Renaissance uses generative AI on AWS to accelerate personalized learning at scale

As a global leader in K12 educational technology (EdTech) solutions, Renaissance supports over 16 million students across more than 100 countries with a mission to improve learning outcomes for all students, regardless of ability or background. Its data-rich platform helps educators make informed decisions and deliver personalized learning experiences for students.

This personalization depends on a complex process to match every piece of content—whether a math problem or reading passage—to the right learning objectives. These objectives are based on educational standards set by states or national frameworks, and they shape how instruction is delivered in schools and districts.

With hundreds of thousands of content items to map and education standards frequently evolving across states and countries, this content alignment work became time-consuming to scale. To keep pace without sacrificing quality or expert oversight, Renaissance worked with Amazon Web Services (AWS) to develop a scalable, human-in-the-loop solution powered by generative artificial intelligence (AI).

A connected ecosystem for data-driven learning

Founded in 1986, Renaissance has grown from a reading practice program into a comprehensive learning ecosystem built on core pillars: assessment, practice, instruction, and insights. These pillars are tightly integrated, enabling the platform to deliver personalized practice recommendations for students, instructional guidance for teachers, and data-driven insights for administrators.

“Because we have the connection between instruction, practice, and assessment, we can meet students where they’re at in their learning journey,” said Julianne Robar, senior director, metadata and product interoperability at Renaissance.

Powering this personalized recommendation system is a rigorous content alignment process. Every piece of content—from test questions to instructional resources—must be mapped to the appropriate learning objective, which is typically derived from state academic standards. These standards set the expectations for what students should learn at each grade level and subject, and they influence curriculum decisions at the district and school level. The alignments between content items and learning objectives are essential because they determine the accuracy of the platform’s recommended guidance for each student’s learning level.

Renaissance’s in-house psychometric and pedagogy experts have long upheld a rigorous content alignment strategy, but maintaining it manually is time-consuming and increasingly complex as standards evolve and the business scales. With a dozen states having recently adopted or working on new standards in 2025 alone, and international frameworks to consider as well, the team needs a scalable way to reduce manual effort without compromising accuracy.

Collaborating with the AWS Generative AI Innovation Center

As generative AI technology matured, Renaissance began exploring how it could support this work. Renaissance’s AWS account team led them to the AWS Generative AI Innovation Center, an initiative that connects AWS science and strategy experts with organizations to help prioritize generative AI use cases, build a roadmap, and move solutions into production.

From the outset, Renaissance made it clear they weren’t looking for another proof of concept. “Getting something more mature was important for us,” said Robert Zieroth, principal architect, AI product strategy and innovation at Renaissance. “That was one of the selling points. AWS was willing to go beyond the POC and knew how to do it.”

Together, Renaissance and AWS AI strategists regularly convened to develop a working solution. “AWS was quick to volunteer new ideas and approaches to test,” Zieroth added. The project also underwent review by Renaissance’s internal Accelerate Innovation by Design (AID) committee, which vets AI initiatives against a rigorous rubric for responsible AI development.

One of the project’s most important decisions was involving its end users—the content alignment experts—throughout the process. “What I really appreciate about Renaissance,” said Nico Tyjeski, a generative AI scientist at AWS who supported the project, “is that they asked the people doing the work what problems AI could solve, and included them in selecting the right data, reviewing outcomes, and refining the model.” While involving frontline users may seem obvious, it’s still uncommon in many AI implementations—making Renaissance’s work a standout example of human-in-the-loop design.

Building a scalable, human-in-the-loop solution on AWS

Over a six-month period, Renaissance worked with AWS to develop a generative AI solution designed to streamline content alignment while preserving pedagogical expert oversight. Built on Renaissance’s existing item-to-objective matching framework, the solution mirrors their established alignment standards and accelerates one of the most time-consuming parts of the content mapping process.

At a high-level, the solution works as follows:

• A user submits a new content resource, such as a math problem or multiple-choice question.
• The generative AI solution analyzes the item and returns a ranked list of learning objectives the item is most likely aligned to, along with a confidence score to help guide expert review.
• An alignment expert reviews the suggestions and makes a final decision.
• The approved results route back into the model to improve future predictions.

“The AI is not making a final determination,” said Robar. “It still goes back to humans for review because there’s a lot of nuances in making those determinations.”

The architecture supporting the solution draws on a range of AWS services, including Amazon Bedrock for its generative AI models, Amazon OpenSearch Service for information retrieval, and AWS Lambda for serverless processing. Other integrated services include Amazon Elastic Container Service (Amazon ECS), Amazon Simple Storage Service (Amazon S3), Amazon DynamoDB, and Amazon API Gateway.

Looking ahead: Faster personalized learning at scale with generative AI

The solution is currently undergoing internal scaling and is expected to go live in the coming months, but feedback from its early content alignment users has been positive. “Our alignment team sees the potential,” Robar said. “They know it can save time, and that time can be reinvested in more complex content design and instructional strategy.”

By accelerating the alignment process, Renaissance can make sure that newly developed content is quickly matched to the right learning objectives, so the platform can deliver relevant, standards-aligned material to educators and students more efficiently. For instructors, that means faster access to classroom-ready assessments; for students, it means a more responsive, personalized learning experience powered by timely, accurate recommendations.

For Renaissance, the content mapping solution has increased awareness and enthusiasm for AI across the company. Teams are now more comfortable engaging with AI development and contributing to training data, which has helped accelerate AI adoption within Renaissance. “The socialization of this project has helped the rest of the company look for other opportunities,” Robar said. “It has shown teams that it’s worth their time to produce training data because it pays back dividends.”

Renaissance’s AI-powered alignment solution shows how generative AI and human expertise can work together to streamline personalized learning recommendations for millions of students around the world.

Learn how AWS helps EdTechs build, deploy, and scale solutions that help students and educators succeed. Contact us today.