Workflow · Production

Duolingo combines human expertise and AI across four course-creation stages to deliver personalized language lessons

The problem

Creating high-quality, personalized language courses at scale requires both creative human judgment and AI-driven personalization — neither alone is sufficient.

Workflow diagram · grounded in source
1
Curriculum design by experts
human_review
“experienced curriculum designers carefully plan what to teach and when for any given course. They design the high-level structure of the course, specifying the order of learning objectives”
2
AI translation generation
ai_action
“we use AI to help create a range of possible translations of all the sentences so that we can later accept learners' responses in cases when there are multiple correct ways of saying the same thing”
3
Automated exercise generation
ai_action
“most of the time we use computer algorithms to automatically create exercises from "raw" content”
4
AI grading and audio generation
ai_action
“we may ask learners to say the sentence out loud and use AI to figure out whether they said it correctly. AI also allows us to generate the audio for exercises”
5
Human-authored comprehension questions
human_review
“our content developers write those questions themselves to make sure we ask about something that aligns with the lesson's learning objective”
6
Birdbrain lesson personalization
ai_action
“we use our Birdbrain model to figure out which exercises in a particular lesson are going to be the best match for a learner's level of knowledge”
Reported outcome

The result of all four course creation stages is a personalized Duolingo lesson that teaches carefully chosen concepts, uses rich language content, converts it into interactive exercises, and adapts to the learner's level.

Reported metrics
Content development speed and accuracywork faster and with fewer mistakes
Reported stack
BirdbrainDuolingo World character voices
Source
https://blog.duolingo.com/how-duolingo-experts-work-with-ai/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The result of all four course creation stages is a personalized Duolingo lesson that teaches carefully chosen concepts, uses rich language content, converts it into interactive exercises, and adapts to the learner's l…

What tools did this team use?

Birdbrain, Duolingo World character voices.

What results were reported?

Content development speed and accuracy: work faster and with fewer mistakes (source-reported, not independently verified).

How is this workflow AI workflow structured?

Curriculum design by experts → AI translation generation → Automated exercise generation → AI grading and audio generation → Human-authored comprehension questions → Birdbrain lesson personalization.