Polaris runs listening sessions across a whole school, district, or organization and turns them into reports you can act on, with privacy protections at every step.
Here's how Polaris works, step by step. Steps marked “AI-assisted” run through our data processing pipeline. No data is trained on or sold.
1. Conversations
What makes you feel like you belong at this school?
Describe a time a teacher made you feel heard.
Choose an image that represents how your classroom feels.
When was the last time you felt proud of something you did at school?
Record a short video describing your ideal classroom.
What does a good day at school look like for you?
“School feels like a second home when teachers listen.”
Question designAI-assisted
Write questions about climate, curriculum, belonging, or anything else. Start from a template, draft with AI, or write your own.
Everyone responds
People share through group empathy interviews, one-on-one conversations, or text, in their own language. Sessions run 20 to 60 minutes.
Game-based activities
Participants choose images, sketch ideas, and respond to interactive prompts, so reflection is something they do with their hands.
2. Post-recording
I think what makes school feel good is when teachers actually remember what you said.
Yeah, like when they follow up. That means they were listening.
“Sometimes I feel like nobody would notice if I just stopped coming...”
Student #4 · Q3 · 2:14Transcription
Recordings become timestamped transcripts, with each speaker's turns kept separate.
TranslationAI-assisted
Conversations in other languages are translated into your primary language, keeping meaning and tone.
Red flag monitoringAI-assisted
Every transcript is scanned for harmful or urgent content before anyone sees it, and flagged for the right people.
Data clean-upAI-assisted
Filler words, false starts, and non-responses are removed before analysis.
3. Analysis
My teacher Mrs. Johnson at Lincoln Middle always makes time to talk.
My teacher [TEACHER] at [SCHOOL] always makes time to talk.
PII removalAI-assisted
Names, locations, and personal identifiers are automatically redacted to protect every participant.
Content analysisAI-assisted
Responses are organized by speaker turn, topic, and pattern across the full data set.
ThemesAI-assisted
Responses that say similar things are grouped into themes, across hundreds of conversations at once.
4. Reporting
Spring Climate Check-In · Grade 8 · 47 responses
78%
Belonging
4.2
Comfort
4
Themes
Student Voice and Agency in School Improvement
Mitra & Gross, 2023
Acting on student voice is linked to higher engagement.
“I feel safe when the teacher makes sure nobody gets made fun of.”
Research connectionAI-assisted
Findings are linked to published research on what has worked elsewhere.
Listening reportAI-assisted
The report lays out the main themes and recommended next steps. Every finding links to the audio it came from.
SearchAI-assisted
Ask questions of every conversation in plain language and get answers backed by audio.
5. Feedback loop
Informed by: Spring Climate Check-In
Based on the “Desire to Be Heard” theme, this follow-up explores specific moments when students felt their voice mattered.
Generated Follow-Up Questions
Tell me about a time you shared an idea and it actually changed something at school.
What stops you from speaking up when you disagree with something?
IterationAI-assisted
Act on what you heard, then write new questions that follow up on the themes that came up.
What makes you feel like you belong at this school?
Describe a time a teacher made you feel heard.
Choose an image that represents how your classroom feels.
When was the last time you felt proud of something you did at school?
Record a short video describing your ideal classroom.
What does a good day at school look like for you?
“School feels like a second home when teachers listen.”
Some of this is being built now, and some of it is still an experiment. If there's something you'd want on this list, tell us at the bottom of the page.
8 items
On the roadmap and under active work.
Translation and multilingual expansion
Full multi-language pipelines: transcribe, translate, and analyze in 50+ languages natively.
Multimedia responses
Images, video, and drawing, so participants choose how they express themselves.
Advanced question builder
Drag-and-drop question flows, branching logic, and a template library.
Additional data integration
Bring in data you already have (staff surveys, discipline records, test results) and Polaris adds it as context to the analysis.
Cross-session trends
Track how sentiment and themes change over weeks and months.
Bulk session management
Create, schedule, and manage hundreds of sessions across sites from one control panel.
Demographic lens
Cross-reference insights with demographics to see the impact on specific groups.
Advanced mixed methods
Anchor vignettes that add context to quantitative and qualitative data together.
7 items
Ideas we're exploring that may find their way into Polaris.
Family portals
Visibility into community-wide themes for families, with no individual data.
Facilitation mode
Guided suggestions and prompts for the facilitator while the conversation unfolds.
Game-based feedback
Party-style games that surface feedback about how an organization works.
Image recognition
Photos and videos from participants feed into the analysis alongside voice and text.
Vignettes
Fully anonymized, fictional day-in-the-life stories built from real themes.
Public dashboards
Shareable dashboards that let a community explore its own themes, voices, and next steps.
Design challenges
Participants work through a hands-on task and think aloud, turning project reflections into insight about their skills.
Still wondering about something? Ask us directly.
Get in touch, or tell us what you'd like Polaris to do next.