Adopting new technology in schools can be nerve-racking, let alone artificial intelligence. We share your concern for student privacy and security, and we feel deeply obligated to good stewardship over the data we’ve been entrusted.

Our school agent is designed to be a thoughtful helper that turbocharges your existing workflows. We combined the power of AI with strict privacy controls and deep integration with your school systems. We’re helping staff work more efficiently while keeping data safe and secure.

What is the difference between a school agent and other generative AI tools?

A school agent has deep knowledge of school operations. Where a tool like ChatGPT is limited to publicly available information, our agent accesses your EHR, your student support data as well as your student information system data to formulate answers. Our school agent can also take actions on the platform on your behalf, like drafting messages to students.

What EHR data is available to the school agent?

The scope of EHR data available to the school agent is configurable. Our school agent can be instructed about conditions, allergies, medications, action plans, immunization records, encounters, screenings, form requests, medication administrations, and student documents.

What SIS data is available to the school agent?

When synced with August Schools, attendance, academics, activities, sections, and demographics are all available to the agent. Our information retrieval model is flexible and can incorporate additional sources and signals, like data or PDF exports from other backend systems.

Can I provide my own content to the school agent?

Yes. At any time, you can upload additional content for the agent, like a student handbook, mission statement, spreadsheet, or stack of report cards. The agent will reference these materials when composing answers and deciding how to react to your prompts.

Is my data used to train a model?

No. Our agent is powered by a large language model or LLM. This kind of model is trained once on massive datasets of publicly available, general content, including sources like textbooks, encyclopedia, and academic literature. These models don’t learn in realtime—rather, they deploy the sum of knowledge they acquired during their training.

How is visibility managed between practitioners?

The agent has the same visibility that the practitioner does. It considers only the students or encounters the practitioner can also access. There is no passive sharing of information between personal assistants, and they do not learn in realtime.

How is visibility controlled between roles?

We use role-based access control. The agent is restricted to those fields on the student and elsewhere that are appropriate to your role. A counselor’s personal assistant does not know about medications, conditions, or allergies.