Syntegon Strategically Brings Generative AI into Daily Work

Moving from early adoption to an enterprise-wide AI platform strategy

Syntegon recognized early that generative AI had great potential for knowledge work. Next, it moved beyond individual use cases and developed an enterprise-wide platform strategy.

The strategy focused on a solution that would enable productive use cases while helping protect sensitive information. The strategy called for controlled integration of internal knowledge from systems such as Confluence and SharePoint into AI-supported workflows in accordance with existing access rights.

This meant supplementing organizational guidelines with technical IT safeguards such as centralized authentication, role-based access controls, traceable logging, and an architecture open enough to integrate different models, knowledge sources, and capabilities.

From the outset, Syntia was conceived as a strategic enterprise platform closely aligned with business teams, open to different models and knowledge sources, and grounded in a reliable governance framework.

An open platform combines security with technological flexibility

Syntegon intentionally chose an open platform architecture that combines security, ease of use, and technological flexibility. Together with SoftwareOne, Syntegon created Syntia as the Company’s central entry point for generative AI. The solution is based on Open WebUI and enables authorized employees to use different large language models and AI capabilities for different tasks using a consistent, multilingual interface.

Governance as part of the user experience

Security and governance are integral to the user experience design. Employees sign in with their Syntegon corporate account. Before using the platform for the first time, they receive the Code of Conduct and must actively accept its terms. Only then can they access the chatbot interface. Relevant interactions and system events are logged in compliance with data privacy requirements. In this way, Syntia provides technical support for the existing AI Guideline and establishes a reliable framework for the productive use of generative AI.

Syntia fulfills Syntegon objectives by bringing together several practical use cases. Employees can summarize documents, translate text, generate images from text prompts, research current information through web search, and interact with connected large language models. In addition, knowledge from Confluence is integrated into the platform, and the incorporation of SharePoint content is being evaluated in a proof of concept.

Flexible architecture and targeted enhancements

Syntia uses containerized cloud architecture technology. Open WebUI provides the open user interface, while AWS Bedrock and Azure OpenAI provide access to various AI models. Entra ID handles authentication; role-based permissions, access controls, IAM mechanisms, and end-to-end TLS encryption to protect the platform and its resources.

The infrastructure is defined as code using Terraform, enabling automated and repeatable deployment. Docker containers and a CI/CD pipeline through GitHub support controlled deployment and ongoing development. The underlying code was version-controlled and provided to Syntegon.

SoftwareOne developed targeted technical enhancements to meet specific requirements. Because Open WebUI cannot communicate natively with AWS Bedrock, a Bedrock Access Gateway was implemented as an intermediary layer. Additional middleware keeps the architecture open and reduces dependence on any single vendor. Syntegon and SoftwareOne also took a pragmatic, controlled approach to integrating SharePoint content. Because no standard interface to Open WebUI was available, they implemented a clearly scoped proof of concept. Relevant documents were exported on a schedule, converted into a searchable format, and imported into the knowledge base using scripts. Based on the results, Syntegon decided how to proceed with the integration.

Agile implementation with knowledge transfer

SoftwareOne supported Syntegon from architecture and requirements alignment through knowledge transfer. The scope of services included project management, infrastructure automation, container build and deployment, configuration and security hardening of Open WebUI, integration of AI models, search capabilities and knowledge sources, architecture and security documentation, complete, editable project documentation in Syntegon’s Confluence environment.

Implementation followed an agile approach with clearly defined steps. A customer-side product owner prioritized the requirements. Regular alignment meetings, a visual work board, reviews, and retrospectives made progress and decisions transparent. The team jointly reviewed individual components with Syntegon during demo sessions against defined acceptance criteria. In this way, Syntia evolved step by step into a functional minimum viable product (MVP) for production use.

After go-live, an additional hypercare phase supported the transition into active use. For two weeks, the SoftwareOne development team answered technical questions, handled service requests, made operational adjustments, and provided extended support. Daily check-ins helped the teams quickly assess issues during the initial phase and address them together.