SaaS Is Not Dead Yet

2026-06-01 · Source: AI & ML – Radar · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

Mike Loukides argues that the widespread belief in the demise of Software as a Service (SaaS) due to AI agents is premature. While agents like Claude, Gemini, or GPT allow individuals to quickly create custom software, such as personal CRM tools, this approach leads to significant challenges in team collaboration and data sharing. The article highlights that individual agent-built tools often result in disparate backends and schemas, making company-wide reporting and consistent metric tracking impossible. Traditional SaaS, like Salesforce, provides shared resources and bundled features that facilitate teamwork, which agentic programming currently lacks. For agentic skills to be truly useful in an enterprise, critical infrastructure for sharing, requirements definition, collaboration, testing, versioning, and security must be developed. The future of SaaS, therefore, lies in evolving to provide structured, machine-readable data via APIs for agents, rather than human-centric dashboards, positioning established SaaS providers as essential systems of record.

Key takeaway

For AI Product Managers evaluating agent-based solutions, recognize that individual agent-built tools create data silos and hinder collaboration. Your strategy should prioritize developing shared infrastructure for agentic skills, focusing on robust APIs that provide structured data for machine consumption, rather than human dashboards. You must also address critical needs like sharing, versioning, and security to ensure enterprise-wide utility and avoid being blindsided by slow adaptation.

Key insights

AI agents enable individual tool creation but lack enterprise-grade collaboration and data sharing capabilities inherent in traditional SaaS.

Principles

• Work is about groups and teams, not individuals.
• Enterprise software requires shared data and metrics.
• Value exists in bundled features beyond immediate needs.

In practice

• Design APIs for agents, not human dashboards.
• Provide structured state and relationship graphs.
• Offer agentic skills as a starting point.

Topics

• SaaS Evolution
• AI Agents
• Enterprise Collaboration
• Data Silos
• API Design
• Agentic Programming

Best for: VP of Engineering/Data, Executive, AI Architect, Director of AI/ML, AI Product Manager, CTO

Related on AIssential

• The Software Collapse & The New Stack — The agentic era shifts software from human-informed systems of record to autonomous systems of action, inverting traditional SaaS value.
• Trusting AI agents should mean putting them on rails, not letting them run free — Enterprise AI agent trust is low due to a lack of guardrails and shared context, not capability.
• Beyond the Chatbot: Practical Frameworks for Agentic Capabilities in SaaS — Integrating AI agents into SaaS requires balancing non-deterministic behavior with SaaS demands for predictability, cost control, and security.
• AI Agent Memory for SaaS: A Builder’s Guide to Context That Does Not Betray Users — Effective AI agent memory in SaaS requires separating context types and a structured architecture to ensure reliability and user trust.
• AI Inside the Enterprise — Enterprise AI adoption faces significant integration hurdles and architectural paralysis, requiring a shift to viewing AI as a new type of user.
• The “Computer” Clash: Perplexity vs DevRev on True AI Autonomy — True AI autonomy requires deep data integration and the ability to take action, not just provide answers.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI & ML – Radar.