Meshed Group improves student insight with AI-driven summaries using AWS

For more than 18 years, Meshed Group has delivered education management solutions that help higher-education providers manage the full student lifecycle. As education institutions expanded and student needs became more diverse, educators and administrative staff struggled to interpret large volumes of student information stored across records such as application, enrolment, academic performance, attendance, assessments, payments, at-risk indicators, and support actions. Staff often relied on multiple reports to understand a student’s academic progress or support needs, making it harder to spot early signs of academic or financial risk and provide timely assistance.

At the same time, generative AI was being rapidly adopted across the sector, influencing how students researched topics and how faculty assessed originality. Meshed recognized a similar opportunity within its own platform to help education providers interpret the data they already had. Pramesh Khadka, founder and CEO of Meshed Group, says, “Our focus is to empower education, enable innovation, and scale globally, and AI is an important part of that direction for us because it helps us make sense of complex student information and identify where support may be needed sooner.”

Meshed Group began developing its first generative AI capability by evaluating how Amazon Bedrock could support a more structured approach to interpreting student information. The team assessed several foundation models and tested how each processed academic, behavioral, and financial data points while keeping personally identifiable information (PII) out of model inputs. AWS solutions architects guided the process by helping Meshed understand model behavior, configure guardrails, and align testing workflows with the data privacy expectations of Australian education providers.

With this foundation, Meshed designed an AI-driven summary workflow that generates a concise 250-word narrative using structured student data. Rather than creating a general-purpose summary, the team refined prompts and model parameters to reflect higher-education context, ensuring the output addressed attendance patterns, assessment performance, submission history, and financial standing. The workflow sends only required non-PII data fields to a selected model in Amazon Bedrock, validates the returned text, and surfaces the summary through an on-demand action within the platform.

To determine the most suitable model, Meshed evaluated multiple foundation models available through Amazon Bedrock, comparing clarity, tone, and how accurately each model synthesized structured fields into a coherent narrative. These evaluations helped the team standardize summary generation and ensure consistent behavior across different student profiles. Khadka says, “Amazon Bedrock gave us a controlled environment to evaluate models, adjust parameters, and shape a workflow that reflects real academic scenarios, creating the foundation for what we plan to build next.”

Since introducing AI-driven student summaries, Meshed Group has given education providers a faster and more consistent method for understanding student progress in one place. Staff can now review consolidated academic, behavioral, and financial indicators without navigating multiple screens. Providers report an 80–85 percent reduction in the time spent synthesizing student information and are now reviewing twice as many students per day. This helps teams to identify learners who need support earlier in their academic journey and respond more proactively.

The structured summaries have also improved how decisions are made across institutions. Instead of relying on varied manual review practices, staff receive a standardized narrative that highlights the most relevant factors for discussion. This consistency has helped reduce review time by 8 minutes per student and strengthened how education providers prioritize learners who need attention. Khadka says, “Amazon Bedrock helps us bring together the information our clients already maintain and present it in a way that gives staff a clearer starting point for discussions with their students.”

As part of Meshed’s AI innovation roadmap, the company is developing an AI-assisted query feature powered by Amazon Bedrock to shape how customers plan for future reporting and analytics. Through demonstrations and early design discussions, customers have gained a clearer sense of how the feature could interpret reporting datasets and produce concise, text-based explanations. Once introduced, it is expected to reduce time spent navigating reports by 80 percent and make larger datasets easier for non-technical staff to interpret.