INSIGHTS
Thinking worth reading
Frameworks, perspective and field notes on marketing and AI transformation.
AI literacy vs fluency: skills that don't expire
Search "AI literacy" and you'll find two kinds of content: tool tips with the shelf life of a supermarket avocado, or predictions about which jobs disappear next. Neither helps with an actual Tuesday afternoon decision.
At 21, Sarthak Dhawan was already using AI to build and ship a real product while still at university. Then one setting he'd switched on ran up $30,000 in a month. He knew how to use the tool. He didn't fully understand what it was doing while he used it.
Three tiers separate tool fluency from AI literacy: the foundations, using AI well day-to-day, and leading its use responsibly. Most people never find out which one their next decision actually needs until it's already gone wrong.
Recognise which tier your next decision actually needs — before it costs you $30,000.
Google, AI slop and the case for better content
AI can produce a polished article that says nothing. Fact-checking is necessary, and it won't tell you whether the piece is worth anyone's attention. Here's what does, and why it has to come from someone with experience.
AI governance has a Marketing problem
A 2022 Australian court fined a company $44.7 million for what its algorithm decided, not what a person wrote. Marketing meets that exposure before most other business functions do — and no new law is required for it to apply. Here's what already puts you on the hook.
Even your AI has 'wait, why did I come in here' moments
Your AI's memory works exactly like yours does — usually a shortcut, sometimes a liability. We watched an AI confidently tell us our filing system was broken. It wasn't. It had guessed, then reported the guess as fact. So when should you trust it?
When Finance asks what your AI spend bought
Everyone can quote the number. Almost nobody can explain what it bought. That gap — between spending on AI and proving what it delivered — is where the real problem lives. It's also the first thing Finance will ask about next quarter.
AI Safety for Small Business
Most AI privacy risk isn't hidden — it's a setting nobody's checked. Activate the free, built-in controls already inside ChatGPT, Claude, Microsoft 365 and Gemini, and your business information stays yours. No extra cost, no IT department needed.
It's not drift. It's spaghetti.
When AI output looks wrong, the usual suspect isn't the model — it's the mess feeding it: stale data, conflicting versions, rules nobody's cleaned out. Here's a framework for spotting which one you're dealing with, before you fix the wrong thing.
AI doesn't know when you can't afford to be wrong
Some AI mistakes cost you ten minutes. Others cost you a client or a claim. When you run the business yourself, there are fewer people between an AI error and its consequences. How to tell which kind of decision you're in before you act.
AI can make your decisions. It can't take your blame.
Air Canada's chatbot gave a customer wrong information. In tribunal, the airline argued the bot was its own legal entity — and lost. The real question for CMOs is whether a decision can be reconstructed and explained six months later.
Claude now watermarks your work. Can you explain who wrote it?
Anthropic's new watermark can tell you Claude was involved in a piece of text, or code. It can't tell you who's accountable for it. Most marketing organisations don't have an answer to that question yet. The gap matters more than the watermark does.
Turns out your AI's memory is exactly as reliable as your uncle's
AI memory is genuinely useful — until it fills gaps with assumption instead of fact. We hit this twice with a decision record: once a minor slip, once a costly one. Same fix both times — stop, check the record. Here's three practical ways to set that boundary.
The information on this website is general in nature and does not constitute professional advice. All rights reserved.
Other names and brands may be claimed as the property of others.
© 2026 The UnConsultants Pty Ltd
ABN 87 689 593 333