Do's and Don'ts for Teaching Intermediate English Speaking
Discover the essential do's and don'ts practice intermediate english speaking for teachers. Learn how to maximize student talk time and correct errors.
ESL/EFL educators often meet the same B1/B2 pattern: students know useful vocabulary, yet their spontaneous speech still stops, restarts, and depends heavily on teacher support. Teaching intermediate speaking matters because outdated drills can let fossilized errors settle while students miss chances to build active speaking skills. This guide gives a practical do's and don'ts framework for in-class facilitation and autonomous practice.
Key Takeaways
Below are the key takeaways in this guide:
- Student talk time: Use open-ended questions so students produce longer answers, not one-word replies.
- Error correction: Protect conversational flow first, then correct patterns after the task.
- Contextual input: Teach words inside useful situations, not as isolated lists.
- Loora for homework: Loora gives proactive real-time pronunciation and grammar feedback during spoken conversations, so students don't need complex prompts.
Do's and Don'ts for Maximizing Student Talk Time
For intermediate speaking practice, Student Talk Time should clearly outweigh Teacher Talk Time. Your role is to set the task, model the first move, and redirect when needed, while students carry the conversation for most of the activity.
In class, I once watched a student freeze during a role-play about a delayed project update. Instead of rescuing him with a full answer, I used the Follow-Up Rule and asked, "What happened after the deadline changed?" That small prompt kept the turn alive, and the student added two more details without losing conversational confidence.
The shift is simple but powerful. Less teacher explanation creates more student production, and more production gives you better evidence for feedback.
Classroom Facilitation Checklist
| Do | Don't | Why it matters |
|---|---|---|
| Use open-ended questions like "What would you suggest to the client?" | Use Yes/No questions that end the turn too fast. | Intermediate learners need space to build ideas across several sentences. Short answers limit real speaking practice. |
| Apply the Follow-Up Rule after each first answer. | Move to the next student as soon as one answer appears. | Follow-up questions train students to extend, clarify, and react. That mirrors real ESL speaking activities. |
| Use thematic clustering for role-plays, such as project updates, client calls, or team decisions. | Teach vocabulary in isolation before asking students to speak. | Words become easier to use when students meet them inside a clear situation. Context supports recall. |
| Give students planning time before spontaneous speech. | Dominate the setup with long Teacher Talk Time. | A short planning stage helps students enter the task with ideas, not scripts. The speaking still stays active. |
| Let pairs run the task before whole-class feedback. | Correct every pair while they are still speaking. | Peer-to-peer interaction increases Student Talk Time and gives quieter students a safer first round. |
Good facilitation does not mean stepping away from the lesson. It means designing the talk so students do the cognitive work, then using your teacher input at the moments where it has the most value.
Do's and Don'ts for Intermediate Error Correction
Accuracy means students choose correct grammar, words, and pronunciation. Fluency means they keep the message moving. At the intermediate plateau, over-correcting every small mistake can break conversational flow, so feedback should protect comprehensibility first and refine grammar accuracy after the speaking task.
In one B1/B2 lesson, I delayed correction during a team-decision role-play and took notes instead. Students spoke for longer turns, negotiated meaning, and stayed inside the task. Afterward, we reviewed three repeated patterns together, which made the feedback clearer and less personal.
Intermediate Error Correction Do's and Don'ts
| Do | Don't | Why it works |
|---|---|---|
| Delay feedback until the end of the speaking task. | Interrupt a student mid-sentence for every grammar slip. | Delayed feedback keeps the conversation alive. It also gives you better evidence of repeated errors. |
| Focus first on comprehensibility. | Treat perfect pronunciation as more important than clear communication. | In communicative language teaching, the listener must understand the message. Minor sound issues can wait if the meaning stays clear. |
| Explain the reason behind the correction. | Give the correct form with no model or answer key. | Students need to know why a form works. A clear model helps prevent fossilization. |
| Correct patterns, not every single mistake. | Turn feedback into a long list of errors. | Pattern-based feedback helps students notice what affects their speaking most. It also protects conversational confidence. |
Error correction has the most value when students can reuse it. A short correct model, followed by one quick transformation task, often works better than a long explanation.
Don't Rely on General Chatbots for Homework and Do Recommend Purpose-Built Artificial Intelligence Tutors
Teachers often suggest AI language tools for autonomous practice, but general chatbots usually need careful prompts before they act like a speaking coach. For intermediate learners, a purpose-built AI English tutor can reduce that setup work and keep homework focused on spoken output.
Artificial Intelligence Tools Comparison: ChatGPT vs Claude vs Loora
| Practice area | ChatGPT | Claude | Loora |
|---|---|---|---|
| Speaking focus | ChatGPT works well for written prompts, grammar explanations, and open-ended text chat. Voice practice depends on how the student sets up the task. | Claude is strong for nuanced text explanations and complex language questions. It is not built mainly as a live English speaking tutor. | Loora focuses on spoken English practice. Students talk with the AI and receive feedback during conversations. |
| Scenario realism | ChatGPT can simulate a project update or client call if the student writes a clear prompt. Without that prompt, the practice can stay general. | Claude can create detailed role-play scripts and explain better responses. The student still has to guide the scenario step by step. | Loora includes professional role-plays for meetings, client calls, presentations, handovers, and critiques. Built-in starter scripts adapt to the student's industry and level. |
| Role-play flexibility | ChatGPT can role-play many situations through text. The quality depends on the student's prompt and follow-up instructions. | Claude can build careful role-play scenes and refine them in text. Students need to keep directing the activity. | Loora lets students create custom scenarios for specific work contexts. A student can rehearse a presentation, a job interview, or a difficult status update. |
| Real back-and-forth talk | ChatGPT supports back-and-forth chat and can ask follow-up questions when prompted. The exchange is mainly text-led. | Claude handles long text exchanges with strong reasoning. It can respond thoughtfully, but the speaking structure is not built in. | Loora runs interactive, multi-turn spoken role-plays. The AI responds dynamically, asks follow-up questions, and gives realistic pushback. |
| Feedback depth | ChatGPT can explain grammar and suggest better wording when asked. It does not provide built-in real-time pronunciation correction from the Loora feature set. | Claude gives detailed grammar breakdowns and careful text feedback. Students must request and organize the feedback themselves. | Loora highlights pronunciation and grammar mistakes during the conversation. Afterward, it gives post-conversation review, native-speaker examples, and vocabulary takeaways. |
| Conversation pressure | ChatGPT can create pressure through text prompts, such as a difficult client question. Students still read and type, so the pressure differs from live speech. | Claude can write challenging follow-up questions and stakeholder replies. The student controls the pace through text. | Loora asks students to answer verbally in the moment. That supports practice for thinking while speaking. |
| Context fit | ChatGPT can adjust to a student's context if the prompt includes enough detail. Without detail, replies may stay broad. | Claude can adapt text to a detailed brief. The student still needs to supply that brief clearly. | Loora uses starter scripts that adapt to the student's industry and level. Students can also set their own custom scenario. |
| Best real-world outcome | ChatGPT can prepare students to draft answers and explore language choices. It works best when they know how to prompt it. | Claude can help students analyze wording and improve prepared responses. It is useful for planning before speaking. | Loora prepares students to rehearse spoken workplace tasks, such as a project update, client call, or presentation. It gives real-time feedback without requiring complex prompt design. |
Conclusion
Learning how to teach intermediate English speaking means moving students from passive knowledge into active, judgment-free speaking practice. Strong language acquisition at this stage comes from combining classroom facilitation, delayed feedback, and structured homework that gives students repeated chances for spontaneous speech.
When teachers pair open-ended ESL speaking activities with an AI English tutor like Loora, students get more chances to speak, review feedback, and return to class ready for deeper conversation. That connection between in-class interaction and autonomous practice is where the intermediate plateau starts to shift into stronger conversational control.