The AI engine that reads your classrooms, plans tomorrow's lessons tonight, and makes every student feel like the programme was built just for them — because it was.
CLARA is the system that closes the loop between what happens in your classrooms and what happens in them tomorrow. It reads, it thinks, it plans — overnight, automatically, for every classroom in your school.
Before CLARA, lesson planning was a teacher's Sunday evening. Class records were filed and occasionally reviewed. Progress testing happened on a fixed date regardless of what had actually been taught. Student progress sat in folders that nobody had time to synthesise.
With CLARA, every record is read every day. Every lesson plan is built from live data about that specific group of students. Every assessment tests what was actually taught. And every teacher receives personalised professional feedback before the next morning.
CLARA does not replace teachers. It removes everything that stands between a teacher and the act of teaching.
CLARA never stops. There is no test week, no review month, no scheduled report date. The loop runs every single night for every classroom in the school.
The system learns from what it reads. The longer CLARA runs in your school, the more accurately it understands your students, your teachers, and your academic rhythms.
Every lesson plan, every assessment, every piece of student feedback is adapted to the specific group — their level, their history, their pace, their needs on that day.
CLARA responds to what it reads — not to a template. Yesterday's class record directly shapes tomorrow's lesson. A student's question on the platform shapes this week's content.
CLARA acts autonomously. It does not wait to be asked. Every night, without prompting, it analyses, generates, and delivers — ready before the first teacher arrives.
Each lesson plan CLARA generates is the intersection of four distinct data sources — all produced the day before, all specific to that classroom and those students. No two lesson plans are the same.
Running nightly · All classrooms · Autonomous
The process runs automatically, in sequence, for every classroom. No human intervention required.
At the end of each day, every teacher submits their class record through the Stella platform — what was taught, how the group responded, which objectives were met, and any observations about individual students.
For every classroom, CLARA analyses the day's record against the historical pattern for that group. It identifies what landed well, what needs reinforcement, which students show signals of difficulty, and how the class is tracking against the syllabus schedule.
It then generates personalised professional feedback for the teacher — specific, constructive, and ready before the next morning.
CLARA pulls the next day's syllabus objective, cross-references it against the class record analysis, reads student platform activity from the past 24 hours, and reviews the accumulated history of that group. It then generates a complete, structured lesson plan — timed, sequenced, and annotated with personalisation notes.
Based on everything covered in that classroom over the previous two weeks, CLARA generates assessment tasks — questions, activities, and writing prompts that test exactly what those students have actually been taught. Not a fixed test. Not a generic benchmark. A live assessment built from the Record of Work.
Teachers arrive to a complete lesson plan waiting in their Stella dashboard. Assessment tasks are queued. Teacher feedback is delivered. The school has a full picture of every classroom's state — without anyone having worked through the night.
Every student receives a personal progress report from CLARA every fortnight — not a generated summary, but a genuinely useful letter that tells them exactly where they are, what they have achieved, and what to focus on next.
What the student has covered in the past two weeks, how they have performed, and how they are tracking against their original learning goals.
Specific areas where the student is performing well — drawn from class record observations, assessment data, and teacher feedback over the period.
Honest, constructive identification of where the student needs more work — with specific suggestions, not generic advice. If a student has struggled with conditionals, CLARA says so — and suggests what to do about it.
Suggested activities, resources, or focus areas for the student's own time — calibrated to their level and to what has been covered in class. Genuinely useful, not generic.
A brief preview of the academic content coming up in the next two weeks — so students can arrive to class prepared rather than surprised.
For Maria · B1 Intermediate · Week 6–7 · 20 May 2026
✓ Your use of reported speech in last Tuesday's writing task was confident and accurate — one of the strongest in the group. Your teacher noted you are beginning to use it naturally in conversation too.
📖 Recommended this week: Read one short news article and underline any passive structures you find. Try to rewrite two of them as active sentences. This takes about 10 minutes and makes a real difference.
CLARA runs on your school's own dedicated NVIDIA hardware. Every analysis, every lesson plan, every student report is generated on-site — nothing is transmitted to external AI providers.
CLARA runs on a dedicated appliance installed in your school. Student records, class data, and personal information are processed locally — never in a shared cloud environment.
CLARA uses open-weight AI models of European origin by preference — no dependency on US hyperscalers, no data flowing across jurisdictions.
All Records of Work, attendance data, and assessment records are stored in auditable formats aligned with the requirements of bodies including QQI, Equals, and the British Council.
Compliance is not a checkbox — it is engineered in. Data cannot leave your infrastructure because there is no mechanism to send it anywhere.
The appliance connects to Buongiorno for updates and remote maintenance via an encrypted WireGuard tunnel. No inbound ports. No data flows outbound.
CLARA was built and tested at Travelling Languages, an accredited English language school in Dublin — with real students, real teachers, and real GDPR obligations from day one.
Every school that adopts CLARA receives its own dedicated hardware, configured and maintained by the Buongiorno team. No student data is ever shared with any external AI provider, cloud platform, or third party. This is an engineering commitment, not a policy one — and it applies to every school, every student, every day that CLARA runs.
Get in touch to discuss a CLARA implementation at your school — or to see a live demonstration of the lesson planning engine.