AI Will Make Bad Decisions Faster: Rory Sutherland

0:00So when you think of the traditional marketer, obviously, I'm sure you're getting peppered with questions about AI. How do you think that is going to change? It all depends on what we, we shouldn't, we should remember, we still have some degree of agency, and it all depends on what we ask the AI to optimize for. And my concern is that AI will fall into the hands of people in procurement, in finance, right? and they will ask it to optimize for short-term profitability. Now, I would argue that AI, when used speculatively in the world of... Right, okay, this all comes down to a philosophical distinction about what a business is for.

0:48And since about 1970, 80... No, since 1980-something, actually since 1990-something, the overwhelming dominant assumption about what is the purpose of a business is shareholder value creation. And therefore, the short-term pursuit, never mind the fact that it's incoherent, because it doesn't ask which shareholder over what time horizon, because everybody's taken that to mean the next financial quarter, okay? Generally, the way you will achieve those things will weight you disproportionately towards cost-cutting to a point which works at first, by the way, and then proves catastrophic. Because it actually...

1:31Okay, Octopus Energy is now bigger than Centrica, than British Gas, right? As an energy provider, I think that's true. What are they focused on? Customer service. All their ads talk about their customer service. What are they doing? They're doing radio ads. Brilliant, because radio is totally underexploited as a medium, okay? You know, what they will do is they will overestimate, over-optimise for efficiency and under-optimise for discovery. It's all to do with the explore-exploit trade-off. It's an algorithmic thing. And if you go back a bit earlier, before Milton Friedman came up with this idea of shareholder value, generally what that means is that the senior management of an organisation are create incentives that align them with the shareholder's own interests.

2:20Those incentives do not trickle down to anybody much below the C-suite. So you end up with a fundamental misalignment in a corporation where there's a group of people at the front, at the top, who are a kind of managerial Ponzi scheme, who have massive share options and things aligned with the shareholder interest. Meanwhile, the people who are trying to do a good job for the customers or look after the workforce are entirely at odds with the management. because nobody normal gets up in the morning and wants to enrich shareholders. You might want to serve customers or make sales or add value, but nobody's motivated by that.

2:53Nobody normal. So you create this autistocracy at the top. If you go back a bit earlier to a guy who won the Nobel Prize for Economics, called Herbert Simon. Herbert Simon had a completely different view of what a business was for. and his view was that profit was not an objective of a business. It was a necessary constraint that in order for a business to do its business, it had to make a profit to maintain its own existence and to justify its existence and to attract a reasonable degree of future investment. But the purpose of a business was the discovery of new sources of value. New sources of value mostly to the consumer, but you could equally argue new sources of value to the workforce because you know you would argue that if you can if you can find a way of rewarding your employees which doesn't involve money flexibility of work you've discovered a new source of value creation you've discovered a new value exchange so the purpose of an organization was to look at its resources and ask how it how its present and future resources could be deployed towards the creation of new forms of valuable exchange.

4:08Zach Thompson, do you ever get sick of calling kebab houses and nanas when you're trying to speak to qualified decision makers? I do. Well, I think I might have found the solution. 98 % correct qualified data, which means when you're cold calling, you're always going to be speaking to the right person. Do you know what the solution is? It can't be Prospio, can it? Prospio, let me spell it for you. P-R-O-S-P-E-O dot I-O. If you're looking for great qualified data, then you know what the solution is. Prospio.io. There'll be so many business owners now that are probably facing that dilemma of, do I become a freelancer?

4:50Do I outsource it somewhere else? Do I bring in AI? Do I hire? How, in today's economy, how would you approach that decision? I would do it so different now. It would almost be crazily different. And I've got a bit of a taster of that because at my agency, we would get 30 new hours of client work and then we'd hire another person and keep going. But now at CoachVox, we're a software company, but we're still creating social media assets for our customers. but it just takes a very advanced prompting system lots of different ai models and a kind of community manager to run it so we can generate in probably it's probably a tenth of the time what would have taken a team member like a long time and a lot of a lot of effort to do and we can just do it so much easier and it's actually good and there's actually a method to it and it's proven and it works so it's almost ridiculous how different that would be right now I think if I was starting an agency from scratch I would think of it as being me plus an army of robots and then I would only hire people when I knew exactly what I wanted them to get the robots to do you're in an army of robots that sounds like a sci-fi movie that's gonna yeah well we say we kind of say this so our coaches right now at CoachVox, we say all you need to run the coaching business of your dreams is you, your VA and your AI.

6:18And it's true. If you look at Peter Drucker, he said the only two valuable activities in a business are marketing and innovation. Everything else is a cost. That chimes completely with the Drucker. And without getting personal, you know Milton Friedman was weird right he he literally I mean you know I mean very very great writer very very interesting man very persuasive man interestingly he couldn't fascinatingly he had a very rare condition he couldn't tell the difference between music and random noise never heard of that before there's somebody else famous it might have been Che Guevara or somebody else or Simon Bolivar somewhere but apparently it's incredibly rare which always strikes me is a bit interesting.

7:07But anyway, I digress. But that strikes me. Okay, the quest for new forms of creative value exchange strikes me as something that everybody in business can get interested in and motivated by. The employees can find it motivating. R &D people, it fundamentally focuses you towards the future and also towards the customer. The shareholder value movement fundamentally will always focus you towards short-term operational efficiency improvements, which are fine as far as it goes, but that's not a five-year ambition, is it? The shareholder value movement has also been a gift to tech because tech comes in on the back of basically, we can get rid of your staff and replace them with some form of automated alternative.

7:59Now, I would argue that in the Drucker-Austrian Herbert Simon model of the world, your call center should be seen as a major source of interesting customer information and competitive advantage and a source of customer loyalty. In the modern vision of the world, it's something to be replaced with AI. Now, by the way, I think customer service can be massively amplified by AI, but that requires an opportunistic mindset, not an efficiency-driven mindset. Oh, God. I'm so tired. I've just been talking to qualified prospects all day without any wait time. How have you been doing it? Because I've been cold calling people on my mobile, and I reckon I've spoke to one person.

8:42No, no, no. Stop that right now and use Nooks.ai. Why? What does Nooks do? It finds me the right people to call, and it calls them, and it connects with them over and over and over again. So you're saying it's nooks.ai and I can just be speaking to qualified decision makers all the time. No other better connection, right? Wow, nooks.ai. So you say understand AI. My mum started using ChatGPT and she sent me a poem the other day about why my 18 month old son won't eat his peas. And she said this was all generated. So that's my mum's understanding of AI. Shout out mum. um what do you think i guess business owners entrepreneurs are fundamentally getting wrong when it comes to using ai i think that i think thinking in terms of prompting is very limited because that can only take you so far so i feel like the entry level is go on to chat gpt try some prompts out see what it can do but then after that it's like you have to think beyond prompts you have to think in the systems so imagine if you previously would have got a team member to follow an SOP and it's an SOP for writing a blog post say and it's like this is what you do to come up with the title this is what you do to come up with the copy meta description etc so now you're getting that team member to use AI to create that copy and because you're giving them a ban list you're teaching them how to use the tone of voice you're putting in the psychological profile of the dream customer but then don't stop there it's like that SOP is now not for the team member to follow that SOP is for the robot to follow so the next stage after that is to get the robot to follow it automatically rather than it having to be going through the team member so I'd almost be like think in think in the systems like think in AI it's almost like take the human out of it and think, how would I get this delivered to me in its complete and final form and then work backwards from there?

10:47And then you learn the systems and the software programs that you need to put in place to make that happen. What we've done, and the Labour Party has hardly helped in this by basically spreading a narrative of doom over the entire economy, is that if the the AI is in the hands of people who have an efficiency mindset, it will actually be catastrophically bad and awful. And those people, by dint of the power they have, by alignment with the dominant shareholder value paradigm, have most of the decision-making power. So actually AI will just make shit decisions faster. I was going to say, so if you've got short-term...

11:26Because it all depends where you start, right? I mean, if you take... This is totally gross in the over-sumplication. But if you take large language models as a very sophisticated form of predictive text, the thing we all know about predictive text, right, is if you start with the wrong letter, you'll never get the word you want. I literally searched on Google Maps for Abergavenny, but I typed the A as a Q. Now, to any normal human, it was pretty obvious. I was looking. No, no, it was coming up with places in fucking Iran. Right? OK? Right? Because if you get the first... I'm so sorry, I'm swearing.

11:58OK. But if you get the first letter wrong in that model, you'll never get what you want. If your foundational assumptions are wrong and you use AI, you'll just get to a stupid place faster. And the worst thing about it is everybody will think it makes perfect sense. Yeah. Yeah, I think that's...