A useful AI assistant needs more than a clever answer. It needs to understand the work: which information matters, what a good result looks like, and where your permission ends. For a creative business, that might mean preparing a client response without losing its voice. For an independent creator, it might mean turning an approvedContinue reading "Plugins and Skills: Giving AI Your Way of Working"

An AI system can look impressive in a demonstration. The harder test comes when it touches customer information, changes a business record or triggers an action someone must later explain. That is why a finding from EY deserves attention: 36% of executives surveyed said their organisation had experienced an AI incident or failure with aContinue reading "EY’s 36% AI incident finding: what Australian business leaders should take from it"

A useful personal AI agent should give us less to manage. It should remember the context we choose to share, help move everyday tasks forward, and recognise when a decision belongs to us. That is the promise behind Meta’s Muse, Grok Bot and OpenAI’s dots. Their public descriptions emphasise different aspects of the same ambition:Continue reading "Muse, Grok Bots and Dots: What We Want From a Personal AI Agent"

Explore a gallery of 853 Dots examples, its most-viewed examples and practical ways to try AI assistance for everyday work, projects and creative production.

How should we judge AI when a conversation becomes personal? OpenAI’s MentalHealthBench offers useful evidence, alongside important questions about trust, judgement and human agency.

Practical lessons from Holly Li’s Office Drama interview: start with useful tasks, set clear boundaries, and understand the difference between access, approval and retained information.

A clearer interface can help you examine an answer. A useful personal agent also needs a clear brief, appropriate access and an explicit point for your decision.

Start with one familiar task and a result you can check. These four prompt ideas show where an interactive answer could earn a place in your working day.

ChatGPT can now shape an answer into something you can explore. The useful question is whether the interface helps you understand the decision in front of you.

Practical agentic workflows that keep client information out

AI literacy is becoming another layer of the digital divide. Here is what separates useful, responsible AI use from guesswork, and four habits that help people close the gap.