Monday, 30 June 2025
Implementing a new Curriculum Management System (CMS) is a major undertaking for any university. These platforms are vital for managing and publishing course catalogues, coordinating academic planning, and ensuring compliance with regulatory and internal governance standards. However, one of the most complex and time-consuming aspects of implementation is preparing the curriculum data for migration into the new system.
Most universities hold curriculum data across a patchwork of unstructured and inconsistent formats—Word documents, Excel files, legacy Student Management System (SMS) records and emails, to name but a few. Bringing this information together, cleaning it, standardising it, and validating it is often a manual process that can take months.
At Prolifics, we understand the challenges facing our Higher Education customers; we’ve explored how GenAI can streamline and accelerate this process. Our Innovation Centre has demonstrated how GenAI can intelligently extract, interpret, cleanse and restructure data ready for CMS ingestion—cutting delivery timelines dramatically and reducing the dependency on overstretched registry and academic teams.
Data Preparation for CMS Implementation
Launching a new CMS typically begins with data mapping and migration activities—but many universities underestimate the scale of work required to get their curriculum data fit for purpose. Curriculum information is often fragmented, inconsistently structured, and heavily reliant on manual collation. This creates a bottleneck that can stall timelines and consume significant internal effort.
Common problems faced during CMS implementations include:
- Unstructured source material: Course data often lives in Word documents, Excel spreadsheets, or PDF forms created and maintained by individual departments, rather than central systems.
- Lack of standardisation: Module templates vary between faculties or change over time, with inconsistent labelling and formatting of key fields like credits, assessments, or prerequisites.
- Manual data extraction and rekeying: Teams frequently resort to copying and pasting content into CMS upload templates, with limited validation at source.
- Inaccessible legacy knowledge: Understanding the rationale behind legacy course structures often depends on a small number of individuals familiar with historic decisions.
- Delayed testing and configuration: Without a reliable set of clean, structured data, it becomes difficult to test the new CMS properly or configure workflows and rules effectively.
These challenges don’t just affect data quality—they increase implementation risk, delay go-live dates and create downstream problems when students and staff begin to use the new system.
By introducing GenAI into the data preparation phase, universities can dramatically reduce the manual overhead, improve consistency, and give project teams a much faster route to configuring and validating their new CMS.
Traditional ETL Approaches Compared
Many universities approach CMS implementations with the expectation that existing ETL (Extract, Transform, Load) processes or scripts can be adapted to migrate curriculum data. While ETL is ideal for structured, well-defined datasets, it quickly becomes unmanageable when dealing with curriculum data that is varied, loosely formatted, and maintained in inconsistent ways.
Common limitations of ETL in this context include an inflexibility with unstructured data, as well as significant effort configuration, setting up rules, mappings and validation logic.
For structured SIS data (e.g. student records or enrolment history), ETL remains valuable. But for curriculum content—modules, assessments, learning outcomes, credit frameworks—a more intelligent, adaptable approach is required. This is where GenAI delivers a significant advantage.
GenAI: A Smarter Way to Prepare Curriculum Data
GenAI offers a fundamentally different approach to preparing curriculum data for CMS migration. Rather than relying on rigid mappings and predefined structures, it uses advanced language models to interpret natural language and infer context—just as a human would.
Using GenAI, for example tools from Google or OpenAI, universities can:
Extract structured fields from unstructured content: GenAI can read through Word documents, PDFs, and other free-text formats, identifying and extracting key fields such as module codes, credit values, assessment methods, and prerequisites.
Standardise and reformat content automatically: Output can be normalised into the required CMS schema, ensuring consistency across faculties and departments.
Flag data quality issues: The model can detect inconsistencies, missing fields, or conflicting values, reducing the need for multiple rounds of manual review.
Create import-ready files: Data can be exported in the correct format for CMS ingestion, significantly reducing the time spent on template population and manual rekeying.
Prolifics’ proof-of-concept exercises have shown that GenAI can reduce the data preparation phase from several months of manual work to just a few days—without compromising on quality. It supports collaboration between registry and academic teams by allowing them to focus on verification rather than data entry.
By embedding GenAI into the CMS implementation process, universities can accelerate delivery, improve data governance, and reduce the operational strain on internal teams.
The Proof of Concept
To validate the use of GenAI in a real-world setting, we ran a series of proof-of-concept exercises focused on preparing typical curriculum data for a CMS implementation. We tested how GenAI could automate the most time-consuming tasks typically faced at the start of these projects.
Key outcomes included:
Rapid processing of unstructured source data: GenAI successfully interpreted module specifications held in Word and PDF documents, extracting key fields such as credit value, module leader, assessments, and prerequisites with high accuracy.
Standardised outputs across departments: Despite inconsistencies in formatting and terminology, the model was able to create structured, uniform datasets aligned to the CMS data schema.
Integrated validation checks: Built-in logic flagged missing data, duplicate entries, and values outside expected ranges, enabling earlier identification of issues that would otherwise surface during testing or go-live preparation.
End-to-end output in under five working days: A process that previously required multiple academic and admin staff over several months was reduced to a few focused days of activity, supported by SME review rather than manual input.
By using GenAI, we removed a major implementation bottleneck and freed up resources to focus on configuration, user training, and governance—accelerating the overall project timeline and improving data confidence ahead of CMS launch.
Benefits for Universities
Adopting GenAI as part of a CMS implementation brings measurable benefits across planning, delivery, and governance:
- Faster implementation timelines: Reducing the time required to cleanse and structure curriculum data allows institutions to meet project milestones without relying on prolonged manual effort.
- Reduced pressure on SMEs and registry teams: Rather than manually rekeying data, staff focus on verification, oversight, and process improvement.
- Improved data quality: GenAI not only extracts data but validates and standardises it, reducing post-migration errors and the risk of downstream issues.
- Greater flexibility in project delivery: With a repeatable approach that works across faculties and formats, GenAI supports staggered or phased CMS rollouts with ease.
- Future readiness: The same models can later be adapted to support change control processes, module updates, and curriculum reviews, offering long-term value beyond initial setup.
Ultimately, this approach enables universities to modernise faster, reduce cost and risk, and shift the focus from manual data handling to strategic curriculum planning and student experience.
Jonathan Binks - Head of Delivery
Prolifics Testing UK
