by CPI Staff | Oct 12, 2026 | AI Agents, AI Governance & Risk Management, Blog, MCP Server
In this blog post Using A2A and MCP Together for Safer Business AI Systems we will look at why many business AI projects get stuck after the first promising pilot, and how two emerging standards can help you build AI systems that are safer, more useful, and easier to...
by CPI Staff | Oct 12, 2026 | AI, AI Agents, AI for Business & AI Strategy, AI Governance & Risk Management, AI Vendor Selection & Implementation, Anthropic, Blog, Claude
Anthropic’s release of Claude Opus 4.7 isn’t just another model refresh. For organisations running agentic workloads — the kind that chain tool calls, browse systems, write and execute code, and make decisions across long-running tasks — this release moves...
by CPI Staff | Oct 12, 2026 | AI Agents, AI for Business & AI Strategy, AI Governance & Risk Management, AI Vendor Selection & Implementation, Blog
OpenClaw has rapidly become the de facto open source framework for autonomous AI agents. Developers love it. The community is thriving. And it is fundamentally changing how organisations think about deploying always-on, self-evolving AI assistants. The problem is that...
by CPI Staff | Oct 12, 2026 | AI Agents, AI for Business & AI Strategy, AI Governance & Risk Management, AI Vendor Selection & Implementation, Anthropic, Blog, Claude
Anthropic just revealed the architecture behind Claude Managed Agents. For any organisation deploying AI agents in production, the engineering decisions they made carry real implications for governance, security, and vendor risk. Here is what Australian IT leaders...
by CPI Staff | Oct 12, 2026 | AI, AI Agents, Blog
Zero Trust is well understood for users. Verify identity, check device health, enforce least privilege, assume breach. Most mature IT organisations have some version of this in place. AI agents break that model. Not because Zero Trust principles are wrong — but...
by CPI Staff | Oct 12, 2026 | AI, AI Agents, Blog, Microsoft AI Foundry
Most Australian organisations that started building AI capabilities in the last two years are hitting the same wall. The proof of concept worked. The board approved the next phase. And now IT teams are drowning in questions nobody planned for. Where do the models run?...
