Forget Use Cases: Where AI Value Really Comes From, with Jack Nugent

How this transcript was made

Transcribed by machine from the published audio, then corrected by hand. It is not a verbatim record: the changes below were made, and are counted.

This episode

72 speaker turns, 76 changes.

• opening voiceover replaced with the authored script (match 96%) and attributed to Max
• closing CTA identified as CTA 2 - Subscribe (v4) (match 100%), replaced with the authored script and attributed to Max

What was changed in these transcripts, and what was not

Generated: 2026-09-20 from 2026-09-20-named/, itself transcribed from the audio that is live on the feed. Every change below is mechanical and counted. No sentence was reworded, no claim altered, nothing removed for being awkward.

The editorial decision

Owner's choice, 2026-09-20: light clean. "Um", "uh" and stammered word repeats ("and, and") are removed, because nobody intends to say them. "You know", "I mean", "kind of" and "sort of" are KEPT - they run to roughly 2,300 instances across the catalogue and stripping them would make the page read as somebody else.

Counts across the whole set

Change	Count
stammered repeat removed	667
um / uh removed	654
space before punctuation	241
split contraction	132
guest turn named (Joe Casabona)	101
organization -> organisation	71
diarisation fragment returned to the host	62
ai -> AI	57
empty turn dropped (was filler only)	36
synthetic outro restored from script	21
organizations -> organisations	18
hyphen followed by space	14
realize -> realise	14
A.I. -> AI	10
recognized -> recognised	9
gray -> grey	9
summarize -> summarise	9
color -> colour	9
behavior -> behaviour	8
synthetic intro restored from script	8
analyze -> analyse	7
behavioral -> behavioural	7
recognize -> recognise	6
percent -> %	5
personalized -> personalised	5
skeptical -> sceptical	5
analyzing -> analysing	5
center -> centre	5
optimization -> optimisation	4
realized -> realised	4
favorite -> favourite	4
realizing -> realising	4
canceled -> cancelled	4
recognizing -> recognising	4
announcer turn relabelled from Guest	4
summarizing -> summarising	4
labor -> labour	4
behaviors -> behaviours	4
colors -> colours	4
organizational -> organisational	4
customized -> customised	3
analog -> analogue	3
optimize -> optimise	3
honor -> honour	2
optimizing -> optimising	2
summarizes -> summarises	2
personalize -> personalise	1
prioritize -> prioritise	1
prioritized -> prioritised	1
organize -> organise	1
rumor -> rumour	1
organized -> organised	1
capitalize -> capitalise	1
centered -> centred	1
organizes -> organises	1
emphasize -> emphasise	1
split name repaired (Schaefer)	1
humor -> humour	1
traveling -> travelling	1
defense -> defence	1
maximize -> maximise	1
Organizations -> Organisations	1
apologize -> apologise	1
digitized -> digitised	1
monetize -> monetise	1
analyzed -> analysed	1
narrator opener re-attributed to Max, words left as transcribed	1
theater -> theatre	1
specialized -> specialised	1
split name repaired (Wadsworth)	1
aluminum -> aluminium	1
specializes -> specialises	1
skeptic -> sceptic	1
customize -> customise	1
personalization -> personalisation	1
optimized -> optimised	1
visualize -> visualise	1
spoken domain -> humanspark.ai	0

What was deliberately left alone

• "Program". A computer program keeps that spelling in UK English; only a broadcast or a scheme is a "programme", and only a reader can tell which was meant.
• "License" and "practice". Both have a noun/verb spelling split in UK English that a rule cannot decide.
• Capitalised words mid-sentence that would otherwise have been respelled: 4 of them, most often part of a name. Left as spoken.
• Retired show and brand names. Spoken by the host and guests. A transcript that removes what was said is not a transcript.
• "Co-pilot". Sometimes Microsoft's product, sometimes the metaphor. A rule would corrupt the honest ones.

The synthetic segments

The opening voiceover on the archive episodes and the closing call to action are read by Max, the show's AI narrator, from a script written in advance. Transcribing them means guessing at words already on file, and the guesses were wrong in places - one opener came back as "an AI that boned him up" where the script says "phoned him up". Those passages carry the authored text instead, and are attributed to Max rather than to the guest: the diariser groups Max with the non-host speaker, which would otherwise put a real person's name on words they never said.

MaxYou're listening to Something About AI.

Jack Nugent's sharpest point in this one is about consultants: if somebody can hand you a use case, they can hand the same one to every competitor you have,

and within three months whatever edge you had is gone. He works in commodity and shipping markets, where the most valuable data a company will own in ten years is still sitting on paper.

One note: the book Alastair mentions partway through was published, as Use AI, Stay Human. Here's the conversation.

Alastair McDermottWrong AI Strategy. A lot of businesses are focusing on quick wins and cost cutting right now, but true innovators are completely reinventing their business model.

So today, I'm going to speak with a fantastic guest.

We're going to be talking about why the real power of AI is not in looking at efficiency, but in actually transforming how your whole business operates. Is AI a fad?

Jack NugentI mean, it depends who you ask today, to be honest. Look, I think there's a few things that, you know, would be quite helpful to kind of, you know, ascertain it to begin with.

I think, you know, AI is a tool, and it's one of many tools, and it might not be the right tool for your business. But it's the most powerful tool we have today.

So I think, you know, management, CEOs, CTOs, they owe it to the shareholders, they owe it to the board to make sure that,

you know, that AI definitely doesn't help their business, because it's such a powerful tool.

I think another way of looking at it is that we quite often have this kind of Northern European or US bias.

But on a global basis, we barely scratch the surface with AI, in terms of the data we can source, the problems we can solve, the people we can work with.

And I think when you look at data, and you think about what we currently have, you know, online and in our hands and everything else,

it's a tiny drop in the ocean from where we might see in five to 10 years from from now, you know, the most exciting data that a lot of these companies are going to have,

they're going to be things that they don't even think of as data today. And that's going to be the best, most valuable propriety data.

So, you know, with new data, suddenly, you can create new models and solve new problems. And I think that that's going to be an evolution that we're going to see.

You know, I think if you look at back back to the Daily Mail headline in 2000, where they said the internet was a fad, I get the feeling we're kind of at a similar point with AI.

Alastair McDermottThat's, that's interesting. And that's a great example. Because I remember seeing, I remember seeing pictures of that floating around a couple of years ago.

And it's like, it's shocking to think that people might think the internet was a fad when it's changed so much of what we do.

But now we're like, I see quite frequently, people saying things like AI is overhyped. And like, personally, I just don't get it.

Because there's so much, you know, there's been, there's, there's so much that it hasn't done yet. And yet what it has done is incredible. So

Jack NugentI think, I think, you know, when you look at what AI, you know, especially things like machine learning and everything else, what it is, in essence, it's prediction.

You know, there's a, there's a really good book, called prediction machines by three economists. And they look at everything that goes into that concept of prediction.

And every single business out there, you know, whether they're in sales or medicine, or whatever it might be, they are making predictions every single day,

then they might not even be aware of what type of predictions they're making, or why or how.

But that's really where the value with AI can come in is when you start breaking down the workflows in your company, and saying, Look, we're making prediction on this, on almost no data.

Okay, so where is the data? What data would be useful? So you know, when you start looking at that, you start thinking about how there are certain substitutes, which will lose value,

like human judgment, and there are complements, which will gain value.

So the the complements of game values data, and sorry, human judgment is compliment, a substitute would be human prediction.

And that's the thing, you have to get the two very clear, what is a prediction? And what is a judgment? And that's a very different thing.

And definitely within my industry, to give you an example from trading, you know, your prediction might be, is the market going to go up or down?

Okay, but your judgment might be, do I put a trade on today, on the basis of that prediction, and what size of trade and everything else.

So you have to be very clear with, you know, the sort of effectiveness of AI, and think of it as a predictive tool.

Alastair McDermottVery interesting. There's something that you mentioned earlier that I want to follow up on, and maybe this little feedback into it. You mentioned new data.

Can you tell me what you're thinking about there?

Jack NugentWell, I just think that whenever we look at, you know, models, like large language models, when we look at anything to do with AI today,

you know, we've seen a huge transformation of the data landscape in terms of actually being able to store it, be able to analyse it, be able to use it more effectively.

But I think there is still so much in the world, which is analogue.

You know, I think about a lot of shipping documents with cargoes, you know, a lot of things that happen, you know, with international supply chains,

supply chains that are still kind of like paper documents, they're not even, you know, digitised yet.

So when I think about where we are in terms of the data landscape, I think there's still a massive amount of potential to tap new sources and use that data in new ways.

So that's really what I mean in terms of, you know, if you think about how much of the data we have out there, it doesn't even cover anything in terms of most of the countries,

most of the companies, most of the economies that we kind of look at.

Alastair McDermottYeah, another example. And this is somebody who I might have on the podcast soon, I was talking with an AI consultant, and we were talking about water flow in,

you know, in pumping stations in kind of municipalities, where like, it's not super interesting to most people.

But if you're if you're trying to figure out particularly in places where you've got, you know, limited water supplies,

if you're able to figure out water flow by putting some sort of capture device, we're able to digitally monitor water flow and amounts and things like that.

And then you're able to monitor that over time. With AI, you can do some really interesting things with that and learn from that.

And so that's an example that I where I was thinking, okay, that's something that's very analogue right now. But we might have data about that in the future that could be useful for

Jack NugentSomebody. Yeah, and I think sometimes, you know, there's such an overlap in terms of what you think data will be useful for, and what it ends up being useful for,

you know, it's sort of two separate things, almost.

You know, a famous example is with a lot of hedge funds that started to look at satellite imagery over, you know, parking lots, to see how well the economy was doing.

And you know, which companies were making money in better sales and everything else.

And I know that there are some people who have used, you know, sentiment analysis with employees, companies and all this sort of stuff.

So you know, there are there is a lot of different ways of looking at data.

We've made huge strides in how we store that data in the cloud and with, you know, better data warehouses and sensors.

But we're still pretty much at the beginning, you know, you talk to a lot of people about their, you know, real time data sets,

and you're lucky if it goes further back than 2020 or 2019, which isn't that long, in terms of a data set.

So yeah, I think we're definitely at the sort of early stages of data collection, let alone actually building the models and solving the problems that we could be looking at.

Alastair McDermottSo your background is in creating value in the commodity and shipping industries. So I know you told me before that your role was to ask, how are we going to make money?

So how does this change how you think about how AI is used and how it can generate value?

Jack NugentYeah, I think that's a really good question. I mean, obviously, as an industry, you know, commodities and shipping can be pretty brutal.

You know, you constantly have to adapt, and you constantly have to think of new ways of making money and generating value.

So it's quite a good training ground for actually, you know, looking through new technologies, looking through new markets and saying, how is this actually going to generate value?

And I think the problem that we have is that, you know, everyone is asking about use cases, you know, they want to be shown use cases, but it kind of misses the really exciting point of AI.

You know, AI is asking questions about ourselves that we've never previously asked.

You know, if we talk about biases, if we talk about, you know, prejudice, we talk about things that are happening socially, you know, we're asking ourselves, why do we make decisions?

Are they fair? You know, are they accurate? Are they accurate? In a way that we never did before.

So I think when people look at use cases, it's, it's almost a bit of a kind of lazy approach, because you're hoping that someone will come along and show you,

okay, do this with AI, and that will make you money. And that's basically what it means. How will AI make you money today, or save you money today?

And these are metrics that we don't really apply to the vast majority of businesses or employees. You know, if you go to 95% of employees, and you say, how much money did you save last year?

Or how much money did you make last year? Most of them couldn't give you a figure.

And, you know, I speak to a lot of CEOs and board members, you know, and I say, okay, with your business, summarising in one sentence,

you know, how do you create value, which basically means how do you generate revenues? And a lot of them can't do it.

So you know, you take this kind of messy combination of a badly articulated corporate strategy with your company, confusing metrics for how you value sort of performance with your employees.

And then you throw in AI, and you say, Oh, it doesn't work. It doesn't add anything.

But you know, the first things first, you have to basically clarify, independent of AI, how do you plan to make money? You know, what is the value proposition of your company?

And once you've got the kind of corporate strategy clear, then you see if there's any predictive element in there. And if you have the capital to deploy to start developing AI tools.

But the first things first is how do you plan to make money, regardless of AI, you know, don't go into it thinking,

okay, well, I'm going to put AI, and that's going to make the money in itself, you need to have a different strategy first. And a lot of people kind of skip that stage.

Alastair McDermottI wonder if we could actually come up with some sort of hypothetical real scenario here, because it's a lot of these concepts are very abstract for for people listening to this or watching

this. I'm just wondering, you know, could you give me an example of what you're talking about here? Because I think that would be useful for people to kind of work through.

Jack NugentIn terms of what I mean, I mean, that basically is saying, you know, you know, okay, I can give you an example where it's where it's done badly.

And then you can see maybe how it could be done better. Okay.

There is a lot of SaaS companies that say, okay, we've got this great tool, okay, and we put a lot of investments as tool, and we're going to use AI to improve the tool.

And that's going to make the business fantastic.

And the problem is actually the the lifeblood of their business is actually nothing to do with the tool is to do with their sales, you know, and it's to do with actually creating revenue.

And it's kind of a bit naive to think that just because you have the best tool or the best platform, that you're therefore going to get the best market share.

Actually, you can have a very average product or average service and do phenomenally well, because your sales is fantastic.

So you know, I think that's the thing, like, I see a lot of people trying to target AI efforts on the product development side,

and almost none on the kind of sales side on thing that actually brings the revenue into the business. And that's what is very important to try and clarify.

But you know, to be honest, to be honest, Alastair, if they can even articulate in one sentence, and it makes sense how they plan to generate value and make make more revenue,

they're already vastly ahead of the vast majority of businesses out there. A lot of companies, they can't do it. They can't say exactly why.

Alastair McDermottThat's a pretty low bar, to be honest.

Jack NugentI mean, but it's been my experience. I mean, it's, it's, it's shocking, but it is true, you know, and people confuse, like I said, whether it's okay, well,

our product needs to be better, our tool needs better with actually sales needs to be better. They confuse what the lifeblood of their business is.

Or they just don't know, they just say, look, we're in this industry, and we are one of them.

And that that's the from that you're supposed to gauge how they're with what the value proposition is.

You know, let's say I'm an advertising, you know, company, I'm in the advertising industry, end of sentence, that's it, you're supposed to now understand how they generate value.

But but that's not really, you know, in like a trading environment, that wouldn't be good enough.

You haven't got a niche, you haven't got something that you are trying to monetise with an edge, you know, you basically need an edge. And you wouldn't be rewarded for that.

So what I'm trying to say is that, like, I see it a lot where people go, I want an AI strategy, but you haven't even got a business strategy, or a clear one.

So that really needs to be done that way around first, you know, see the real value from AI.

Alastair McDermottRight. So I mean, this, and I see this quite a lot, you know, implementing almost any new kind of tactic or even strategy in any business in any in any sphere,

usually requires that at some level, you need to go back to basics.

And so that's kind of what you're saying here is you need to go back to what's what's your basic overall strategy, your value proposition? What's your positioning?

And understand that before you start to try and implement AI?

Jack NugentWell, yeah, because I think that that will give you the best shot of actually implementing AI successfully.

I think if you can summarise in one sentence, what it is that you're doing to, you know, hopefully be your competitors at or what are you trying to make revenue from?

If you can then turbocharge that with AI, it focuses everything. Okay, what data do we need? You know, what sort of problem are we trying to solve?

Okay, well, that informs what what model you need. So it kind of everything stems from that. Whereas I see a lot of companies who haven't really kind of nailed that down.

And then they try to throw AI into that. And it doesn't it doesn't go anywhere. You know, it's, it's, it's, it's too difficult that way.

Alastair McDermottI think that's where they say, okay, let's look for the low hanging fruit first. And so they start to implement it kind of at that level.

Jack NugentYeah. And I always think that's a bit of a shame, because I think the low hanging fruit is basically sort of code word for removing it from the commercial angle of your business.

You know, I mean, you know, the best AI gains you're going to see is if you know, your best commercial people are trying to solve commercial problems.

I think if you just sort of palm it off to a different, you know, IT department within your company, or something like that, you know, you're not really going to get much value.

It's not going to generate that excitement of what what's possible and that sort of creativity of what data sets you could leverage,

you know, what problems you could solve, because you're just trying to go for efficiencies.

And I think also it depends kind of where in the business cycle you'll put, you know, your company is, you know, if you are a very big,

large company, and there's a lot of, you know, bloated costs, yeah, you probably will save a lot of money doing that.

But from the vast majority of companies who are still trying to generate revenue, still try to get market share, still try to grow,

you know, just to focusing on cost savings and efficiencies is not enough, it's not going to get you to where you want to be.

So I think, yeah, that's why going for the low hanging fruit can be a bit self defeating, to be honest.

Alastair McDermottSo the way I'm thinking about this now, and please correct me if I've got if I'm articulating this wrong, but I'm thinking of this as a top down approach and a bottom up approach,

the bottom up approach would be maybe how you describe, you know, picking the low hanging fruit and starting to make things more efficient, kind of across the

Jack NugentBoard.

Alastair McDermottAnd then the top down approach might be kind of looking at it as a more strategic, you know, here's our, here's our business model, here's our value proposition.

And where can we implement AI, you know, to go beyond the efficiency stuff and into kind of like the big picture of where this can help with,

you know, with decision making and things like that. Is that a fair enough way of describing what you're talking about?

Jack NugentYeah, I think so.

I think if you're looking for value, if you're looking for business value with AI, if someone can come in and tell you this is the use case, they can also tell all of your competitors.

And that means that you really haven't got any value, because it will just be gone with within a couple of months.

So I think, in terms of, you know, just looking at the efficiencies, looking at use cases, keeping everything simple. Yeah, that's great.

But it's not really going to change your business, it's not going to transform your business model.

And I think, you know, one of the other added advantages of the AI landscape is that there is this whole new ecosystem of sort of collaboration.

You know, you realise that that company over there is better at something than I am.

You know, I might have a traditional business model, it's very good, very successful, you know, we make a lot of money doing it.

But we should be innovating, and we should also be collaborating with people who are better at things than we are.

And, you know, I think that if you're just looking for low-hanging fruit, and you're just trying to do it that way,

it closes the door to all of that kind of exploration that you can be doing that will benefit your business and create value.

Alastair McDermottSo why not do both then?

Jack NugentYeah, I mean, look, there's an argument for doing both, for sure.

But I think that I see 95% of the sort of conversation out there is about the low-hanging fruit and the efficiencies and the cost savings.

And, you know, very little is about, okay, how do I actually use this creatively, you know? We used to be able to just do everything with numbers and machine learning.

Now we have this whole language component, you know, and every industry, every business is part numerical, part linguistic.

And we have this huge, great tool that we, you know, we don't really know what to do with it yet.

A bit like, you know, the internet in the 1990s, you know, we didn't know exactly what to do with it.

But that's how, if you can look at it in a new creative way and say, okay, where does language operate within our business? Where does it operate with our customers or counterparties?

And try to leverage it in that way. You will, you'll create more value. And I think as well, I said, like, you know, this is a difficult conversation. You know, this is much harder work.

I'm not sugarcoating it. I'm not saying, you know, follow these 10 steps and you make loads of money with AI.

What I'm saying is that if you are creative with looking at your business, irrespective of AI at first, and you find what your edge is,

and you turbocharge the edge with AI, that is the best way to generate value. It's not to hire someone to give you use cases or just to look for efficiencies.

I mean, like I said, if you're a huge company and you've got lots of inefficiencies, then yes, maybe there's more incentive.

But for a lot of companies that are looking to grow, then you really want to be trying to turbocharge your, the life blood of your business, basically. Right.

Alastair McDermottI was thinking of this in the context, because I was kind of thinking of this in the context of another conversation I had on the podcast with a lady called Joanna Pineda.

And she has a development company. They have 50 employees. And so what she said to her employees is, look, nobody's going to lose their job because of AI.

We may not make new hires or hires that we were intending to make, but nobody's going to lose their job. And I think that's a really, I think that's a really good

Jack NugentApproach.

Alastair McDermottAnd so I was thinking about this in the context of, well, first of all, in the context of, well, how do we prevent massive job losses?

Apart from creating value for anybody, I was just thinking of like the impacts on society of this.

But I started working out and I picked a hypothetical company with, I made up a hypothetical pharma biotech kind of health company with 200 employees where they laid off 80 % of the staff

because they were able to basically produce, with 20% of the staff, they could produce the same output, right? That's one scenario of the hypothetical.

The other scenario of the hypothetical is they keep all 200 staff and they arm all of those with AI and train them up and make them more AI literate.

And in that scenario, what happens with productivity? Well, obviously productivity goes up, but innovation goes up as well.

And so what you end up seeing if you, if you graph the two, and I must publish this as a blog post or something, but if you graph the two of those out,

the company that fired 80% of their staff do make savings and they continue on this very slow flat rise, you know, just true kind of natural,

you know, 2% per year kind of rise, like they would have done pre AI, whereas the company who keeps all the staff and arms them with AI,

their increase in value is much more kind of like hockey stick or something like that.

Because they're, they've got the same amount of people, they've got the same cost structure as they had before, but their income potential is now so much greater.

Their value potential is so much greater. So that's how I think about this. I'm thinking about it in terms of like, what are your choices as a business leader?

Well, one choice is we can save a lot of money by firing lots of people and replacing them with AI or another choice is how about we take those great people that we have and arm them with

these amazing tools. So I'm just wondering how you think about, how you think

Jack NugentAbout that. You know, I think again, we have to be very clear about, you know, where AI starts and where it stops, you know, like it's,

it's not necessarily a kind of tool which you could give to every, single employee in that, in that scenario. And it would just yield massive value.

But, you know, what is really interesting is when you start to think about the mechanics of, you know, prediction, you know,

machine learning, you know, genitive AI, you know, you need a data set and you need to be able to train it. Okay.

And if it doesn't have the training data, then it's not very good with its predictions.

Whereas humans, we're, we're, we're very good at trying to make assumptions or predictions with very little data. Okay.

So if you look at that example there, you know, you might have 20% of your staff, but as soon as the market changes,

as soon as a scientific, you know, advancement changes, as soon as something adapts in the economy, you won't have the people to respond to that.

You know, you would just be basically, you'd have trained everything on, you know, what you previously knew, what was previously true and things adapt too quickly.

So you will always need that human element. And I think there's kind of like two other ways of looking at it as well.

You know, there's a, there's a famous sort of advertising executive called Rory Sutherland, and he wrote about the doorman fallacy.

And, you know, he says that, you know, if you look at the, you know, a doorman outside of a hotel and he's opening and closing the door,

you might say, okay, well to save costs and to make it more efficient. You know, we fire the doorman and we install an open, you know, an automatic door opener.

But then you suddenly realised that he was actually bringing all this other values and added value to the, to the hotel that you didn't previously sort of comprehend.

You know, whether it's like security, customer recognition, hailing taxis, you know, even the prestige of the hotel,

you know, you might be able to charge more per, you know, per night, per room because you have a doorman, you know, just from the sort of prestige of it.

So this kind of doorman fallacy of like, you know, you're looking at costs and efficiencies and you miss the other value that's in the kind of,

you know, the different types of roles that that kind of person can play. So I think that's one element.

And then if, you know, if that hasn't convinced you and you think, no, I could fire everyone tomorrow and the business would still do really good, you know, really well.

You know, again, you want to think about these data sets. Okay. If you go and ask your employees how they do, what they do and why, it probably would be very bad.

You know, they won't really be able to explain what they do and why and how and everything else.

But if you monitor what they're doing and you learn from following them and you, and you can try to build a data set from how your company actually functions, that could be quite useful.

So, you know, within a five to 10 year period, you could have propriety data on how your business actually functions with your employees,

which will better inform your own processes, how you do, you know, what you do going forward. And also, it might be valuable to someone else for a completely different reason.

You know, like we say, hedge funds might say, oh, we're really interested in knowing about this workforce. Can we buy your data? And it's a data asset.

So, you know, again, like if you either look at it and go, we might miscalculate the actual value of our employees. That's one risk, the doorman fallacy.

But equally, you're missing an opportunity to build your own data set by actually bringing them into the kind of AI journey.

And, you know, there are some research papers which are discussing about how do you actually compensate these, you know, employees for contributing to AI models and data sets.

So, you know, I think that that for me, if I was to make the call, I'd go with that one because, you know, you're far better positioned to create more value across your business.

Yeah, absolutely.

Alastair McDermottAnd I spoke with, I think it was Justin Collery about this on a previous episode. And he called it tacit knowledge and, you know, the value of that is massive.

And I think, you know, so right now I'm just about to publish a book called the AI Powered Thought Leader, which is based in a major part on the previous iteration of this podcast.

We call the AI Powered Thought Leader rather than the AI Powered Business Leader. So I'm writing that book in a major way augmented by AI. So AI is helping me write the book.

But it's very much based on the conversations that I've had and having transcripts. And it's something I've noticed from podcasting for so long

So I have transcripts of approximately, I'd say, 270 podcast episodes, right?

Jack NugentYeah.

Alastair McDermottIt's a mountain of data. It's a mountain of knowledge.

And being able to data mine those transcripts and pull insights and pull information from those, I'm able to use it in so many different ways.

But having all of that is just so incredibly valuable. And it's only now that we're seeing AI tools with big enough context windows.

So they're only developing to the point where they can start to process that amount of data. But it's incredible what you can do when you've got a lot of information like that.

And this is where, like Justin was talking about, you know, starting to record every meeting that you have at work.

And it's just, you know, and he was talking about the concept of having the off the record and the on the record,

Jack NugentRight?

Alastair McDermottWe're going to hit record now where everybody, you know, so you have your small talk first and then you can go on the

Jack NugentRecord.

Alastair McDermottI actually wonder about that. And I think you just record the whole thing and just tell it to ignore all the small talk.

But then you get into kind of all sorts of like societal kind of social things of, well, is it fair to be recording everybody at work all the time?

Jack NugentWhat are we using with that?

Alastair McDermottWhat about the privacy implications? You know, court cases and what if somebody gets fired and then they say, I want access to all the transcripts or

Jack NugentSomething?

Alastair McDermottSo there's all sorts of messiness that comes with that. But in general, the concept of having these transcripts and having all this knowledge to feed to the machine.

It's incredible that like that information and that is where that's part of what we're talking about here.

That's part of what you miss if you decide to fire all your staff and replace them with ChatGPT. Is it going to miss out on all of that value, you know?

Jack NugentYeah, no, of course, you know. And, you know, like I say, these tools are still predictive, you know, it's basically predicting the next word and everything else.

And it's not concerned with the kind of causal logic and, you know, challenging sort of counterfactuals and everything else.

So, you know, there are definitely the role of the human is paramount.

And I think, again, if you look at the business value discussion, you know, we're having and you say, right, OK, our business value is AI. There is no value.

You know, the value doesn't come from the kind of the model itself. It comes from how unique the data set is and how you're uniquely solving a problem.

You know, the technique itself, you know, that you can get very good data scientists and machine learning engineers and everything else. And that's fantastic.

But you need the kind of unique data set within a kind of unique industry or niche. And you need to be solving the problem in a different way.

And without those, you're not going to be able to create the value.

So that's why, you know, going back to your previous question of like an example, you know, you need to sit with a domain expert in the industry that they're concerned with and really

understand how the dynamics work and see if they even have a business approach or, you know, a kind of strategy for generating value.

Which, you know, as I said, you say it sets the bar quite low, but it's funny how actually few do that, actually have a kind of articulated value proposition.

Alastair McDermottAnd very ironically, that is something that you can actually get AI to help you with quite easily. And I have a video online somewhere.

I did a live stream about a year ago now at this point where I actually used it to do just that live.

And I created value proposition and a full content strategy in, I think it was in around 17 minutes.

And that was from scratch because I actually did this live and I did it as a demo of the technology. And I was using the voice app. So I shared my screen and did that whole thing.

Jack NugentSo you can find that live.

Alastair McDermottI'll try and link it in the show notes. But this particular problem is something that you can get it to help you with.

Overall, I mean, yeah, we still need smart people who know their industries, who are in these leadership positions to figure out, okay, here's what the capabilities are of these things.

Here's how we could actually get an edge with it.

Jack NugentHere's how we can use it.

Alastair McDermottAnd that part, yeah, I think you're right. Like if it's easy, then it's like it's going back to the old thing of if it's easy, it's probably not worth as much.

You know, it's only the truly hard things to do that are truly valuable.

Jack NugentA hundred percent. A hundred percent. And, you know, the market is very good at rewarding those who can think differently and come up with new ideas.

And it's very good at punishing people who just try to reg urgitate what everyone else does. And that's always been the case and always will be the case.

So, yeah, I mean, for sure, if you're just hoping that someone tells you, okay, this is what to do with AI, you've missed the point.

The point is that you're going to have to do some hard work and look around your industry and say, okay, this is how things normally function. What if I challenge this?

And, you know, it takes kind of twofold approach as well. I think, you know, from your technical team, you're going to need people who are naturally curious.

People who kind of go, okay, look, I'm very good at Python. I'm very good at my data science techniques. I've got all those things.

But I'm also curious about other industries or other problems or other issues. So they're going to have to be curious enough to try and develop something that might be of value.

And then from the sort of domain side, the domain expert, they have to be comfortable with being told they're wrong. Again, this is another rare quality that not many of them have.

But, you know, when you become an expert, you have all of these kind of preconceptions of how things work, how they always will work. This is just, you know, what's done.

And basically, when you look at the data and you look at the model and it tells you something completely different, you have to be able to park all of that and follow what the model says.

And it's very difficult to do that. You know, a lot of people say, well, I'm not interested in doing that. You know, I know better than any data or model.

But pretty much every single industry, you know, whether you're looking at sports or medicine or education, whatever it is,

the data approach will yield better results than the kind of expert, you know, kind of gut instincts. But it's a difficult one.

So you kind of need both sides of the equation within a company, the technical side and the domain sides, to have this kind of flexible curiosity and be okay with being told they're wrong,

which is, yeah, not always very common.

Alastair McDermottThis is kind of the money ball approach, which I think a lot of people will not be comfortable with. You know, trust the data, not your good instinct.

Jack NugentYeah, I think human judgment is very important. You know, so like once you've got the prediction, is it the right risk appetite for your company? You know, is the team structure correct?

Do you have enough capital? Do you have, you know, enough risk appetite to act on those predictions or not? And that's completely okay.

You know, you can say, look, we haven't got A, B and C set up yet. So we can't really follow the model or listen to what it says.

But if you, the aim should be that you kind of get to a place where you can set that up as well. So compliment it. And you do have to use that human judgment.

I'm not for one minute saying that, you know, you just follow the data and you do nothing else. You know, the judgment is just as important, if not more important.

But, you know, the good thing about AI and the good thing about this data centric approach is that people are being far more clear about what decisions they're making within the workflow and

when and how. So they'll say, okay, we do this and that's a prediction, that's a judgment call. You know, you've really got to be so clear with where AI starts and where it finishes.

You can't just have this kind of, oh, it just permeates the whole business. It might be used by a lot of people in the business.

But, you know, what functions, what risk appetite, have you made the tough calls and decisions that the rest of your business needs to function if you just took AI out of it?

You know, that's kind of what you want to get to, that actually the whole business functions irrespective, which again is the tough part, you know, that's not the easy part.

Alastair McDermottI'm fascinated in the fact that your entire background is in shipping and commodity because they don't sound like industries that are particularly innovative.

Jack NugentYeah, I mean, look, it's a fantastic industry and, you know, the people in it are the hardest working, you know, most intelligent people I've ever come across in my life.

You know, they will always look at a way of finding a commercial solution to things. They'll look at ways of optimising logistics. You know, it's a very dynamic international industry.

You know, a lot of travel involved and different stakeholders. So, no, I mean, it is actually the most innovative industry from a value point of view.

You're constantly having to recreate value, constantly face new problems, adapt to new situations. And really try to thrive in that environment from a from a tech point of view.

I mean, there's been a lot of developments over the last few years.

It's far more kind of tools and applications out there, whether it's data focused, whether it's, you know, from an operations point of view with shipping and everything else.

And I think, you know, some people look at the industry and they say, oh, it's like dinosaurs. You know, they're still doing the same old thing and everything else.

There's this natural tension in commodities and shipping where actually you can make money from inefficiencies and from things not being transparent.

So, when that collides with the tech people who love transparency and love efficiency, it doesn't necessarily merge well together because basically you want to be a bit more efficient than

your competitors in quite an inefficient market. That's how you make the most money.

So that's why I think a lot of people, when they bring the sort of tech technical solutions, they get a bit per plexed. They say, well, why do people not want to use our tool?

It would, it makes things more transparent and more efficient.

Well, because there's a lot of money to be made from being inefficient and not being transparent, which is the commodity and shipping industry.

Alastair McDermottSo, that's fascinating.

Jack NugentOkay.

Alastair McDermottSo, I just want to try and wrap up the whole conversation and summarise because there's a lot of different things that we've covered.

So, I think that if we go back to the premise of use cases. Yeah. Why do you say forget use cases? Why do you say like use cases are not where you should

Jack NugentStart? Well, I think basically the reason why is because use case means how does AI make money today or save money today?

And I think that that is a very narrow metric for looking at anything in your business. And it's also a metric you don't apply to your employees.

And it's also a metric that you don't even apply to your own CEO to articulate exactly how the business is going to make money.

So, I think there's this, you know, business is far more fluid and far more, you know, ambiguous than like, you know, this is exactly the value we made by using this tool.

So, I think that's one element is the metrics. You know, you're trying to say, how does it make money? How does it save money today?

But you don't go and ask 95% of your employees, what did you make last year or what did you save last year?

And you don't force your board to completely articulate exactly how they're going to make money in the next five years. So, that's one element.

The other element is if someone can show you a use case and say, this will save you money, it's not an edge.

You haven't created any value in terms of, you know, doing something different to make more money or make more revenues.

You're just getting a service provider to show you how to do something and within three months, all of your competitors might be doing exactly the same.

So, it's not a kind of edge from that point of view. So, I just feel like use cases, it shuts down the complexity of the AI conversation.

You know, the AI conversation is, why do we make decisions? Are those decisions fair? Are they accurate? How are we making predictions?

You know, what value can we create from doing this in a different way? You know, and these are really exciting questions.

You know, we've never previously looked at, you know, let's take, you know, biases and everything else.

You know, before when people were hiring for jobs and they were being biased against one candidate over another, it wasn't even discussed. It wasn't even looked at or analysed.

But now with new data, with people looking at, you know, you know, models more accurately, they're saying, you know, were we being biased in that decision making?

You know, which is a fantastic question to ask. And I think that when it comes to business value, AI is asking us that. Are you clear of how you plan to make money before you use me?

Before you use AI, are you clear? And, you know, that's a really exciting question to have. And if you've got the courage to do it, your business will thrive.

So that's those are the sort of reasons why I think that use cases is a very kind of lazy term personally.

Alastair McDermottTo follow on from that, then in order to move beyond use cases and to look at this kind of top down, as I was I referred to it earlier.

If we're going to do it that way, does that not mean that your leadership need to be very AI literate and have a good understanding of what the current and even potential future capabilities

of these AI are?

Jack NugentFor sure. For sure. I mean, 100%. Because if anything, they just need to be very clear about what AI can do and what it can't do.

You know, again, I think there's this kind of trend where I think they use it as a buzzword. They want an AI strategy.

They just want to put AI everywhere, but they don't really know what it can and can't do or why.

So for sure, they either need to train people internally on AI fundamentals or they need to rely on outsourced, you know,

support staff or consultants, whatever it might be, to make sure that it's very clear what it can and can't do within their own business model.

That is absolutely vital to try and progress from here, because otherwise it will just lead to more people saying, well, we use AI in this part of the business, even though it doesn't work.

We can use it in this sphere, even though it doesn't work. And it will waste a lot of time and money.

But I think it is really important to have commercial people involved, because again, you know, if you go to your technical team and they are geniuses,

they are still very much, you know, numerical focused, you know, very much about that might be the efficiencies and the cost savings.

But they aren't tasked day in, day out, month in, month out with how are we going to make money as a business?

So you want the people that are asking themselves that question to also say, well, what about if you use AI here, you know,

and they will have to park some of their cynicism and think, okay, this is a long shot, maybe it won't work, but what if it does? And that's, it does need that commercial input.

I think if it just stays with people who are just sort of very much kind of technically focused, and they feel more comfortable in the kind of numerical sphere,

then that won't necessarily benefit the business in terms of creating more value.

Alastair McDermottAnd my concern here is that I see this kind of pushback against AI at the moment, where I see people putting up some sort of example of somebody using AI very badly.

And of course, it gives you like garbage in, garbage out, gives you some sort of terrible results. And then saying, hey, look at this, AI is terrible.

And my concern with that is that people might look at that and think, okay, that's the extent of the capabilities of

Jack NugentThese systems.

Alastair McDermottI don't need to look at this as a leader, as a business leader, I don't need to look at this. And I think that's very dangerous.

I don't think that anybody maybe consuming this podcast is going to be thinking like that.

Jack NugentI certainly hope not.

Alastair McDermottBut that's, that's my fear is that people could look at these things and think, okay, the capabilities of these systems are just not really all that great.

Jack NugentI think it's a real balance. It's kind of what we said before.

I think that you have to be honest that maybe AI doesn't really add anything to your business, you know, maybe it doesn't add any value. That's a possibility.

But it is the most powerful tool we currently have. So you do owe it to yourself and your shareholders and your board to make sure that definitely isn't the case.

It's too lazy to look at someone else messing around with it and saying, oh, it's pointless. You know, you owe it to your shareholders to make sure it definitely doesn't.

And, you know, going back to the Moneyball example and everything else, chances are, if you can be creative with the problem you're trying to solve,

and you've got a good data set that other people haven't got, you will create value. And those are the two kind of components really that will, you know, will create value.

So, yeah, I think it's definitely something that people have to consider.

Alastair McDermottWell, Jack, this has been fascinating. Thank you so much for coming on and chatting with you about

Jack NugentThis.

Alastair McDermottCan you just tell me briefly about Trativex and what you do

Jack Nugent?

Alastair McDermottYeah.

Jack NugentSo basically, Trativex is a consultancy firm which provides strategic advice and basically bridges the gap between technology and the commodity market.

So we basically have sort of two types of clients, one of the sort of traditional business models, you know, people like ship owners,

ship brokers, inspection companies, they've got the kind of traditional model, but they don't really know enough about technology like AI,

machine learning, and they don't have the manpower to really, you know, look into it on a deep way. So I work with them and help those guys out.

On the flip side as well, there are the technology companies who don't really have the in-house trading and chartering experience that we can leverage with Trativex.

So basically, they come and they get help with understanding exactly how people buy and sell cargos, how they charter ships,

how they trade futures, because that can then help them with their product development, pricing, you know, positioning, all that sort of stuff.

So yeah, so it's, it's, it's, it's exciting business, and it's growing fast. Awesome.

Alastair McDermottWell, Jack Nugent, thank you so much for taking the time and coming to chat about this with us today.

MaxThanks for listening to Something About AI.

If you are new here, there is a back catalogue worth your time and a lot more on the way, so hit subscribe wherever you are listening and you will not miss it.

And if you want to go deeper, Alastair keeps a growing library of free resources over at humanspark.ai.