This submit is co-written with Stefan Walter from msg.
With greater than 10,000 specialists in 34 nations, msg is each an impartial software program vendor and a system integrator working in extremely regulated industries, with over 40 years of domain-specific experience. msg.ProfileMap is a software program as a service (SaaS) answer for talent and competency administration. It’s an AWS Companion certified software program obtainable on AWS Market, at present serving greater than 7,500 customers. HR and technique departments use msg.ProfileMap for mission staffing and workforce transformation initiatives. By providing a centralized view of abilities and competencies, msg.ProfileMap helps organizations map their workforce’s capabilities, determine talent gaps, and implement focused growth methods. This helps more practical mission execution, higher alignment of expertise to roles, and long-term workforce planning.
On this submit, we share how msg automated knowledge harmonization for msg.ProfileMap, utilizing Amazon Bedrock to energy its giant language mannequin (LLM)-driven knowledge enrichment workflows, leading to larger accuracy in HR idea matching, decreased handbook workload, and improved alignment with compliance necessities beneath the EU AI Act and GDPR.
The significance of AI-based knowledge harmonization
HR departments face rising strain to function as data-driven organizations, however are sometimes constrained by the inconsistent, fragmented nature of their knowledge. Vital HR paperwork are unstructured, and legacy methods use mismatched codecs and knowledge fashions. This not solely impairs knowledge high quality but in addition results in inefficiencies and decision-making blind spots.Correct and harmonized HR knowledge is foundational for key actions corresponding to matching candidates to roles, figuring out inside mobility alternatives, conducting abilities hole evaluation, and planning workforce growth. msg recognized that with out automated, scalable strategies to course of and unify this knowledge, organizations would proceed to battle with handbook overhead and inconsistent outcomes.
Answer overview
HR knowledge is often scattered throughout numerous sources and codecs, starting from relational databases to Excel information, Phrase paperwork, and PDFs. Moreover, entities corresponding to personnel numbers or competencies have totally different distinctive identifiers in addition to totally different textual content descriptions, though with the identical semantics. msg addressed this problem with a modular structure, tailor-made for IT workforce eventualities. As illustrated within the following diagram, on the core of msg.ProfileMap is a sturdy textual content extraction layer, which transforms heterogeneous inputs into structured knowledge. That is then handed to an AI-powered harmonization engine that gives consistency throughout knowledge sources by avoiding duplication and aligning disparate ideas.
The harmonization course of makes use of a hybrid retrieval strategy that mixes vector-based semantic similarity and string-based matching strategies. These strategies align incoming knowledge with present entities within the system. Amazon Bedrock is used to semantically enrich knowledge, enhancing cross-source compatibility and matching precision. Extracted and enriched knowledge is listed and saved utilizing Amazon OpenSearch Service and Amazon DynamoDB, facilitating quick and correct retrieval, as proven within the following diagram.
The framework is designed to be unsupervised and area impartial. Though it’s optimized for IT workforce use instances, it has demonstrated robust generalization capabilities in different domains as properly.
msg.ProfileMap is a cloud-based software that makes use of a number of AWS companies, notably Amazon Neptune, Amazon DynamoDB, and Amazon Bedrock. The next diagram illustrates the complete answer structure.
Outcomes and technical validation
msg evaluated the effectiveness of the information harmonization framework by way of inside testing on IT workforce ideas and exterior benchmarking within the Bio-ML Monitor of the Ontology Alignment Analysis Initiative (OAEI), a world and EU-funded analysis initiative that evaluates ontology matching applied sciences since 2004.
Throughout inside testing, the system processed 2,248 ideas throughout a number of suggestion sorts. Excessive-probability merge suggestions reached 95.5% accuracy, overlaying practically 60% of all inputs. This helped msg cut back handbook validation workload by over 70%, considerably enhancing time-to-value for HR groups.
Throughout OAEI 2024, msg.ProfileMap ranked on the high of the 2024 Bio-ML benchmark, outperforming different methods throughout a number of biomedical datasets. On NCIT-DOID, it achieved a 0.918 F1 rating, with Hits@1 exceeding 92%, validating the engine’s generalizability past the HR area. Further particulars can be found within the official check outcomes.
Why Amazon Bedrock
msg depends on LLMs to semantically enrich knowledge in close to actual time. These workloads require low-latency inference, versatile scaling, and operational simplicity. Amazon Bedrock met these wants by offering a totally managed, serverless interface to main basis fashions—with out the necessity to handle infrastructure or deploy customized machine studying stacks.
Not like internet hosting fashions on Amazon Elastic Compute Cloud (Amazon EC2) or Amazon SageMaker, Amazon Bedrock abstracts away provisioning, versioning, scaling, and mannequin choice. Its consumption-based pricing aligns instantly with msg’s SaaS supply mannequin—sources are used (and billed) solely when wanted. This simplified integration decreased overhead and helped msg scale elastically as buyer demand grew.
Amazon Bedrock additionally helped msg meet compliance objectives beneath the EU AI Act and GDPR by enabling tightly scoped, auditable interactions with mannequin APIs—essential for HR use instances that deal with delicate workforce knowledge.
Conclusion
msg’s profitable integration of Amazon Bedrock into msg.ProfileMap demonstrates that large-scale AI adoption doesn’t require complicated infrastructure or specialised mannequin coaching. By combining modular design, ontology-based harmonization, and the absolutely managed LLM capabilities of Amazon Bedrock, msg delivered an AI-powered workforce intelligence platform that’s correct, scalable, and compliant.This answer improved idea match precision and achieved high marks in worldwide AI benchmarks, demonstrating what’s potential when generative AI is paired with the best cloud-based service. With Amazon Bedrock, msg has constructed a platform that’s prepared for at this time’s HR challenges—and tomorrow’s.
msg.ProfileMap is obtainable as a SaaS providing on AWS Market. In case you are all in favour of realizing extra, you may attain out to msg.hcm.backoffice@msg.group.
The content material and opinions on this weblog submit are these of the third-party writer and AWS just isn’t accountable for the content material or accuracy of this submit.
In regards to the authors
Stefan Walter is Senior Vice President of AI SaaS Options at msg. With over 25 years of expertise in IT software program growth, structure, and consulting, Stefan Walter leads with a imaginative and prescient for scalable SaaS innovation and operational excellence. As a BU lead at msg, Stefan has spearheaded transformative initiatives that bridge enterprise technique with expertise execution, particularly in complicated, multi-entity environments.
Gianluca Vegetti is a Senior Enterprise Architect within the AWS Companion Group, aligned to Strategic Partnership Collaboration and Governance (SPCG) engagements. In his position, he helps the definition and execution of Strategic Collaboration Agreements with chosen AWS companions.
Yuriy Bezsonov is a Senior Companion Answer Architect at AWS. With over 25 years within the tech, Yuriy has progressed from a software program developer to an engineering supervisor and Options Architect. Now, as a Senior Options Architect at AWS, he assists companions and prospects in creating cloud options, specializing in container applied sciences, Kubernetes, Java, software modernization, SaaS, developer expertise, and GenAI. Yuriy holds AWS and Kubernetes certifications, and he’s a recipient of the AWS Golden Jacket and the CNCF Kubestronaut Blue Jacket.
