Some conversations you leave with a few notes. Kerri O’Neill left me with a phrase I can’t unsee, a question I now ask before hitting send, and a reading list I wasn’t expecting.
In short: On The Deep Dive, Kerri O’Neill, Chief People Officer at Ipsos UK and Ireland, explains cognitive surrender: the point where we stop questioning what AI hands us. Her case is to use AI to build people up rather than cut them, and to protect the thinking that’s yours.
Kerri O’Neill belongs to a group that barely existed a few years ago: the HR leader who is also steering her organization through the biggest technology shift it has ever faced. Those two jobs used to sit in different functions, run by different people who rarely even had lunch together. I’d been looking forward to this one all week, and it did not disappoint. Here are some of my main takeaways.
To introduce her properly: Kerri is Chief People Officer and Global AI Workforce Transformation Lead at Ipsos UK and Ireland, she sits on the CIPD’s AI Advisory Board, and through her seat on the board at Oxfam she’s currently working through what AI means for global inequality. I was, and remain, slightly in awe.
Her career started in insurance, and she draws a line from that world straight into how she thinks about transformation now. Insurance prices risk by taking a 50- or 100-year view and working backwards. She calls it future-back thinking: decide what you want to be true in a decade, then look hard at what this year’s decision does to it. In a world where short-termism feels more prevalent than ever, that’s a mindset we could all do with borrowing.
There’s a name for the thing most of us do and none of us mention. Cognitive surrender: the point where you stop interrogating what AI has handed you and simply pass it on.
Kerri notices it in herself when she’s tired, but her fix is almost annoyingly simple: ask AI to do the thing, then leave it alone until tomorrow. She’s right, it works. I’ve done it twice since and the mini reset brings you back round with a clearer mind.
What makes her worth listening to here is that her answer isn’t to use it less. It’s to remember what you’re bringing to it. Her line on this is the one I’ve quoted most since:
The biggest supercomputer on the planet is sitting behind your own two eyes.
Kerri O’Neill, Chief People Officer, Ipsos UK and Ireland
We come with 5 senses and a brain running something like a hundred billion neurons. AI has language, vision, pattern and a great deal of speed. It doesn’t have the rest of it. Her point, made considerably more warmly than I’m making it here, is that we’ve become oddly quick to sell ourselves short. People are not tasks. Anyone who does their job well already knows the difference.
We covered an awful lot of ground, and I’d encourage you to listen to the whole thing, but another takeaway I keep coming back to is the compromise organizations seem increasingly willing to make.
Faster, cheaper and leaner on one side. Better on the other. More often than not the first three are winning, and the bill for that tends to arrive later.
Kerri makes the case for augmenting rather than subtracting. She isn’t claiming every business can protect every role; economies rise and fall and she says so plainly. Her argument is that if cutting headcount is the entirety of your AI ambition, you’re missing the far bigger prize, and you can only bank that saving once. Reskilling is cheaper than rehiring, and there are already organizations quietly buying back the experience they let go of.
So the better question isn’t “what can we cut?” It’s “what would this free our people to do?” When she puts that to teams, the answers are rarely about output. People ask for more time with clients, with customers, with each other. Kerri’s optimistic reading is that AI might clear out enough administrative sludge to cause a genuine resurgence of human connection at work, which isn’t a sentence I expected to be writing about a technology rollout.
Learning to use AI properly takes effort, and people only make that effort if life looks better on the other side. Frame it as subtraction and you’re asking your teams to work hard at their own diminishment. Frame it as what it frees them to do and you’ve given them a reason to bother, which is the difference between a tool everyone has and a tool everyone actually uses.
If your only ambition for AI is to cut your workforce, you’re missing a huge trick.
Kerri O’Neill, Chief People Officer, Ipsos UK and Ireland
What I hadn’t planned for was quite how far we’d travel from HR: the knock-on effects for society, for developing economies, and for what work looks like for the generation coming up behind us (how many times do I have to tell you it’s a fascinating conversation? Go. And. Listen.)
The impact on early careers is one I keep coming back to. There are nearly a million young people in the UK not in work or education (ONS, 2026), and that number has every chance of growing as organizations hollow out the positions AI can now cover.
Kerri is clear this is a business problem as much as a government one, and Ipsos are acting on it. When a senior role falls vacant, they don’t always backfill at the same level. Sometimes that headcount goes into early-career hiring instead.
I asked what single habit we should all protect in a world of AI. I was expecting a system, something with steps.
Kerri said to work out where you get lost in your own thoughts, and then guard it. For her that’s reading, and increasingly poetry, including a poet she came across at Oxford who writes about the corporate world of work. She was specific that this isn’t about switching off. It’s about protecting the one place where the thinking is entirely yours, wherever that may be.
Where do you get lost in your own thoughts?
Kerri O’Neill, Chief People Officer, Ipsos UK and Ireland
I opened the episode by asking whether the difficult bit of AI is really the technology, or whether the difficult bit is us. By the end I was fairly convinced it’s us, and I think that’s the good news, because it’s the part we can actually do something about.
Kerri’s parting advice was characteristically practical: stay open-minded about what the technology could do for your organization, and get your hands dirty with it so you understand both what it’s capable of and where it falls over. She’s adamant you can’t learn this one theoretically.
Mine is smaller. Before you hit send or save, ask whether what you’re about to release is what you’d have produced if AI hadn’t been available to you.
I’ve had to stop and check myself more than once since this conversation. It takes about 10 seconds and it’s occasionally humbling, but I’ve yet to regret doing it.
Watch the full episode of The Deep Dive with Kerri O’Neill, and enter to win a copy of her new book.