Why Pearl Started with an AI Admin Assistant for Education Leaders
By: Chris Kerr
Vice President, Product, Pearl
Every technology company right now is grappling with the same question: how do we harness the power of AI in a way that actually adds value?
It’s a genuinely compelling challenge. We’re watching an entire industry figure out how to engage with something that feels meaningfully different from past waves of innovation. Some companies’ implementations have been celebrated. Others have been met with skepticism. Many fall somewhere in between — well-intentioned, but ultimately checkbox exercises. Still, everyone is experimenting, and for good reason. This isn’t blockchain or NFTs. It isn’t a speculative technology looking for a problem to solve.
AI has the potential to be deeply impactful across nearly every domain and application. The real question isn’t whether to integrate AI — it’s how to do so thoughtfully, especially in education, where trust, evidence, and human judgment matter.
At Pearl, our answer starts with a very deliberate first step: the AI Admin Assistant.
Pearl’s First AI Move: A PhD-Level AI System for Administrators
Pearl’s AI Admin Assistant is a PhD-level AI assistant built directly into the Pearl platform and designed specifically for education administrators to help them explore and understand their data. Its purpose is simple but powerful: reduce insight latency, ground decisions in evidence, and multiply human impact.
Our initial AI implementation is a conversational tool embedded right where administrators already work. Through a natural chat interface, admins can ask questions about their tutoring and student support programs — enrollment trends, session completion rates, student progress — all within the context they define. For more complex questions that require deeper analysis, Pearl AI guides admins to the relevant insights pages and generates tailored visualizations on the spot.
This isn’t a bolted-on chatbot or a generic AI assistant.. It’s an AI analytics tool woven directly into the platform.
Why Start with Administrators and Tutoring Data?
So why start here? Why administrators? Why data exploration?
Administrators have an outsized impact on the organizations and systems they lead. The decisions they make and the clarity they have when making them shape how programs operate day to day. When administrators have better insight, that understanding carries through to coordinators, tutors, and ultimately students.
An organization doesn’t become “AI-powered” because of a tool alone; it does so when its leaders can see more clearly and act with confidence.
Building the Right Foundation for AI First
From a technical foundation standpoint, this is also the right place to start. Future capabilities like AI-supported tutoring coaches, intelligent product guidance, or automated knowledge bases all depend on the same underlying data foundation. Building a strong administrative data layer first gives us the building blocks to move forward deliberately.
Another important factor is our partners. At Pearl, we’re fortunate to work with districts and organizations that have deeply engaged administrative teams. Nearly everyone we speak with is eager to be part of shaping the tools they rely on. That collaborative approach has been key to Pearl’s growth so far, and we see it as essential going forward.
While feedback from tutors, parents, and students is incredibly valuable, the reality is that engaging those groups at scale is different. With administrators, we have real, frequent conversations with the people who will use this tool every day. That proximity helps us build something grounded in real workflows and real needs.
Human-in-the-Loop by Design
This approach also reflects how we think about AI more broadly. Our goal is to build human-in-the-loop products that put people and their goals first and support judgment rather than replace it.
I like to think of it as an Iron Man suit: the suit makes Tony Stark incredibly capable, but the suit alone isn’t the story – it’s Iron Man inside. Our AI tools are designed to extend human capacity, not replace human judgment.
The administrator remains in control, asking questions, interpreting results, and ultimately making decisions. The AI simply reduces friction and expands what’s possible in the time they have.
Internally, we talk about this challenge as insight latency — the gap between when tutoring happens and when leaders can actually act on what they’re seeing. Historically, that gap stretches months. Our goal with the AI Admin Assistant is to compress that cycle dramatically, helping administrators move from action to understanding in near real time.
That’s also why our AI is grounded in the existing evidence base for high-impact tutoring, with citations and source material available so administrators can understand not just the answer, but the reasoning behind it.
Conclusion: Why This Matters
We’re already beginning to see the impact this can have on the organizations we serve, and the momentum is energizing. Not because of the technology itself, but because of what it enables people to do. At the end of the day, behind every system and every dataset are real humans trying to do important work. Those are the people we’re here to support.