From Visible Platforms to Dark Software Infrastructure
ERP software evolution in the AI agent era describes the shift from feature-heavy, user-facing systems to invisible infrastructure that AI agents call through APIs to run core business processes, making the software itself largely hidden while its logic and data remain essential. In this model, users talk to AI agents that orchestrate workflows across finance, HR, manufacturing, and operations, while the ERP runs quietly in the background. Younglimwon Soft Lab executive Ho Woong-ki describes this as “dark software,” drawing an analogy to fully automated “dark factories” with no human workers. He argues that legacy software transformation is now about becoming selectable by AI, not humans. AI agents will increasingly retrieve data, interpret causal relationships within enterprise information, and execute tasks end-to-end, so ERP vendors that remain tied to screens and menus risk losing relevance as interaction shifts to conversational and autonomous agents.
AI Agents as the New Enterprise Interface
As AI agents enterprise adoption accelerates, the primary interface for work is moving from dashboards to dialog. Ho Woong-ki notes that the industry has advanced from search utilities and generic conversational tools to agents that can autonomously retrieve data, analyze context, and complete business processes without direct human clicks. This is reshaping expectations for ERP software evolution: instead of designing features around user interfaces, vendors must expose clean, reliable APIs, ontologies, and business rules that agents can understand and chain together. Ho stresses that “having high-quality data is important, but utilizing that data is even more critical,” highlighting the need for frameworks that encode business context and causality so AI does not misinterpret enterprise records. In an AI-native world, the winning ERP is the one that agents prefer because it is easier to call, more consistent, and better structured for autonomous execution.
The $89 Billion Mid-Market ERP Gap and AI-Native Challengers
While large enterprises wrestle with complex legacy suites, an $89 billion mid-market ERP gap has opened among companies that have outgrown small-business tools but are not ready for traditional enterprise deployments. Intuit Enterprise Suite targets this space as an AI-native platform for businesses in the USD 10 million–USD 100 million (approx. RM46 million–RM460 million) range, offering enterprise-grade depth without the heavy implementation burden. According to Intuit, its Enterprise Suite runs at roughly USD 12,000 (approx. RM55,200) per year versus USD 80,000 (approx. RM368,000) or more for legacy ERPs, with over 90% of customers live within 30 days. By embedding AI agents on its GenOS platform to act on financial issues rather than only flag them, Intuit positions itself as a “dark” operational layer that mid-market AI agents can tap quickly, compressing sales cycles and lowering total cost of ownership.
Reinventing Legacy ERP as Agent-Enablement Platforms
To survive the rise of AI-native platforms, traditional ERP vendors must pivot from selling visible feature lists to providing reliable agent-enablement infrastructure. This means exposing modular APIs, event streams, and ontology-aware schemas that AI agents enterprise deployments can stitch into tailored workflows. Instead of competing on breadth alone, vendors need to prioritize usability through seamless integration, short deployment cycles, and predictable behavior at the API level. Intuit Enterprise Suite’s focus on a single CFO stack and deep, vertical-specific capabilities in areas like construction shows how targeted, AI-first design can outcompete heavier suites in the mid-market ERP gap. Legacy providers that fail to adapt may end up as commoditized data stores behind more agile AI-native platforms. Those that succeed will accept their role as dark software: mostly invisible to users, but indispensable to the agents running modern enterprises.
