Treat every follow-up as the next turn, not a new speech: answer the new question, link to one useful detail from what you already said, then add one fresh reason or example.
The scary part isn’t speaking for 35 seconds. It’s hearing question #4 and realizing question #4 remembers question #2. 😅
The whole method in three moves
- Answer: respond to the new question immediately.
- Link: reuse one relevant detail from your previous answer.
- Extend: add one new reason, example, consequence, contrast, or limitation.
Call it Answer → Link → Extend. If you remember only one thing from this page, remember that.
What DET Interactive Speaking actually does
Duolingo’s Interactive Speaking task is a simulated conversation rather than a row of unrelated monologues. According to the official Duolingo English Test explanation, a conversation contains 6–8 turns, you get 35 seconds to respond to each prompt, and there is no preparation time. The system can select personalized follow-up questions based on what you have already said.
That last part is the plot twist. Your previous answer is not dead. It is now part of the conversation.
Check the current format on the official DET Interactive Speaking page and Duolingo’s official Interactive Speaking introduction.
Duolingo describes the task as looking at speaking qualities including fluency, grammar, pronunciation, vocabulary, coherence, and task completion. Your practical job is simpler: answer the actual question, stay connected, and keep your English understandable.
Why decent speakers suddenly blank
Imagine your first question is about where you like to study. You say:
I usually study in the library because it’s quiet, and seeing other students work helps me concentrate.
Then the next question asks:
What specifically makes that environment productive for you?
A common reaction is to mentally press the giant red RESTART button:
I like studying in the library because the library is quiet and I can study there...
That sentence is grammatically valid. The problem is strategic: you are spending your new turn rebuilding information the conversation already has.
A stronger continuation is:
The biggest thing is the lack of interruptions. Like I mentioned, I focus better when other people are working around me, and in the library I’m much less likely to check my phone every two minutes.
That answer moves. It answers the follow-up, links back to one earlier detail, then gives the conversation somewhere new to go.
Answer → Link → Extend in real time
Answer the new question first
If the prompt asks why, start with a reason. If it asks for a disadvantage, give the disadvantage. If it asks what changed, name the change. Do not spend your opening sentence driving around the neighborhood looking for the question.
Link one earlier detail
You do not need a summary of your previous answer. One detail is enough. Bridges such as “Like I mentioned…” or “That connects to…” can help, but do not force them into every turn.
Extend with one new layer
Add a concrete example, personal consequence, contrast, downside, or condition.
Old detail: I like working in teams.
Follow-up: When can teamwork become difficult?
Extension: It becomes difficult when responsibilities are vague, because two people may assume someone else is handling the same task.
One conversation, three turns
Turn 1
Question: What kind of student club would you join?
Answer: I’d probably join a photography club because I like exploring a city with a purpose. Taking photos makes me notice details I would normally walk past.
Turn 2
Follow-up: Why would doing that with other students be better than doing it alone?
Answer → Link → Extend: The social part would push me to go out more regularly. Since I already enjoy noticing small details, comparing photos with other students would also show me what I completely missed.
Turn 3
Follow-up: Could being in a group ever make the activity less enjoyable?
Answer → Link → Extend: Definitely. If everyone wanted to stop at different places, it could become slow and frustrating. I’d probably prefer a small group so we still get the social benefit without turning one photo walk into a committee meeting.
Adaptive Follow-Up Lab
Speak your answer aloud before opening each reveal. Your goal is not to copy the model wording. Check whether you answered, linked, and extended.
Scenario A: study habits
Your previous answer: You said you review difficult material early in the morning because your mind feels clearer.
Follow-up: What makes that routine difficult to maintain?
Check a strong continuation
Answer: The hardest part is going to bed early enough. Link: Morning study only works for me when I’m actually alert. Extend: If I sleep late, I usually move the difficult work to the afternoon instead of forcing a useless 6 a.m. session.
Scenario B: technology
Your previous answer: You said online maps make travel less stressful because you can change plans quickly.
Follow-up: Can relying on that technology create any problems?
Check a strong continuation
Answer: Yes—people can stop paying attention to their surroundings. Link: I like maps because they let me change plans quickly, but extend: if I follow the screen blindly, I may miss signs, landmarks, or a much nicer route right in front of me.
Scenario C: teamwork
Your previous answer: You said group projects can be useful because people bring different skills.
Follow-up: How would you deal with a teammate who was not contributing?
Check a strong continuation
Answer: I’d talk to them privately first instead of accusing them in front of the group. Link: Different skills only help when everyone has a clear role. Extend: I’d ask whether the problem was workload, confusion, or motivation and then agree on one specific task and deadline.
Four mistakes that make adaptive follow-ups harder
Restarting the whole topic
The conversation already has context. Rebuild only what the new answer needs.
Repeating because repetition feels safe
Reusing an earlier point can create coherence, but it needs a new layer. Think old detail + new consequence, not old paragraph again.
Forcing a giant template
A rigid “Firstly, secondly, in conclusion” structure can sound strange when the prompt is simply asking you to clarify one thing you just said. A conversation is not a five-paragraph essay wearing headphones.
Manufacturing a grand conclusion
You do not need to finish every turn like you are ending a TED Talk. Once you have answered and developed the point clearly, finish naturally.
A practical 35-second training rhythm
The official task gives you 35 seconds to record each response. For practice, use a flexible rhythm: answer directly, connect one relevant detail, develop with an example/reason/contrast, and finish when the idea is complete. This is a training scaffold, not an official scoring formula.
Micro-challenge: carry one detail
Answer: What is one habit that helps you learn effectively? Speak for about 20–30 seconds, then write down one noun from your answer. Invent a follow-up and answer it while using that noun exactly once and adding one new idea.
Want more spontaneous speaking reps?
Once the manual method makes sense, you can choose a speaking-practice path in FunFluen for more English output practice. It is general speaking practice—the exact DET lesson on this page is not preloaded there.
For broader learning workflows, browse FunFluen’s media-based language-learning guides.
Quick questions
Do I need to fill all 35 seconds?
The official task page gives you 35 seconds to record each answer. Treat that as available response time, not a reason to pad an idea after you have answered the question clearly.
Are the follow-ups really based on what I say?
Yes. Duolingo states that Interactive Speaking can select personalized follow-up questions based on previous responses, which is why continuity matters.
Should I memorize an answer template?
A rigid full-answer template is a poor fit for an adaptive conversation. A lightweight thinking pattern—Answer, Link, Extend—is more flexible because the content still has to respond to the actual question.
Don’t prepare six speeches. Prepare to continue.
Interactive Speaking feels unpredictable because the next turn can react to the conversation you are creating. That is exactly why the safest strategy is not to predict every question.
Answer what was asked. Carry one useful detail forward. Add one new layer.
Question #4 may remember question #2. Good. Now you know what to do when it does.