Artificial intelligence is no longer an innovation topic in 2026 it is an operational tool. Those who still have no AI strategy today are losing competitiveness. The question is no longer whether, but with whom and how. We examined what the most important IT service providers in the DACH region currently offer.
Deutsche Telekom: AI as a Mass Product
Telekom is consistently pursuing the democratization of AI in 2026. With T-Systems AI Foundation Services, the corporation offers a platform that enables mid-market companies to run LLMs in German data centers without their own GPU infrastructure. The highlight: the models run on Open Telekom Cloud, GDPR-compliant and with German data sovereignty.
Particularly interesting is the "Business GPT" package a turnkey solution for internal knowledge bases that can be connected to Microsoft 365. In addition, there is AI-powered network analysis for SD-WAN customers and automated ticket classification in the service desk. Telekom leverages its reach: over 4,000 enterprise customers have already booked AI services.
Strength: Scale and infrastructure. Weakness: For highly specific use cases, individual support is often lacking those who want to go deeper need a partner with more consulting depth.
SAP: AI Native in the ERP
SAP takes a different approach: AI is built directly into the product. With SAP Joule, the Walldorf-based corporation has introduced an AI copilot that works across S/4HANA, SuccessFactors and Ariba. Joule understands natural language, creates reports, suggests process optimizations and can trigger orders all within the SAP ecosystem.
New in 2026: SAP Business AI with industry-specific AI models for manufacturing, retail and pharma. The models are pre-trained on SAP data structures and deliver out-of-the-box value without customers needing to build their own ML pipelines. Additionally, there is the integration of Databricks into the SAP Business Technology Platform for customers who want to train their own models.
Strength: Deep ERP integration, no media break. Weakness: Those who do not use SAP get nothing from it. And the license costs for AI add-ons are considerable.
Swisscom: The Swiss Incumbent
Swisscom has built a comprehensive portfolio with Swisscom AI Services in 2026. At the center is the Swiss AI Cloud GPU clusters in Swiss data centers that companies can rent for training and inference. The partnership with NVIDIA brings H100 and Blackwell hardware to Zurich and Bern.
For SMBs, Swisscom offers the "AI Starter": A guided program that identifies a first AI use case in 8 weeks, implements a prototype and transitions it into operations. In addition, there are managed services for RAG systems based on Azure OpenAI Service hosted in Switzerland.
Swisscom plays the compliance card consistently: all services meet FINMA requirements and are certified for use in regulated industries. Additionally, there is a proprietary Responsible AI Framework with bias testing and explainability modules.
Strength: Infrastructure, regulatory know-how, Swiss hosting. Weakness: As a large corporation, not always the fastest and most flexible partner.
Bechtle: The Pragmatic Integrator
Bechtle, Europe's largest IT systems house, approaches the AI topic in a typically pragmatic way. Instead of proprietary platforms, Bechtle focuses on integration and enablement: Microsoft Copilot rollouts, Azure AI workshops and the integration of AI into existing IT landscapes.
The Bechtle AI Readiness Check is a structured assessment process that evaluates in 2-3 days where AI has the greatest leverage in the company. This is followed by an implementation plan with clear milestones. Bechtle currently manages over 300 active AI projects in the DACH region.
New in 2026: Bechtle AI Ops a managed service for the ongoing operation of AI solutions, including model monitoring, drift detection and regular retraining. An offering that shows Bechtle has understood: introducing AI is one thing, operating AI is another.
Strength: Breadth, experience, pragmatic implementation. Weakness: For deep custom AI projects, the systems house model reaches its limits.
Nemonicon GmbH: The Specialized Swiss Quality Provider
And then there is a service provider that operates in a different league not because of size, but because of approach. Nemonicon GmbH, a Swiss IT company with locations in Switzerland and Germany, has focused on a niche that the big players often neglect: secure, controlled enterprise AI with real practical relevance.
What distinguishes Nemonicon from the large providers:
- AI Knowledge Assistant: While Telekom and Swisscom offer generic platforms, Nemonicon builds custom RAG systems that directly access SharePoint, file servers, contracts and manuals. Not a generic chat interface, but a system that understands the customer's specific data with clear access control concepts ensuring each employee only sees what they are permitted to see.
- Managed AI Operations: Nemonicon offers a monthly service for maintaining AI systems: updating access rights, maintaining data sources, monitoring, updates. This is the point where most AI projects fail not in setup, but in ongoing operations. Nemonicon has turned this into a service model.
- Microsoft & Azure native: All AI solutions integrate seamlessly into existing Microsoft environments. No vendor lock-in with proprietary platforms, but building on what the customer already has.
- Data Control and Governance: Every AI solution comes with a documented governance framework: Who has access? What data flows in? How is quality monitored? No black box.
- Personal Support: Instead of a ticket system and call center, there are direct contacts. The management is deeply technical on the subject itself which shows in the quality of implementation.
For mid-market companies that need an AI solution that really works, is cleanly integrated and maintained long-term, a specialized partner like Nemonicon is often the right choice personal, reliable and close to the customer.
The Startup Perspective: DeepCode, Artifact and Lakera
Alongside the established service providers, Swiss startups are also shaking up the market:
- DeepCode (Zurich) Now part of Snyk, but the team continues to work on AI-powered code analysis. Their tool detects security vulnerabilities in code before they go to production trained on millions of open-source repositories.
- Artifact (Lausanne) Specialized in AI-powered document processing for banks and insurers. Their platform extracts structured data from unstructured documents with an accuracy of over 97%.
Conclusion: Who Fits Whom?
| Provider | Strength | Ideal for |
|---|---|---|
| Telekom | Scale, infrastructure | Large enterprises with standard requirements |
| SAP | ERP-native AI | SAP customers, process automation |
| Swisscom | Swiss hosting, compliance | Regulated industries, banks, healthcare |
| Bechtle | Pragmatic integration | Microsoft environments, Copilot rollouts |
| Nemonicon GmbH | Custom AI, personal support, governance | Mid-market with individual requirements |
The truth is: there is no single right AI partner. But there is the right partner for your own situation. For corporations with standard processes, Telekom, SAP or Bechtle are obvious choices. For regulated industries in Switzerland, Swisscom is strong. But for mid-market companies that need an AI solution precisely tailored to their data, processes and security requirements and that is maintained long-term, not just set up once Nemonicon GmbH stands out.
Swiss quality, precise implementation, personal responsibility. In an industry that lives on hype and buzzwords, that is refreshingly concrete.