One classroom, a dozen home languages, one lesson written only in English.

Teaching English Language Learners in their own language,
while they build English skills.

English language learners and multilingual learners often sit through instruction, assessments, and assignments built entirely in English, with no way to show what they actually know. Colleague AI differentiates lesson plans for ELL and MLL students, builds diagnostic assessments in a student's home language, and gives every student an AI tutor that speaks and listens in more than 50 languages, removing the language barrier between a student and the lesson while it builds the English skills they need.

Try it now
Free to try. No credit card required.
The reality of a multilingual classroom

One lesson, written in English. A dozen home languages in the room.

A single class might include native speakers of Spanish, Mandarin, Arabic, Vietnamese, and Somali, all learning the same grade-level content, all assessed the same way, in a language most of them are still acquiring.

The lesson you're handed: English only
Entering
new to English, strong home-language literacy
Emerging
early English, home language still primary
Developing
conversational, academic English still forming
Expanding
reads grade-level text with scaffolds
Bridging
near grade level, building academic vocabulary

This is a real classroom. Five WIDA English proficiency levels in one room, and a lesson, an assessment, and a tutor that only work in English reach none of them the same way.

Differentiate, assess, and tutor, all in a student's own language

Reach every student's English level, without leaving their home language behind.

You don't build a separate class for every language or proficiency level in the room. Colleague AI builds the differentiated lesson, the home-language assessment, and the AI tutor from the same student profile.

1

Differentiate every lesson for where a student's English actually is.

Tell the AI agent a student's WIDA level or home language, and Colleague AI builds a version of the lesson with the vocabulary support, visuals, sentence frames, and pacing that student needs, without lowering the standard the lesson is teaching to.

2

Assess what a student knows, not just what they can say in English.

Colleague AI builds diagnostic assessments in a student's home language, so a language barrier doesn't get mistaken for a learning gap. A teacher sees where a student actually stands in the content, separate from where they stand in English.

3

Give every student an AI tutor that speaks their language.

The AI tutor works in more than 50 languages, with speech-to-text and text-to-speech built in. A student can ask a question out loud in Spanish, Tagalog, or Pashto and get an answer back the same way, no typing required, no language they don't have yet.

4

Support the home language while still building English.

None of this replaces English language development. It protects it. Colleague AI scaffolds in a student's home language only as far as needed, then steps the support back as English proficiency grows, so a student ends up doing the work in English, on their own.

Built for language acquisition, not translation

Language support that's grounded in how ELL and MLL students actually acquire English.

Treating a student's home language as scaffolding, not a shortcut, is a pedagogical choice about how English develops. Colleague AI is built on that distinction.

The AI tutor

Speaks and listens back, in real time.

Speech-to-text and text-to-speech mean a student who isn't ready to write in English yet can still ask questions and work through a problem out loud, in their home language, in English, or in whatever mix fits where they are.

Diagnostic assessments

Separate the language barrier from the learning gap.

A student assessed only in English can look behind in content they've actually mastered. Assessing in a student's home language first shows a teacher what's really a skill gap and what's just an English gap.

Every scaffold

Built to fade as English develops.

Support in a student's home language isn't the end goal. Colleague AI hands more of the work back in English as a student's proficiency grows, so the scaffolding shrinks while the student's English keeps growing.

50+
languages supported by the AI tutor, with speech-to-text and text-to-speech
50%
average reduction in prep time
24/7
AI tutor access, in a student's own language
$18M
in federal R&D from IES & NSF
Backed by research with the University of Washington and the AmplifyLearn.AI Research Center, led by Dr. Min Sun. Compliant with SOC 2, FERPA, COPPA, SOPIPA, and GDPR.
Every student, every language

Teach ELL and MLL students in their own language. Build their English at the same time.

Differentiate the lesson, assess in a student's home language, and hand them an AI tutor that speaks more than 50 languages, so a student's English level stops deciding what they're able to learn.

Try it now
Free to try. No credit card required.

Why Teachers Love Colleague AI

Create and Improve Assessments

Create assessments aligned to standards and generate rubrics in minutes.

start using ai for grading

Significantly Reduce Time on Grading

Use AI for grading, without removing human involvement. Improve feedback quality without sacrificing time.

Differentiate Instruction at Scale, Faster

Start Your Free Trial Now

Start using Colleague AI today. Sign up and explore our free trial.