Enterprise AI shifts toward open-weight models and agentic workflows

The enterprise landscape is undergoing a significant shift in how artificial intelligence is deployed and utilized. Organizations are increasingly moving away from closed frontier models in favor of open-weight models. This transition is driven by the need for modularity, greater control, deeper customization, and enhanced protection of proprietary intellectual property. While open models offer strategic independence and the ability to encode specific enterprise context into the AI stack, the economic benefits extend beyond simple token savings to include the complexities of operating and maintaining these models internally.

Simultaneously, business software is evolving from traditional systems of record into an agentic business fabric. Rather than employees manually navigating applications like CRM or ERP, AI agents are beginning to orchestrate workflows and execute tasks across various functions. This shift is expected to blur departmental boundaries and transform software economics, moving vendors from per-seat licensing toward consumption- or outcome-based pricing models. Front-office applications, such as customer service and sales platforms, are leading this transition due to their proximity to conversational data and measurable business outcomes.