Novagentica launches platform to let industry run AI agents inside their own systems
Novagentica launches a platform allowing industrial firms to run AI agents inside their own systems, reducing vendor lock-in and addressing geopolitical model restrictions.
Martin Hofmann, the former Volkswagen and Salesforce executive, has founded Novagentica and introduced a software layer that enables companies to deploy AI agents across their internal IT landscapes without being tied to single-provider models. The platform connects customers to multiple AI models and executes agent workflows within a client’s infrastructure, a design intended to preserve operational control and data governance. Hofmann positions the product as a response to both commercial dependence on U.S. model providers and the practical challenges of integrating generative AI into industrial processes.
Platform purpose and technical approach
Novagentica provides an orchestration interface that routes tasks to different underlying AI models while running agent logic inside customers’ systems. This allows companies to keep data transfers minimal and maintain control over where sensitive processing occurs. The company’s approach is to support proprietary cloud models, vendor APIs, and open-source alternatives so businesses can switch providers or fall back to local models if an external service becomes unavailable.
The architecture emphasizes an “expert in the loop” oversight model where human operators retain final decision authority while agents handle routine or structured operations. Novagentica’s agents are built to automate recurring tasks such as ordering components, monitoring quality metrics, and pre-screening candidate resumes, according to descriptions of typical use cases. The product thus aims to convert large-language-model capabilities into deterministic, auditable workflows suited to regulated and safety-critical environments.
Founder background and market positioning
Martin Hofmann brings nine years of IT leadership at Volkswagen and a subsequent stint at Salesforce to his new venture, blending automotive systems experience with Silicon Valley product practices. He has framed Novagentica as an attempt to translate large-model advances into industrial-grade applications that respect the operational constraints of manufacturing, defense and financial services. Hofmann has said his time at a major automaker exposed him to the friction companies face when adopting new information technologies at scale.
Novagentica’s value proposition leans on that operational credibility: the platform is pitched to IT departments that need predictable integrations and strong governance rather than experimental, externally hosted AI prototypes. The company argues that Europe may have ceded early leadership in base model development but can compete in the practical deployment and responsible adoption of AI agents across enterprise systems.
Early customer pipeline and industry reach
Hofmann has disclosed that Novagentica has already engaged a mix of industrial clients, including a major automotive supplier headquartered in southern Germany, a prominent defense contractor, and a division of a southern German carmaker. Discussions are reportedly underway with banking institutions and other operators of critical infrastructure. The target market spans large DAX corporations down to Mittelstand firms that manage complex manufacturing and logistics workflows.
These initial customer relationships suggest a focus on sectors where uptime, compliance and data sovereignty are non-negotiable. For many such organizations, the ability to run AI agents within their existing ERP, PLM and manufacturing-execution systems is a practical prerequisite for adoption. Novagentica’s positioning emphasizes integration into established IT landscapes rather than requiring customers to migrate workloads to new cloud-native stacks.
Funding plans and investor strategy
Hofmann self-funded the company’s initial build and says there is strong interest from U.S. investors for the upcoming seed round. Despite that demand, Novagentica intends to prioritize “globally operating European investors” to align financing with a strategy that stresses European data governance norms and market access. The choice reflects a desire to avoid strategic dependencies that could arise from investor influence or technology commitments favoring a single foreign provider.
The seed round will be aimed at expanding engineering capacity, accelerating enterprise sales, and beefing up compliance and security features that enterprise clients expect. Observers note that early traction with blue-chip customers can materially increase the company’s leverage when negotiating terms with lead investors and enterprise partners.
Geopolitics, model access and vendor risk
Recent U.S. export restrictions and access controls on advanced models have sharpened European corporate concerns about reliance on a narrow set of providers. Measures that temporarily limited foreign access to leading models such as Anthropic’s advanced releases and selective distribution of newer OpenAI models have highlighted how policy decisions can affect commercial availability. Novagentica frames its platform as a hedge against such vendor or geopolitical risk by enabling multi-model deployments and local hosting where appropriate.
The company’s messaging stresses flexibility: customers can connect to cutting-edge hosted models when available, but they can also pivot to open-source or locally deployed alternatives if access is curtailed. This strategy aims to preserve continuity of service and to minimize the business impact of sudden policy changes or platform decisions by third parties.
Open-source models and performance trade-offs
Novagentica also promotes the use of mature open-source models where they satisfy enterprise requirements, even while acknowledging these models typically trail leading commercial offerings in raw capability. The platform accommodates this trade-off by allowing companies to select different models for different tasks—using top-tier hosted models for complex reasoning and open-source or on-premise models for routine automation or privacy-sensitive workloads.
The company argues that flexibility and the ability to orchestrate model selection per workflow are more valuable to most industrial users than dependence on the absolute cutting edge. For regulated sectors, determinism, auditability and the ability to demonstrate consistent outcomes often outweigh marginal improvements in general-purpose language performance.
Novagentica’s launch marks a pragmatic entry into the enterprise AI market, targeting customers that demand control, continuity and integration rather than speculative innovation alone. The company’s progress over the next twelve months—measured by customer deployments, regulatory compliance capabilities and the composition of its investor base—will determine whether a multi-model, in-system agent strategy gains traction across Europe’s industrial heartland.
