.png)
Most childcare software has helped schools digitize their work. The next step is software that does some of that work for you. illumine NEXT was the launch of that idea: a panel with Tim Seldin of The Montessori Foundation and early years consultant Lara Hudson on what leaders need from AI, followed by live demos of three AI agents from illumine's co-founders, and an open discussion on data security and keeping people in the loop. This is the written recap. The full recording is at the top of this page.
Speakers
Tim Seldin, President, The Montessori Foundation
Lara Hudson, Early Years Leader and Consultant
Navneet Rastogi, Co-Founder and CEO, illumine
Sourabh Agarwal, Co-Founder and CTO, illumine
Purva Goyal, Co-Founder and CPO, illumine
Best for
Childcare, preschool and Montessori owners, directors and administrators
Session length
82 minutes
In this recap
What leaders need from AI, demos of three AI agents, how children's data is protected, and the questions answered live
Key takeaways
- The shift is from software that stores your work to AI that completes parts of it, with a person reviewing the result.
- Ease of use decides whether staff adopt new technology. Tim's test is a tired teacher writing a report at the end of the day in a language that is not her first.
- Parents are better informed and expect instant updates, and schools compete harder for every family.
- illumine's AI work is grouped into three areas: grow (enrollment), operate (data and finance) and connect (families).
- Three agents were demoed live: a billing agent that answers questions about your numbers, an enrollment agent that answers calls 24/7, and an activities agent that drafts daily updates from photos.
- AI is only as good as the data it is given and the people using it. Teachers and leaders stay responsible for what goes out.
Why illumine NEXT
Navneet opened by explaining the thinking behind the event. Software has helped schools move from paper to digital records, but the work itself still has to be done by people. The question illumine asked was whether today's technology could take on part of that work too.
What school leaders need from AI
Before the demos, Tim and Lara described the pressures they see in schools.
Change is hard
Tim said that changing technology in a school is difficult, so ease of use matters most. He described a teacher at 5:30 in the evening who has to write a report about a child, and whose first language is not English. Tools have to work for her.
Parents are more informed
Lara said parents now arrive with more knowledge and higher expectations. They want instant updates, often on messaging apps like WhatsApp, and nurseries compete hard for each family.
One platform, not many
Lara stressed the value of having enrollment, communication and operations in one place rather than spread across tools that do not talk to each other.
AI is arriving in the curriculum
Lara noted that the UAE is mandating AI in the curriculum under its 2031 framework. She also described a nature-based nursery owner who was excited, not put off, by what AI could do for her setting.
Montessori educators are cautious
Tim said many Montessori educators feel anxious about AI. The Foundation's view is a balanced one: AI is a tool, and it can be set up to work ethically within clear guardrails.
Three areas: grow, operate and connect
Navneet framed illumine's AI roadmap around three jobs every school leader has.
Grow
Respond to every inquiry, book tours and enroll more families.
Operate
See your numbers clearly, from billing and finance to compliance.
Connect
Keep families informed with less manual effort from teachers.
Demo 1: the billing agent
Navneet showed an AI agent built into illumine's billing. Leaders ask questions about their finances in plain language and get answers back as tables or charts.
- Ask in plain language, then ask follow-up questions. The agent keeps the context of the conversation.
- Turn answers into charts and pin them to your dashboard.
- Save reports with filters so you can run them again, and export to Excel or PDF.
- Role-based access means people only see the data their role allows.
Navneet said the same approach is being extended to enrollment and compliance data, such as vaccinations, emergency contacts and allergies.
See childcare billing software for how billing works in illumine today.
Demo 2: the AI enrollment agent
Purva demoed an agent that answers a school's phone and captures every inquiry, day or night.
- A parent calls
- The agent answers from the school's knowledge base
- It books a tour on the calendar
- The call is logged on the inquiry
Each call is recorded, transcribed, summarized and tagged with the caller's sentiment, and added to the inquiry's activity log. Questions the agent cannot answer are forwarded to the team.
- You choose when it answers. It can take calls only after hours, or after three or four rings when nobody picks up.
- Parents can ask for a person. At any point, the call can be handed to a member of staff.
- It switches language. In the demo it moved into Spanish when the caller did, while the transcript stayed in English for the team.
See the AI enrollment agent, or read more about how an AI enrollment agent works in childcare.
Demo 3: the AI activities agent
Sourabh explained the problem with an example from his own daughter's school. Teachers share several photos of each child every day, and Sourabh estimated that a school with 100 children can end up with around 2,000 photos a day to sort and tag. He said that work can take around three hours a day, and the agent brings it down to a few minutes.
Groups the moments
The agent recognizes children, groups related photos of the same activity together and flags what is happening in them.
Uses the teacher's notes
Teachers add short typed or voice notes, and the agent turns them into a narrative for each child.
Teacher stays in charge
Teachers can edit a draft, reject it with feedback, or ask the AI to improve it before anything is sent.
Learns your school
Over time it picks up the school's tone and philosophy, and it can write to each parent in their own language.
Sourabh added that face recognition kept recognizing his daughter as she grew. See AI daily reports, or read about AI daily reports and parent communication.
Which agent matters most
Asked which demo stood out, Lara picked a different one for each role.
For leaders
The enrollment agent, because no inquiry goes unanswered.
For teachers
The activities agent, because it gives time back to the classroom.
For owners
The finance agent, because early years is becoming more business-driven.
How children's data is protected
Sourabh walked through how illumine's AI handles school data.
- Trained on mock data. The AI features are built using mock data, not school data.
- No direct database access. Every request passes through permission checks before any data is retrieved.
- Faces stay out of the language model. Face recognition uses separate, non-LLM models. The language model is used only at the final step, with the minimum data it needs.
- Enterprise models that do not train on your data. Each school's data is kept isolated, with separate schemas by country, region and school.
- Opt-in learning only. If a school chooses, its data can be used to improve its own agent and nothing else.
- Consent is respected. Children without photo consent are not processed, and their faces can be blurred.
Read more about how illumine approaches child privacy and protection.
Keeping people in the loop
The panel agreed that AI should support staff, not replace their judgment. Tim said the teacher should always review what AI produces, and ideally a director double-checks it too. Lara said leaders still have to decide how AI is used and add their own input, and that results depend on what goes in:
AI is usually as good as the data that it's presented with.
She advised involving the whole team and setting clear policies on how AI is used. Purva added that the agents flag their confidence, so staff know which drafts need a closer look.
Questions from the session
Do you use cameras or phones to take pictures automatically?
No. Navneet said illumine does not automatically capture photos, video or voice. The approach is consent first, so schools know what is being used and how.
Which languages does the platform support?
The whole platform can be switched to another language, including Arabic, French, English and Spanish. Right-to-left languages such as Hebrew are supported too. Schools can also turn features on or off to suit their curriculum and country.
How can we share ideas for where AI should help next?
Navneet asked attendees to share ideas in the chat or contact the team directly, and said the team would follow up.
What will AI do next for directors and owners?
Lara asked for help with compliance, especially keeping documentation current, which she said is a big focus in the UAE and easy to lose track of. Navneet said compliance is what the team will pick up next. In an internal 24-hour hackathon, the team built a prototype that detects new regulations and suggests what needs to be done based on the school's own data. It has not launched yet.
Could AI support self-evaluation against the new UAE quality framework?
Lara asked whether AI could take uploaded evidence, map it to the framework and suggest a likely grade. She noted the framework is new and qualitative rather than compliance-based. Navneet said it is possible from a capability point of view, since the building blocks, such as image recognition and turning evidence into a narrative, already exist in the daily reports work, and it is worth exploring. Tim added that AI is a logical tool for compliance and could help Montessori schools prepare for accreditation, or support a voluntary, self-reported seal of excellence. Navneet cautioned that AI cannot decide what best practice is, so experts from the community need to guide it.
Can AI pick up emotion in a teacher's voice?
An attendee pointed out that voice carries a lot of emotion. Navneet agreed: when teachers give voice input for daily reports, the AI can sense whether they sound excited or worried and shape the draft accordingly. It also checks tone, for example keeping announcements formal and making sure incident reports are written sensitively.
What's coming next
Inquiry follow-up
Agents that follow up with inquiries the team may have missed and help convert them.
Newsletters
Drafts built from the daily reports and observations teachers have already written.
Lesson plans
Recommended or generated lesson plans, personalized for each child based on what they have learned and what has been observed.
Narrative reports
Being developed with Tim. Instead of a long list of lesson names parents do not understand, a coherent report on where the child is, guided by educators and the evidence on record.
What to do next
- List the tasks that eat most of your team's week, and start with the most repetitive one.
- Check how often calls go unanswered after hours or when the office is busy.
- Agree what AI can draft and what a person must always review, and write it down.
- Ask any AI vendor how children's data, photos and consent are handled.
- Involve teachers early, since they will decide whether a tool sticks.
For more on this theme, read the recap of AI and the Future of Montessori Leadership or how AI in early childhood education is changing childcare. To see the agents in your own setting, explore daycare management software or book a live demo.