I saw an announcement by a new company coming out of stealth called Traversal. They are a native AI application to assist SRE teams in identifying and resolving service issues. It got me thinking  about enterprise software. The future isn’t just about bolting AI agents onto legacy platforms, it’s about building AI-native applications from the ground up.

Vendors like SAP, Oracle, and Salesforce are starting to introduce AI assistants, but the real leap will come when the core of these systems is reimagined around AI decision-making engines. Think of the difference between cruise control and full self-driving. Call it Full-Self-Drive for the enterprise (I should copyright that).

AI-native systems won’t just support human decision-making, they’ll drive it by learning and adapting continuously. This shift will unlock massive productivity gains, reduce waste, and increase agility in ways traditional systems never could.

We’re just at the beginning, and I think the next wave of transformative enterprise software will be built by those who embrace this AI-first approach.

💡AI-Native Applications: The Next Evolution of Enterprise Software

There’s been a rush lately to add AI agents and copilots to existing enterprise systems, from CRM to ERP to analytics platforms. While these integrations offer real value, they also miss the broader opportunity: the future of enterprise software isn’t just AI-enhanced, it’s AI-native.

What do we mean by that?

An AI-native application isn’t just a legacy system with a chatbot on top. It’s software designed from the ground up with AI as the decision-making and operational engine. These applications use intelligent agents to observe, analyze, decide, and act – continuously improving over time.

Let’s take ERP systems as an example. Most of today’s platforms still rely on static configurations, predefined workflows, and manual oversight. An AI-native ERP would instead:

• Monitor real-time data across operations

• Continuously predict demand and refine inventory models and needs

• Automate procurement and production scheduling

• Optimize logistics, transportation planning, and provisioning

• Continuously learn and adapt based on outcomes

In other words, it wouldn’t just recommend decisions, it would make them — similar to how full self-driving software can handle an entire trip without human input.

In the future, AI-native systems won’t just optimize workflows. They will install, configure, and maintain themselves via networks of agents that communicate and coordinate autonomously. These agents will handle everything from API negotiation to infrastructure provisioning through machine-to-machine dialogue. Just as protocols like POP3 and SMTP standardized email, emerging context-sharing standards for agents (e.g., MCP, A2A) will define how they exchange memory, goals, and actions. This will form the backbone for a fully automated, interoperable software ecosystem.

Yes, companies like SAP and Oracle are embedding AI features, but the architectural core of these systems remains largely rule-based. To unlock the full potential of AI, we need to move beyond integrations and embrace a new software paradigm – one where AI doesn’t just augment workflows but orchestrates and optimizes them end-to-end.

This shift won’t happen overnight, but it’s already underway. We’ll start to see entirely new platforms emerge. They’ll be built AI-first and cloud-native, free from the constraints of legacy design. As these systems mature, they will offer superior agility, precision, and automation.

We believe this is where the real productivity gains lie. It's not just about reducing waste and staffing. It’s about using the power of AI to free teams from the complexity of low-value repetitive tasks, and together with AI, finding new and innovative ways to improve the business.

Companies that embrace this new paradigm will see true enterprise transformation.