13 women on the mistakes they made starting their own business | ‘Lobster learning’

13 women on the mistakes they made starting their own business | 'Lobster learning' | Increase your work visibility

How to get strong career feedback from AI

A letter from editorial director Lisa Smosarski.

My obsession with AI has continued to grow over the past few weeks; I've planned holidays, analysed blood test results and tailored nutrition plans for "pescatarians who need more protein but don't have time to prep" (chickpeas, lentils and nuts are my pals here). At work, I've been experimenting too, analysing data, finessing strategies and immersing myself in deep research, but one of my favourite new uses is to ask AI to challenge my thinking or give feedback on my work.

I've written to you before about the importance of being open to feedback, as well as the bizarre accuracy of an AI PDR, but using AI to challenge my assumptions, improve my thinking and identify my blind spots is proving to be the biggest help yet. Why? Because AI will say to me what people are afraid to; despite its human-like desire to please. Think of it as polite candour.

Here are my four top tips for getting strong feedback from AI:

1) Personalise. Ask AI to imagine itself as someone you want feedback from and the style you want. I'm not thinking a specific person – although it would be fun to get some Gordon Ramsay-style musings – more the role of the person. Prompts like: "Act as my manager giving gentle feedback. "Imagine you are the recipient of my negotiation and want to challenge me". "Respond as a Financial Director who is pedantic about detail."

2) Be clear about what you want. Use prompts like:

" What questions would you ask me if I handed this work into you?"
"What flaws are there in my thinking?"
"Have I missed anything obvious?"
"Weigh up the Pros and Cons".

3) Include any briefs or instructions you've been given alongside your work and ask how well you have answered the problem. AI is great at identifying omissions or missed opportunities.

4) Challenge back. Hopefully, it goes without saying, you should always interrogate AI's sources and rationale. If it asserts that you have got something wrong, check links or calculations to see if you have the same findings. If you don't agree, debate back. Even if you don't change a thing in your work, by asking and working through the questions, you'll have a more robust understanding of your own work and will be more set up for success.

Learning like a lobster is 2026's career superpower, but what kind of lobster-like learner are you?

Learn like a lobster

I was lucky enough to read Sarah Ellis and Helen Tupper's new book, Learn Like A Lobster, before it came out, and was immediately hooked. Partly because it's packed full of genuinely useful advice for work and life beyond it, and partly because of all the information about lobsters.

Did you know that lobsters never stop growing? Every year or so, a lobster will get to a point where it recognises its shell is a tad tight and constraining. So rather than putting up with that discomfort, they crack open their shell and crawl out. When that shell breaks open, the lobster is a naked jelly-like thing and is thus vulnerable. But soon enough its shell grows back, and now it's bigger, better and stronger than it was before. Also, they eat the old shell in order to do this – resourceful, right? 

Lobsters are a great work role model for us all, in that case. They recognise when they've outgrown a job, a workplace, a mindset, an industry, and don't stay put in it any longer. Despite the vulnerability that comes from breaking free, lobsters do it anyway, safe in the knowledge that they can come back stronger.

We could all be a bit more lobster-like.