 | โ Full |
| Perplexity | AI-powered research with inline source citations; content verification | Content | Stage 1 (research) | B1 | โ Limited |
| Tool | Best Use in EFL Writing | CLEAR Compatible | L2 Restriction Support | Privacy Notes | Free Tier |
|---|---|---|---|---|---|
| Gemini | CLEAR prompts, vocabulary generation, Ask Don't Tell feedback, classroom integration with Google Workspace | โ Full | โ Responds well to B1/B2 level restrictions | Check Google Workspace for Education settings; use anonymisation codes for submitted work | โ Yes |
| ChatGPT | Essay planning, paragraph modelling, feedback generation; Custom GPTs for specialised tasks | โ Full | โ Consistent with explicit CEFR-level restrictions | Do not submit student-identified writing; use three-digit codes per Saudi PDPL | โ Limited (GPT-3.5) |
| Claude | Extended writing feedback, nuanced rubric application, metalinguistic explanation | โ Full | โ Strong adherence to level and restriction prompts | No persistent memory by default; suitable for anonymised essay submission | โ Limited |
| Copilot | Microsoft 365 integration; Bing-sourced research tasks; available in institutional Microsoft accounts | โ Partial | โ ๏ธ Variable; test restrictions before student use | Institutional accounts offer enterprise data protection | โ Via Microsoft accounts |
| Component | What to Include | Why It Matters for L2 Learners | Example |
|---|---|---|---|
| C - Context | Who you are, your level, your class, your task | Allows AI to calibrate register and complexity appropriately | "I am a B1 EFL student writing an opinion essay for my university writing class." |
| L - Level / Length / Limits | CEFR level, word count, specific constraints | Prevents AI from defaulting to C1/C2 native-speaker output | "Write at B1 vocabulary level. Maximum 150 words. Use only present simple and present perfect tenses." |
| E - Example | A sample sentence, your thesis, or a model of what you want | Anchors AI output to the student's own voice and level | "My thesis is: Social media has more negative than positive effects on students." |
| A - Action | The specific task: generate, improve, explain, ask questions | Defines what the AI should produce - prevents vague, comprehensive output | "Give me one supporting paragraph with a topic sentence and two pieces of evidence." |
| R - Role / Restriction | The AI's persona; what it must NOT do | The most critical component for L2: prevents AI from replacing student thinking | "Act as a writing coach. Do not write the paragraph for me. Ask me two questions about my evidence instead." |
| Tool | Type | Best For | Workflow Step | Key Feature | Free |
|---|---|---|---|---|---|
| Gemini / ChatGPT | AI assistant | Generating vocabulary options, synonyms, collocations at specified CEFR level | Step 1 - Generate | Can generate with B1/B2 CEFR level restriction; produces context-specific suggestions | โ |
| SKELL | Corpus tool | Collocation verification; seeing how a word is actually used in native English | Step 2 - Verify | Shows frequency-ranked collocations and authentic example sentences from a large corpus | โ |
| Ludwig | Corpus tool | Phrase-level verification; checking whether a specific phrase sounds natural | Step 2 - Verify | Searches journalistic and academic corpora; shows alternative phrasings | โ Limited |
| COCA | Academic corpus | Genre-specific verification - checking whether a word is more common in academic vs. informal writing | Step 2 - Advanced | 400M+ word Corpus of Contemporary American English; genre filter (academic, news, fiction) | โ Limited |
| Grammarly | Writing assistant | In-draft vocabulary enhancement suggestions; clarity and conciseness flags | Stage 4โ5 (in-draft) | Integrates with Google Docs; provides vocabulary alternatives in context | โ Basic |
| Stuck Point | When It Occurs | Recommended Tools | What the Tool Does | What AI Must NOT Do |
|---|---|---|---|---|
| Engagement | Before writing begins; student reluctance, blank page | Padlet, Gemini (brainstorming), topic choice menus | Lowers anxiety; generates topic options; scaffolds first thinking in either language | Generate the essay topic or thesis for the student |
| Content Knowledge | Stage 1โ2; student knows what to say in L1 but not in English | Perplexity, Gemini (content exploration), Google Scholar | Provides topic information at appropriate level; surfaces vocabulary for known concepts | Write the argument or provide evidence the student accepts without verification |
| Language | Stages 2โ3; vocabulary gaps, collocation errors, sentence errors | Gemini (CLEAR prompts), SKELL, Ludwig, Grammarly | Generates vocabulary options; verifies collocations; identifies sentence errors | Correct sentences for the student without asking the student to think first (see Ask Don't Tell) |
| Structure | Stages 2โ3; difficulty organising paragraphs or whole essay | Google Slides (clause ordering), essay outliners, Gemini (outline feedback) | Scaffolds paragraph structure; provides templates; gives feedback on organisation | Write the outline or paragraph plan in the student's place |
| Revision | Stages 4โ5; student unsure how to improve their draft | Multi-Agent Feedback System (Ch 21), Gemini (targeted revision prompts), peer review via Google Docs | Identifies specific weaknesses by dimension; produces prioritised revision targets | Rewrite the draft; AI-free Stage 4 is non-negotiable |
| Tool | Activity Type | Target Error / Skill | AI-free Stage 4? | Notes |
|---|---|---|---|---|
| Google Slides | Clause drag-and-drop ordering | Complex sentences; concessive, causal, conditional clause order | โ - original sentences written after | Teacher pre-builds slides with scrambled clause boxes; students drag to reconstruct |
| Gemini - Ask Don't Tell | AI-mediated error feedback | Any sentence-level error; hypothesis-testing about English grammar | โ - student writes revised sentence independently | Student instructs AI: "Ask me two questions about this sentence rather than correcting it." |
| Google Docs - Colour Coding | Paragraph structure diagnosis | Topic sentence identification; trench coat problem; paragraph unity | โ - diagnosis and revision entirely student-driven | Highlight tool: blue = topic sentence, green = support, gold = conclusion/link |
| Grammarly | Surface error flagging | Spelling, grammar, punctuation; vocabulary suggestions | โ ๏ธ Use with caution - student must evaluate each suggestion, not accept all | Teach students to assess whether each Grammarly suggestion improves their intended meaning |
| Padlet | Collaborative sentence gallery | Peer review of sentence variety; sharing strong topic sentences | โ | Students post their best sentence from the lesson; class votes on clearest topic sentence |
| Essay Type | Chapter | Key Tool(s) | Stuck Point Addressed | Critical AI Restriction |
|---|---|---|---|---|
| Narrative | 9 | Gemini (vocabulary bank), Majd Story Audit template | Content Structure | AI must not generate any part of the story itself - the narrative is entirely human |
| Descriptive | 10 | Gemini (sensory word bank), SKELL (collocation check) | Language | Restrict AI output to vocabulary only; no spatial organisation generated by AI |
| Expository | 11 | Perplexity (sourced research), Gemini (And-Then replacement) | Content Language | AI provides language; student provides verified content - never accept AI-generated facts without source check |
| Argumentative | 12 | Perplexity (evidence), Gemini (Khalid Counterargument Cards) | Content Structure | Student must construct the argument; AI may only help identify the counterargument to research |
| Cause and Effect | 13 | Gemini (causal connector practice), Google Slides (event ordering) | Language Structure | AI must not generate the causal analysis; use AI only to practise connector vocabulary |
| Compare & Contrast | 14 | Gemini (comparison paragraph template), Google Slides (criterion sorting) | Structure Language | AI provides the comparison sentence template only; student generates all content for both subjects |
| Definition | 15 * | Gemini (Three-Layer prompt), SKELL (genre-specific usage) | Structure | AI may check structure only; the three layers must be written entirely by the student |
| Problem-Solution | 16 | Gemini (solution generation for feasibility ranking), Perplexity (solution research) | Content Engagement | Use AI to generate solutions students then evaluate; never accept AI's feasibility judgment |
| Opinion | 17 | Gemini (Reason Ladder), arguability test prompt | Structure Content | The opinion must be genuinely the student's own; AI may only check that the thesis is arguable |
| Classification | 18 * | Padlet (Sort-Label activity), Gemini (Three-Test System check) | Structure | Classification principle must be identified by the student; AI may check exhaustive/exclusive/consistent only |
| Tool / System | Function | Chapter | Teacher vs. Student Use | Privacy Requirement |
|---|---|---|---|---|
| AI Collaboration Log | Structured student record of AI use at each writing stage; transparency and metacognition tool | 19, 21 | Student - completed alongside every draft | No personal data submitted to external AI; log remains within the institution |
| Multi-Agent Essay Feedback System | Five-agent Google Apps Script pipeline (Grammar, Vocabulary, Structure, Organisation, Synthesiser) producing dimensional feedback on submitted drafts | 21 | Teacher deploys; students submit via anonymised three-digit code | Three-digit submission codes replace student names; compliant with Saudi PDPL |
| Google Apps Script | Automation layer connecting Google Forms (submission), Gemini API (five agents), and Google Sheets (grade recording) | 21 | Teacher - setup and maintenance; students see only the output | All data stored in teacher-controlled Google Drive; no third-party access |
| Process Portfolio (Google Drive) | Folder structure collecting handwritten Draft 1 scan, revision plan, Draft 2, AI Collaboration Log, and reflection | 19 | Student collects; teacher reviews and confirms grade | Student-controlled Drive folder shared with teacher only; no AI tools access the full portfolio |
| Gemini (Four-Horizon PD Tool) | CLEAR-prompted tool evaluation for teacher professional development; planning AI integration at four time horizons (this week, this term, this year, longer term) | 20 | Teacher - professional reflection tool, not student-facing | No student data involved; teacher uses their own account |