How Nexford University is building an AI-native university with Andela & Anthropic

SNAPSHOT

Customer: Nexford University
Industry: Higher education
Use case: AI-powered university operations, curriculum design, assessment, and software development
Tools: Claude Code, Claude API, Model Context Protocol (MCP), Claude Code Skills

What was built: Nex-OS, an AI-powered operating system spanning curriculum, assessment, enrollment, learner success, finance, and internal operations. Claude supports both the development of the platform and AI-powered workflows within it, with human review and approval built into consequential decisions.

Early impact:

• 525,424 enrollment contacts analyzed and scored, bringing full-population analysis to enrollment operations
• 2,000 hours of analyst time saved by building a taxonomy through Claude vs building manually
• 1,447 learner submissions graded through Nexford’s AI-supported assessment system
• 425 curriculum gaps across 69 courses identified through AI-powered course diagnostics

Problem: personalization was becoming harder to deliver at scale

Nexford University is a global online university built around a fundamental challenge in higher education: how to deliver education that stays relevant to a changing labor market while giving learners meaningful, individualized support at scale.

Keeping curriculum current requires continually understanding how employer demand, skills, and professional standards are changing. Across a portfolio of roughly 200 courses, that means processing a volume of labor-market information that is difficult to analyze manually and even harder to revisit frequently enough to keep pace with change.

Assessment presents a similar challenge. Personalized feedback and opportunities to apply that feedback through multiple attempts can improve the learning experience, but providing that level of individualized assessment across thousands of learners creates significant faculty workload. The same scale challenge extends beyond academics. Enrollment teams work across hundreds of thousands of prospective learners, while finance, learner success, and other university functions depend on large amounts of information spread across different systems and workflows.

For Nexford, the opportunity was not simply to automate individual tasks. It was to find a way to make evidence-driven, personalized operations possible across the university without requiring staff and faculty workload to grow at the same rate.

Solution: an OS connecting AI across curriculum, assessment, and operations

Nexford worked with Andela engineers to build Nex-OS, a connected set of AI-powered products supporting work across the university. Andela engineers worked alongside Nexford’s team to build and scale the platform using Claude, bringing AI into both the software development process and the workflows running inside Nex-OS.

Claude plays two roles in the model: Claude Code helps the engineering team build and maintain the platform, while the Claude API powers AI-supported workflows across curriculum, assessment, enrollment, finance, and learner operations. Nexford also uses Model Context Protocol (MCP) to give Claude Code access to engineering tools and context across its development environment, while centrally managed Claude Code Skills standardize how agents work across the platform. Together, these capabilities let Nexford use Claude both to build Nex-OS and to power the workflows running inside it. 

One of the clearest examples is Nexford’s Workplace Alignment Model, or WAM. The system analyzes vetted labor-market and employer-demand signals and translates them into structured capabilities that inform curriculum and assessment design. Nexford’s Learning Design team retains authority over curriculum decisions, reviewing sources and approving mappings before they become part of a course.

“We’re not bolting AI onto a university. We’ve built it into how Nexford operates — using Claude to analyze what employers are hiring for every week, keep our curriculum current, and give our teams better evidence while people remain in control.”

— Fadl Al Tarzi, CEO, Nexford University

That labor-market context connects directly to Nexford’s Assessment Management System. Claude helps build assessments and grade learner submissions against Nexford-defined rubrics, creating structured feedback learners can act on and making repeat attempts more practical at scale. Higher-stakes capstone assessments remain faculty-graded.

Together, the teams extended this model across enrollment, finance, course quality, learner success, and internal operations.

The Nex-OS Claude operating model:

• Connect curriculum to the labor market: WAM continuously analyzes vetted employer and labor-market signals, helping Nexford’s Learning Design team identify how skills are changing and where curriculum may need to evolve.
• Make personalized assessment scalable: Claude supports rubric-based grading and structured feedback through Nexford’s Assessment Management System, giving learners actionable feedback while maintaining consistent assessment rules.
• Analyze operations at population scale: The Enrollment Intelligence Center analyzes enrollment data across hundreds of thousands of contacts, helping teams identify patterns and prioritize where human attention is most valuable.
• Continuously diagnose course quality: Course Diagnostics analyzes live course content for gaps and alignment issues, surfacing findings for human reviewers rather than changing courses autonomously.
• Build the platform with Claude Code: Claude Code is embedded throughout Nexford’s engineering workflow. Claude co-authorship appears on 90% of commits touching WAM, 87% of FinOps commits, 83% of WorQ commits, and 95% of Nexus commits.
• Standardize how agents build: Nexford centrally maintains 31 Claude Code Skills that encode its engineering practices and make them available across connected agents. Updates can be made once and propagated across the environment, while Nexford’s design system is distributed as a versioned Claude Code plugin.
• Give Claude Code access to the engineering environment: Nexford’s nxu-tools platform uses MCP servers to connect Claude Code to code intelligence, work items, pull requests, build pipelines, environments, and governance across roughly 100 repositories. Its assessment infrastructure also uses a dedicated MCP server.
• Keep people in control: Each Nex-OS product has a defined human-judgment layer. Curriculum designers approve capability mappings, faculty retain control of high-stakes assessments, finance AI is read-only, and AI-supported decisions are logged through Nexford’s central work and decision system.

Impact: More analysis, broader coverage, less manual work

Nex-OS has allowed Nexford to apply AI across work that previously depended on significant manual analysis, while preserving human judgment around curriculum, assessment, financial reporting, and learner support.

Working alongside Nexford, Andela engineers helped turn that vision into an operating platform, using Claude to build and scale AI-powered workflows across the university. The impact is visible in the volume of work Nexford can now analyze and support:

• 525,424 enrollment contacts analyzed and scored, giving enrollment teams population-level insight rather than relying on a limited manual sample.
• 2,000 hours of analyst time saved by building a taxonomy through Claude vs building manually
• 1,447 learner submissions graded through Nexford’s AI-supported assessment infrastructure, expanding the university’s ability to provide structured, rubric-based feedback.
• 425 curriculum gaps identified across 69 courses, helping Nexford systematically evaluate where course content needs attention.
• 140 capability statements across 15 disciplines, connecting labor-market evidence to curriculum and assessment design.
• 14 university departments using WorQ, Nexford’s shared decision log and human work queue for AI-supported operations.

The larger change is how continuously Nexford can operate. Labor-market analysis can happen as the employer needs to change. Course diagnostics can extend across a growing portfolio. Enrollment analysis can cover the full population rather than a sample. And structured assessment feedback can reach more learners without requiring faculty effort to increase proportionally.

For Andela, the engagement also demonstrates how AI-skilled engineering teams can help organizations move beyond isolated AI use cases and build AI into the systems that run the business. By combining Nexford’s expertise in education with Andela’s engineering capabilities and Claude, the teams have created an operating model that can continue to expand as new AI use cases emerge.

“Nexford shows what’s possible when you combine deep domain expertise with AI-native engineering. Working alongside Nexford, our engineers used Claude to help turn an ambitious vision into an operating platform that now spans curriculum, assessment, enrollment, and university operations. What stands out is the scale: AI isn’t supporting one isolated workflow. It’s becoming part of how the university operates, while keeping human judgment at the center.”

— Kennith Jackson, SVP, AI Solutions, Andela 

‍