Independent thinking. Shared around a table.Meet the IN ON network ↗
← All key insights
Roundtable recap · IN ON K-12 · 29 October 2026

AI in schools: what districts need from EdTech partners

Three US superintendents on parent trust, teacher time and the work that starts after a pilot is signed

Rick Surrency, Dr. Curtis Finch & Martin Madrid
Superintendents from Putnam County, Florida · Deer Valley, Arizona · Santa Rosa, New Mexico
TL;DR
  • Start with a learning problem or a teacher workload problem. Explain what changes in the classroom when your tool is used.
  • Help schools earn parent trust. Make the purpose, safeguards and use of student data clear before a rollout.
  • Plan for integration and staff training. A useful product still has to work alongside the systems a district already uses.
  • Research the district before approaching it. Rural schools can offer closer access to decision-makers, but need committed support.
  • A pilot needs both sides to show up after launch. Agree who supports teachers, how the tool will be used and how progress will be reviewed.

What would a superintendent need to see before saying yes to an AI product? That was the question at our IN ON K-12 EdTech roundtable on 29 October, hosted by Alex Sardinha and sponsored by Pubbly. EdTech founders and leaders joined three US superintendents to discuss how schools can use AI while protecting student data and maintaining parents' trust.

Rick Surrency of Putnam County School District in Florida, Dr. Curtis Finch of Deer Valley Unified School District in Arizona, and Martin Madrid of Santa Rosa Consolidated Schools in New Mexico brought different district experiences. All three saw opportunities in AI. Their discussion kept returning to what it would take to make those opportunities useful in a real school.

Written by INONDA from the 29 October 2026 roundtable.

1. Make the classroom benefit clear

Martin described a recent purchase of an AI-supported project-based learning tool. Its appeal was practical: help teachers create projects, offer different ways for students to demonstrate learning, and connect activities to work and careers. Presentations, videos and hands-on projects gave students more ways to engage than another worksheet or essay.

His argument was that changing the task matters alongside changing the technology. If an assignment can be completed by handing the whole thing to AI, schools need to think about what they are asking students to do. He saw scope for AI to support richer projects where students still have to apply and demonstrate their understanding.

Rick added that AI could help teachers tailor instruction to individual students. For a vendor, the useful starting point is a specific lesson, planning task or learning need. Show how the tool helps the teacher and what the student still has to learn and do.

2. Parent trust is part of implementation

Martin stressed the importance of explaining changes to parents before introducing them. A rollout should not come as a surprise. Rick described the practical work his district was considering around parental consent, reviewing vendor contracts and providing alternative lessons for students who do not use an AI activity.

The discussion showed how much an adoption decision reaches beyond a product demo. Schools need to explain why a tool is being used, what data it handles and how they have assessed it. Vendors can support that work with clear information about the product's purpose and safeguards, and a realistic account of what teachers need to do.

Screen time also came up. Curtis wanted the conversation to consider how technology is being used for teaching, alongside children's use outside school. The practical lesson is to explain the learning activity clearly enough that parents can understand its value and ask informed questions.

3. Give teachers more room for the human work

Martin was clear that relationships, care and helping children develop interpersonal skills remain central to teaching. He saw time saved through AI as an opportunity for teachers to give more attention to those parts of the job. Rick's emphasis on individual instruction pointed in the same direction: support the teacher's work with each child.

In the breakout feedback, Rick also highlighted the potential to make learning more accessible to students with additional needs. His condition was that students must still demonstrate what they have learned. Helping a child access information and assessing their understanding are both part of the task.

4. Integration and privacy belong in the first conversation

Curtis reported that his breakout group discussed protected student data, system integration and the burden of learning multiple tools. He described the value of a consistent system that staff can learn together and help one another use. A new product also needs to fit the tools the district already has.

Rick said a data privacy agreement was part of his district's contracting process. Together, these points give vendors a concrete preparation list: understand the district's existing systems, explain how student data is handled, identify the integration work and be clear about the training and support involved. These questions affect whether a school can use a product successfully.

5. Learn the district before asking for a meeting

Curtis's group discussed how EdTech companies can build relationships in the US market. He wanted to know whether another school or district in his state was already using a product. A relevant local reference can help a buyer understand how a tool works in a familiar setting.

Both Curtis and Martin suggested that rural districts could be a place to start building those relationships. Martin explained that an approach to his smaller district was more likely to reach him directly, while a larger district might have several layers between a vendor and the superintendent. That access comes with the reality that a small team wears many hats.

Martin's advice was to research the district and its leadership before making contact. Understand the setting, priorities and people. The implication for founders is to approach a school with a relevant problem to discuss and the capacity to support it through implementation.

6. The partnership is tested after the pilot starts

Martin summarised what he needs in three words: trust, purpose, partnership. Trust that the product will do what is promised and protect data. A purpose that matters to the district, whether improving learning or saving teachers time. And a partner who stays involved after the agreement is signed.

He described pilots that began with enthusiasm but stalled when support disappeared and staff did not know how to use the product. He also put responsibility on schools: if staff are not using a tool as intended, the district has work to do too. A useful partnership makes it possible to raise those problems, work through them and improve the implementation together.

Key takeaway

The thread through the discussion was practical: make AI useful for students and teachers, and help schools do the work needed to adopt it. For EdTech founders preparing their next district conversation, our takeaways are:

  1. Name the learning or workload problem you solve, and demonstrate the classroom activity that changes.
  2. Bring a clear explanation of student data use, safeguards and how the school can communicate the purpose to parents.
  3. Ask about existing systems, staff capacity and training before proposing a rollout.
  4. Agree responsibilities for the pilot, ongoing support and how both sides will review progress.
  5. Research the district and build the relationship around its students, teachers and priorities.
Continue the conversation

Scaling your EdTech: the certification edition

26 November · Digital roundtable · 5:00pm UK

Request a seat