AI and productivity

In a world where we’re all striving to do more, AI promises to boost our productivity and supercharge our workflows. But how can we get the most out of it? That’s what we asked Dr Amantha Imber, who has a Doctorate in Organisational Psychology and helps businesses use AI to genuinely boost productivity.

Here, she shares practical tips on turning AI into a powerful ally for your day-to day work – and explains why we need to treat it less like a robot, and more like a smart but sometimes flaky colleague.

Name your AI

It might sound odd – or even a bit creepy – to try and humanise AI. But Dr Imber says giving your chosen AI chatbot a name (hers is Sunny) helps to remind us that it isn’t some infallible robot that always gets things right.

The best way to think about AI is that it’s like a really smart PhD level consultant mixed with a wacky, unreliable coworker.”

- Dr Amantha Imber - psychologist, speaker, author, AI Consultant of the Year 2025, founder - Inventium (Doctorate Organisational Psychology, 2002)

We know even the smartest humans have biases and limitations, and we need to remember that’s the same for AI, too. “If you give it a human name, you’re more likely to treat it like it is an imperfect human.” If you tell ChatGPT: “Remember your name is XXX” it will store it in its long-term memory. Stuck for a name? Try asking it to choose its own!

Tell it who’s in charge

That’s you, in case you were wondering. And being a successful boss requires specific qualities – specifically, delegation, says Dr Imber. “Good AI delegation happens when you give it a clear purpose, when it knows the outcome you’re looking for, when there is enough context. Giving it context is super, super important, far more important than good prompts.” So don’t just tell it what you’re looking for but also explain why. After all, it’s not a mind-reader – at least, not yet.

Let AI grill you

Again, it’s all about context, and the context here is you. Dr Imber suggests letting your AI bot ‘interview” you, using the following prompt: Run a short, adaptive interview that captures only what matters about my role, context, stakeholders, and how I work, then produce a single structured briefing that covers all information. Be efficient, inquisitive rather than prescriptive, and tailor the depth and questions to my role and signals. At the end, it will produce a briefing document that you can use at the start of any future AI chats where background information might be important. This approach can be useful in other scenarios too. For example, next time you have a problem, don’t ask AI to solve it for you – ask it to quiz you about it. You might discover the answers were there in your head all the time.

Start an argument

Added context will help you achieve more useful outcomes. But it doesn’t mean everything your AI says or recommends is going to be correct. Assuming that AI is always the smartest person in the room is a mistake, says Dr Imber, especially when your reputation – or your company’s – is at stake.

You always need to use your human judgment. Don’t accept the first answer from AI. You need to argue with it to get a better answer and to check bias.”

- Dr Amantha Imber

Thankfully, this doesn’t mean screaming at your computer. “For example, ask it: ‘What are the other ways that this could be done?’ Or: ‘Review your last response, perform a critical analysis and identify areas that are either incorrect or could be improved.’”

Talk to your AI

Most of us are used to just typing stuff into our computers: why should we start talking to them? “Science says that when we talk rather than type, it reduces the strain on our brain. Behaviourally, it’s much faster to talk, too,” says Dr Imber. But conversation mode works much better for specific instances, she adds: generating ideas, problem solving or exploratory thinking – playing the role of the colleague or friend you’re throwing ideas around with. When it comes to more nuts-and-bolts work, it makes sense to switch back to your keyboard. “Typing is better when you’re thinking through the logic of something, adding to work, editing or when you need that precision of semantics.”

Skip the free versions

Privacy and handing AI your data are big concerns. “There are two ways that you can use AI,” says Dr Imber. “You can use it to generate something from nothing, or you can use it to transform something into something else. If you’re using it to transform, that means you are giving it input.” This can include personal or confidential information and, if you’re relying on a free version of ChatGPT or similar, this information can then be used to train the AI model.

If you are not paying, you are the product and anything you upload will be used. It’s money well spent if you are using the AI well.”

- Dr Amantha Imber

Once you are paying, adjust the settings to ask it not to use your data for training.

Be open about using it

It’s tempting to just not mention your AI help. But Dr Imber advises against this. “I think it is rude to write a five-page strategy report with one basic prompt and send that to your teammates with no human judgment or critique.” A better approach would be to tell colleagues you used AI to produce a draft, and suggest you work on making it better together. Your workplace may also have specific policies in place for declaring AI use. (For example, Monash University has set guidelines for acknowledging the contribution of AI in research.) But there’s another reason to be open, too: knowledge. “We recommend all our clients share how they use AI so other people can learn from it and produce better and faster work. AI isn’t just a solo sport. It is absolutely a team endeavour.”