Students log into Amira on a computer with a headset. Students are given a choice of stories to read; as they read the story Amira, uses speech recognition technology to assess student reading gaps and needs. Amira responds in the moment of struggle by providing tutoring micro-interventions. At the conclusion of each reading session, Amira recalibrates so the next time the student logs in, they are presented with a choice of updated stories that are the appropriate level of challenge for the student.

The AI uses speech recognition to process a student's reading in real time, identifying errors down to the individual phoneme as the student reads. When it detects a specific error, it responds the way a tutor would, choosing from a range of coaching moves, like giving a hint, modeling the sound, or prompting the student to try again, to fit that specific error. These responses come from a closed system built on speech recognition, not an open-ended generative AI model, so the AI draws from a fixed repertoire of coaching moves rather than generating replies on its own. Over the first few sessions, it offers strategies based on aggregate information about what tends to work for students at similar reading levels, and over time it adapts to individual students as it learns which particular strategies are most effective for them. It maps each error to specific reading skills and state standards, listening for several thousand discrete skills, and updates the teacher's reports to the phoneme level after every session. It administers and scores Amira's reading assessments, including tasks like oral reading fluency and word identification.

Teachers decide how to build Amira into the class routine, whether as a station that students rotate through during small-group instruction or as independent work. They work from the reports Amira produces after each session to decide what to reteach, how to group students for small-group instruction, and which resources to pull. Teachers can play back a student's recording after a session to hear exactly where they struggled. Teachers have the ability to override and correct Amira’s score. A teacher can assign specific assessments or subtests to a student, can assign specific micro-lessons and skill-targeted tutoring sessions to reinforce areas of need, and can pull individual reports to share with students so they can track their own progress. Teachers can also draw on Amira's screening data, including its indicators of dyslexia risk, to inform next steps.

Leaders adopt and set the scope of use across a school or school system. Because Amira screens students' reading and is an approved screener in a number of states, leaders can use it as their assessment for early reading and dyslexia risk, giving them a shared source of reading data across classrooms. Leaders can use that data to see how reading is progressing across classrooms and grades, identify where students are falling behind, and target support and coaching accordingly. Leaders can also connect Amira to the school system’s core reading curriculum so its assessment, instruction, and tutoring align to the school system’s scope and sequence.