Stay up to date

The HMS Blog

The HMS Blog offers insights, solutions, and inspiration on how companies succeed with data and analytics.

Business professional using a tablet; digital hologram symbolizing AI and automated document generation.

October 28, 2025

How Multi-Agent Language Models Improve Complex Document Generation

How Multi-Agent Language Models transform the creation of complex documents via specialized AI agents that collaborate in planning, researching, verifying and writing.

...

Portrait photo of Cem Özcan, author of the article on Multi-Agent Language Models and AI-powered document generation.

Özgür Cem Özcan

October 22, 2025

Digital Sovereignty for AI and Analytics Platforms – From Principles to Practical Solutions

Sovereign IT architectures ensure control over data and systems without limiting flexibility. Through practical examples, we show how organizations reduce dependencies and achieve long-term resilience through deliberate technology choices.

...

Mark Bangert

Philipp Jakoby

Abstract 3D visualization of blue data points and connections representing a vector space and semantic search in GenAI applications

October 21, 2025

Optimizing Vector Stores: Lessons Learned from Real-World Projects

Vector stores are at the heart of modern GenAI applications – but scaling them efficiently is harder than it looks. From metadata management to hybrid search and ingestion pipelines: here are the common pitfalls and how to optimize them based on real project experience.

...

Lukas Neuhauser

Hand pressing digital TESTING button, symbolizing early LLM tracing and quality assurance in AI projects.

October 14, 2025

Early Tracing and Testing of LLM Calls: The Best Investment You Can Make

Early tracing and testing of LLM calls uncover issues before they get costly – ensuring faster time-to-market, lower risks, and scalable GenAI success.

...

Lorenz Jaenike

Fabian Wahren

Audience attending panel discussion at Snowflake World Tour & SPN Connect 2025 – HMS Blog insights on AI, data platforms and enterprise innovation.

October 9, 2025

Snowflake AI: Transforming Industries – Key Insights from World Tour & SPN Connect 2025

At SPN Connect and the Snowflake World Tour 2025, key insights revealed how industries like healthcare, media, and logistics are turning AI, AISQL, and Cortex into real business value. Read our recap to see what matters now for data teams and enterprise leaders.

...

Fabian Stadler

Dennis Stolp

Abstract wireframe with geometric mesh – symbol for data and technology

September 30, 2025

The Model Context Protocol (MCP): Deep Dive

This article explains how the Model Context Protocol (MCP) structures the interaction between host, client, and server, highlights key security aspects, and illustrates typical application workflows. It is aimed at architects, developers, and IT decision-makers looking to integrate LLMs securely and at scale into enterprise processes.

...

Kilian Schneider

Symbolische Darstellung von Wissen – Bücherregale voller Literatur, Metapher für Informationsfülle und Datenverarbeitung

September 23, 2025

Processing Large Document Sets with LLMs: Balancing Speed, Cost and Completeness

How can Large Language Models process massive document sets effectively? In our new blog post, we highlight the key challenges: context limits, throughput bottlenecks and balancing speed with completeness and share proven strategies from real-world projects.

...

Fabian Kaiser

Business professional using a digital chatbot application powered by AI for enterprise knowledge management

September 9, 2025

Scaling RAG Chatbots: Key Challenges and Best Practices

Global enterprises manage vast amounts of complex knowledge, from regulatory filings to sales materials. Traditional search often falls short. Retrieval-Augmented Generation (RAG) chatbots provide a breakthrough, but scaling them requires careful planning. In this article, we outline the five biggest challenges and the proven best practices that enable sustainable, compliant deployments across business units worldwide.

...

Lukas Neuhauser

Server racks in a modern data center with green LED lighting, symbolizing data processing and IT infrastructure

September 2, 2025

Optimizing Infrastructure for Large Language Models: What Really Matters on the Hardware Side

LLMs hold great promise, but hardware often becomes the real bottleneck. Discover what truly matters in balancing performance, stability, and cost when scaling AI infrastructure.

...

Robin Mader

Abstrakte digitale Schaltkreise in Blau- und Orangetönen, symbolisieren moderne Dateninfrastruktur und den Wandel analytischer Systeme im Kontext strategischer Entscheidungen zu SAS 9

March 17, 2025

SAS 9 Development Is Limited: Three Strategies for Shaping Your Analytics Future

How to Respond to the Limited Development of SAS 9? Explore the three key strategic options – with advantages and drawbacks – to support your next move. Find out how to reduce risk and cost while building a future-ready analytics environment.

...

Karsten Wohlgefahrt

Business-Meeting mit Laptop und Dashboards – symbolisiert datengetriebene Entscheidungen und strategische Vorbereitung auf eine SAS-Migration

April 29, 2025

Lessons Learned: How to Avoid Mistakes When Migrating from SAS

SAS has long been a core component of the analytics infrastructure in many organizations. However, rising license costs, the shift toward cloud-based architectures, and the desire for greater flexibility and long-term viability are prompting many companies to re-evaluate their existing SAS environments.

...

Karsten Wohlgefahrt

Abstract visualization of digital infrastructure: networked data lines overlaying an urban skyline at sunset – symbolizing technological connectivity and digital sovereignty.

July 30, 2025

IT Sovereignty in Practice: Identifying Risks, Reducing Dependencies

How much technological dependency is acceptable? This article offers guidance – with real-world relevance, clear fields of action, and key questions to strengthen digital sovereignty.

...

Mark Bangert

Philipp Jakoby

Abstrakte Netzwerkvisualisierung als Symbol für das Model Context Protocol (MCP) zur standardisierten Kontextverarbeitung in LLM-basierten KI-Systemen.

July 29, 2025

Model Context Protocol (MCP): Enabling Scalable and Standardized LLM Integration for Enterprises

Autonomous agents are on the rise – and the Model Context Protocol (MCP) is a key enabler. This article introduces MCP, explains why context management needs a standard, and shows how MCP simplifies integration and unlocks new use cases.

...

Kilian Schneider

Agentic AI visualisiert: Autonome Intelligenz in einer digitalen Gehirnarchitektur

July 8, 2025

What is Agentic AI?

Agentic AI is one of the most exciting concepts in artificial intelligence. It empowers autonomous AI agents to take on tasks independently, make decisions, and interact with systems – all with the goal of making complex processes more flexible and efficient. Discover how Agentic AI differs from traditional workflows, the advantages it offers businesses, and real-world examples of its application.

...

Luis Wirth

Nikola Ðondraš

Managing Director ,
Head of Sales, Marketing & HR

Have we piqued your interest?

Our sales team will be happy to advise you and give you even deeper insights into our project landscape.