This post is co-authored with Stefan Walter of MSG.
With over 10,000 experts in 34 countries, MSG is both an independent software vendor and a highly regulated industry-operating system integrator with over 40 years of domain-specific expertise. MSG.ProfileMap is software as a service (SAAS) solution for skill and competence management. It is an AWS Partner-Certified Software available on the AWS Marketplace and currently serves more than 7,500 users. HR and Strategies are using MSG.ProfileMap for project staffing and workforce transformation initiatives. By providing a focused view of skills and capabilities, MSG.ProfileMap helps organizations map workforce capabilities, identify skill gaps, and implement target development strategies. This supports more effective project execution, better alignment of talent for roles, and long-term workforce planning.
This post shares how to use MSG Automated Data Harmonization in MSG.ProfileMap to enhance large-scale language model (LLM)-driven data enrichment workflows using Amazon Bedrock, providing greater accuracy in HR concept match and improved coordination with EU AI ACT and GDPR configuration requirements.
The importance of AI-based data harmony
The HR department faces increased pressure to act as a data-driven organization, but is often constrained by the inconsistent, fragmented nature of data. Important HR documents are unstructured, with legacy systems using formats of inconsistencies and data models. This not only impairs data quality, but also leads to inefficiency and blind spots in decision-making. Explanatory and harmonious HR data are fundamental for key activities such as matching candidates to roles, identifying internal mobility opportunities, conducting skill gap analysis, and planning workforce development. MSG has identified that without an automated, scalable way to process and integrate this data, organizations will continue to struggle with manual overhead and inconsistent outcomes.
Solution overview
HR data is usually scattered across a wide variety of sources and formats, ranging from relational databases to files, word documents, and PDF Excel. Furthermore, entities such as people and competency use the same semantics, but have different unique identifiers and different textual descriptions. MSG addressed this challenge with a modular architecture tailored to the IT workforce scenario. As shown in the following diagram, the MSG.ProfileMap core has a robust text extraction layer that converts non-uniform input into structured data. This is passed to an AI-powered harmonization engine that provides consistency across data sources by avoiding duplication and coordinating different concepts.
The harmonic process uses a hybrid search approach that combines vector-based semantic similarity with string-based matching techniques. These methods line up incoming data with existing entities in the system. Amazon Bedrock is used to semantically enrich data and improve accuracy that matches cross-source compatibility. The extracted enrichment data is indexed and stored using Amazon Opensearch Service and Amazon Dynamodb to facilitate quick and accurate searches as shown in the following diagram.
This framework is designed to be unsupervised and domain-independent. It is optimized for use cases in the IT workforce, but also demonstrates strong generalization capabilities in other domains.
MSG.ProfileMap is a cloud-based application that uses several AWS services, particularly Amazon Neptune, Amazon Dynamodb and Amazon Bedrock. The following diagram illustrates the complete solution architecture.
Results and technical verification
Since 2004, MSG has evaluated the effectiveness of the data harmonization framework through internal testing with IT workforce concepts and external benchmarks on the BioML Track of the Ontology Alignment Assessment Initiative (OAEI), a research initiative funded by the Ontology Alignment and EU.
During internal testing, the system processed 2,248 concepts across multiple proposal types. The highly probable merge recommendation reaches 95.5% accuracy, covering almost 60% of all inputs. This has significantly improved the value of HR teams by reducing manual verification workloads by more than 70%.
OAEI 2024, MSG.ProfileMap ranked top of the 2024 BIO-ML benchmark, surpassing other systems in multiple biomedical datasets. NCIT-DOID has tested the generalizability of engines that have achieved a 0.918 F1 score and exceeded 92% with HIT@1 exceeding the HR domain. Additional details are available in official test results.
Why Amazon's rocks
MSG relies on LLMS to semantically enrich data in near real-time. These workloads require low latency inference, flexible scaling, and operational simplicity. Amazon Bedrock met these needs by providing a fully managed serverless interface to its leading foundation model, as it requires managing its infrastructure and deploying custom machine learning stacks.
Unlike the Amazon Elastic Compute Cloud (Amazon EC2) or Amazon Sagemaker hosting models, Amazon Bedrock abstracts provisioning, versioning, scaling, and model selection. Its consumption-based pricing matches directly with MSG's SaaS delivery model. Resources are only used (and billed) when needed. This simplified integration reduced overhead and eased MSG scale as customer demand increased.
Amazon Bedrock has also helped MSG achieve its compliance goals under EU AI Act and GDPR. This allows for close auditable interactions with the model API for HR use cases that process sensitive workforce data.
Conclusion
The success of MSG's integration of Amazon Bedrock into MSG.ProfileMap shows that large-scale AI adoption does not require complex infrastructure or specialized model training. Combining modular design, ontology-based harmony, and Amazon Bedrock's fully managed LLM capabilities, MSG has provided an accurate, scalable, compliant, AI-powered workforce intelligence platform. With Amazon Bedrock, MSG has built a platform that can address today's HR challenges and tomorrow's challenges.
MSG.ProfileMap is available as a SaaS offering on the AWS Marketplace. If you want to know more, you can visit msg.hcm.backoffice@msg.group.
The content and opinions of this blog post are those of third party authors and AWS is not responsible for the content or accuracy of this post.
About the author
Stephen Walter He is Senior Vice President of AI SAAS Solutions at MSG. With over 25 years of experience in IT software development, architecture and consulting, Stefan Walter leads the vision of scalable SaaS innovation and operational excellence. As MSG's BU lead, Stefan led a transformational initiative that bridges business strategy to technology implementation, particularly in complex, multi-center environments.
Genruka Vegiti He is a senior enterprise architect for AWS Partner Organizations, tailors to Strategic Partnership Collaboration and Governance (SPCG) engagement. In his role, he supports the definition and execution of strategic collaboration agreements with Selected AWS Partners.
Yuriy Bezsonov I am AWS Senior Partner Solutions Architect. For over 25 years, Yuriy has since forged from software developer to engineering manager and solution architect. Currently, as a senior solution architect at AWS, he helps partners and customers develop cloud solutions, focusing on container technology, Kubernetes, Java, application modernization, SAAS, developer experience and Genai. Yuriy is AWS and Kubernetes certified and is a recipient of AWS Golden Jacket and CNCF Kubestronaut Blue Jacket.
