Fix the Data, Not the AI

March 24, 2026 | The Zeaware Team

AI quality improves fastest by fixing data, not prompts. POCs can tolerate workarounds, but production requires structured data improvement with AI-assisted analysis, human review, and governed workflows to deliver consistent, trustworthy outcomes.

Prompt Repeat - a simple option with measurable impact

February 20, 2026 | Zeaware Engineering

Zeaware Avalon now supports Prompt Repeat, a simple, configurable technique that can improve accuracy and predictability for classification, extraction, and validation tasks, informed by recent research, with token usage trade-offs.

From Interoperability to Agency: A Zeaware View on Practical AI Sovereignty

February 20, 2026 | Tony Bain (Zeaware CEO)

AI sovereignty is exercised through system design. By abstracting models behind governed orchestration, interoperability becomes practical, enabling control, auditability, and real choice without dependence on any single AI provider.

Anthropic Claude Sonnet 4.6 Support in Zeaware Avalon

February 20, 2026 | Zeaware Engineering

Zeaware Avalon now supports Anthropic Claude Sonnet 4.6, providing a cost-efficient, high-performance model option for agent workflows, coding tasks, and long-context enterprise use cases.

Anthropic Claude Opus 4.6 Support in Zeaware Avalon

February 10, 2026 | Zeaware Engineering

Zeaware Avalon now supports Anthropic Claude Opus 4.6 as a model service, enabling governed use of advanced reasoning, long-context, and agentic workflows without application code changes.

Create a Legal Document Search Engine with Zeaware Avalon and Isaacus Kanon 2 Embedder

February 09, 2026 | The Zeaware Team

Zeaware Avalon now supports Isaacus Kanon 2 Embedder, enabling practical, high-quality legal search using Azure AI Search, hybrid retrieval, and Australian-built AI innovation.

Tool Exfiltration Attacks, GenAI, and Why Control Matters

January 05, 2026 | Zeaware Engineering

Enterprise AI risk isn’t about models acting autonomously - it’s about system design. Controlling tool access, validating execution, and enforcing least privilege reduces exfiltration risk and improves trust.

Objectives for 2026

January 02, 2026 | The Zeaware Team

2026 marks the shift from AI experimentation to trusted, scalable production, where governed agentic systems, measurable outcomes, and confidence in execution become the true competitive advantage.

Free eBook on Agentic Governance

December 17, 2025 | The Zeaware Team

Zeaware is offering a free AI governance eBook by Jesper Lowgren, helping leaders design trustworthy agentic systems with practical, real-world guidance, just in time for holiday planning and strategic reflection.

GPT-5.2 now supported in Zeaware Avalon

December 12, 2025 | The Zeaware Team

GPT-5.2 support in Zeaware Avalon unlocks advanced reasoning, long-context handling, richer multimodal workflows and improved agentic tool use, delivering smarter, more reliable enterprise AI automation

Zeaware Blog Posts