190. Competing in the Age of AI

--- 

We love hearing from our readers and listeners. So if you have questions about the content or working with us, just get in touch on [email protected]

Say hi to Sophia on Twitter and follow her on LinkedIn.

Following us on YouTube, Facebook, Instagram and TikTok will make you smarter. 

Episode Transcript

Hello smart people!

How are you today?

I have been spending an extraordinary amount of time with lawyers recently. No, I am not on trial for something scandalous (because I haven’t been caught!).

I have been approached by a couple of law firms who work with Big Tech clients, and they wanted to get some tech knowledge, to understand their clients better. I thought this was such a lovely and thoughtful thing to do, because in both cases, the law firms came to me - the partners reached out and asked me about workshops and learning programs.

And there is a lesson here for all professionals. The lawyers who reached out to me already work with the biggest tech clients, so they are already extremely extremely successful. And yet, they are still looking for ways to understand the sector and the clients they work with. They don’t have to, but this is literally what being proactive and going the extra mile looks like.

In my experience, the people who are the most successful are just more proactive and make more effort. Another example is that renowned business authors reach out to me themselves and pitch to go on this show. They don’t have to, but they go that extra mile.

So, where can you go that extra mile to reach your ambitions? It doesn’t necessarily mean just working longer hours. It often means just writing that email you’ve been putting off.

And also, for those of you who are lawyers, I’m teaching a free class called: Get Tech Clients: Introduction to tech for lawyers. It’s based on what I’m teaching to the big law firms I told you about. So, if you are a lawyer, or have lawyer friends, send it to them The link is in the show notes, or just go to techfornontechies.co/events

Today, we are going to cover lessons from a book called Competing in the Age of AI: Strategy and Leadership When Algorithms and Networks Run the World.

We are currently in a hype cycle around artificial intelligence, as you may have noticed. This does not mean that the technology itself is not going to be revolutionary, but there is a difference between the potential of a technology in general and the value of a specific company.

And, from what I am seeing, there are real tech bubble characteristics around AI companies in general.

So, I would like you, my dear smart listener to keep your heads in this AI whirlwind. Learn to differentiate real lasting innovation, that could have business and commercial impact, and between basically good PR.

This is why today, we are going back to basics and learning what AI actually is and how it changes business models, and this helps companies make a lot more money.

“Competing in the Age of AI” is my favourite book on how AI changes business. It was written by two Harvard Business School professors: Marco Iansiti and Karim Lakhani.

Professor Iansiti actually came on this podcast to talk about his research, so if you want a free HBS level lesson, listen to episode 9. The Business of AI with Harvard Business School Prof Marco Iansiti

Competing in the Age of AI came out in 2020, so way before ChatGPT became the most downloaded app of all time. BUT:

1. OpenAI was working on generative AI back then, so the fact that it’s new to you doesn’t mean that it’s actually new.
2. To understand how the current wave of AI will change business and bring investors billions of dollars, we need to see how AI is already changing companies from the inside.

The main premise of the book is that relatively simple AI has a massive impact, and even before the advent of ChatGPT on the public arena, it was changing industries from the inside. And I’ll share a couple of case studies from the book in today’s lesson.

Listen to this (I am quoting from the professors’ Harvard Business Review article):

“In 2019, just five years after the Ant Financial Services Group was launched, the number of consumers using its services passed the one billion mark.

Spun out of Alibaba, Ant Financial uses artificial intelligence and data from Alipay—its core mobile-payments platform—to run an extraordinary variety of businesses, including consumer lending, money market funds, wealth management, health insurance, credit-rating services, and even an online game that encourages people to reduce their carbon footprint.

The company serves more than 10 times as many customers as the largest U.S. banks—with less than one-tenth the number of employees.”

10 times more customers, with less than 1 tenth of the employees. So lots more revenue, much less cost. That is living the dream my friends.

Now let’s see how this happens.

We will begin by defining terms: weak AI and strong AI.

Weak AI does one task. For example, the Netflix content algorithm can suggest a movie you want to watch. It can’t also suggest that you should take the train instead of a taxi because there has been a traffic accident. 

It’s called Weak AI, but it has already changed the world. Weak AI, for example, also refers to the content algorithms in Facebook and TikTok, and these have literally changed societies.

Strong AI is a system capable of simulating human reasoning. It can be argued that generative AI is strong AI because it can create something new. Strong AI is the next generation of artificial intelligence, and we are now in the new next ten age.

But watch out - these are the definitions of AI as they stand today, in 2024. These definitions change, and they are not static, like the law of gravity. What was known was considered to be AI 30 years ago is not considered to be AI today.

And now that we’ve defined current tech terms, let’s move on to defining business terms. Yes, I know, this episode is going to be quite thick with knowledge and you may have to listen to it again.

Also, we now paste the episode transcripts in the show notes on techfornontechies.co. So, if you want to go over the concepts, the written format might help.

Ok: business! Basically, to run a business, you need two main components:

•  a business strategy, which defines what you are going to make and whom you are going to sell it to
•  An operating strategy, which defines how you are going to do this

So business strategy is what. Operating strategy is how. Got it?

For an enterprise of any size to be successful, you need both: an aim and a plan.

For example, if KFC decided to expand to the health and wellness market (God forbid), they could have a strategy to market to health conscious people, get some recipes that they plan to cook and have a vision of how they are now going to make even more money than they do today.

But, if they do not update their operating model this whole plan is going to be a total disaster. They would need to change their kitchens, have a new hiring and training process for chefs, because they now have new recipes, and KFC would need to create relationships with a whole new bunch of suppliers, who are going to sell them kale and quinoa, not just batter and chips.

There is a quote in the book that I like “Strategy, without a consistent operating model, is where the rubber meets the air.”

So, to sum up:

•   a business model is defined by how a company creates and captures value from its customers
•  An operating model delivers that value to customers: it is the plan for how to get it done.

AI has been changing how we get things done for decades already. The bit of the business world that it has been impacting is the operating model - how we get things done.

This is the core point I want you to get from this episode: the reason why Ant Financial can serve 10 times more customers with 1/10 of the staff is because they have a different operating model, which is powered by AI.

In the book, the authors share  a case study of Ocado, a British supermarket delivery company. I know this show has an international audience, so just think of Amazon, because there are similarities, and also, you can all imagine what an online supermarket is.

The business model for an online supermarket is that customers order groceries online and get deliveries to their homes. So there are no physical shops, and the shop front is a website or a mobile app.

Great, but we can agree that this isn’t exactly revolutionary anymore.

But what is revolutionary is how the business works from the inside: this is the operating model. It uses AI to make its supply chain incredibly efficient.

I’m going to read directly from the book here:

“The key to the business is Ocado’s centralised data platform, containing unrivalled detail on its products, customers, partners, supply chain and delivery environment. The data is accumulated in the cloud and is exposed through easy to use interfaces for use by agile teams deployed to optimise every kind of application, from delivery routing to robotics, and from fraud detection to spoilage prediction.

(Ok, that is a mouthful, but they use AI to tell when the milk will go off. Isn’t that cool?)

Ok, let’s continue: “All this has combined to build a rapidly growing and profitable operation with a record of 98.5% on time delivery.

AI algorithms are in the driver’s seat of Ocado’s operational execution. Running thousands of routing calculations per second, AI makes sure the company has a highly predictable delivery model, optimised across its fleet of thousands of trucks, delivering in all weather (which is notoriously terrible in the UK - aside from me) and traffic conditions across the entire United Kingdom. The algorithms optimise truck routing in real time and make sure the products delivered are fresh.”

So if you use AI in boring business processes within your organisation, the business can run better and at a much lower cost than the competition. So for the investors here: ask your portfolio companies how they are using AI in their operating models, and what their plans are in this area.

I know that there is all this rage about generative AI and it is really amazing. BUT, my dear smart people, the reason I am bringing you this book now and teaching you these concepts now, is that lots of companies aren’t even using Weak AI in their operating models. There is so much room for opportunity here, which is why I do think we are living in a very unique age, and it is very exciting!

Ok, now I’m going to teach you the last bit of this lesson. It’s going to be a bit full of information and terms, and if it goes over your head - do not worry. Listen once, absorb the general point, and if you need to, you can always go back and listen again to get the specifics.

In order to have an AI based operating model in a business, you need what the authors call an AI factory. I’ll show you how Netflix does it, because that’s something most of you are familiar with.

Netflix uses data and AI to decide which original content to make. Isn’t that amazing? The first time they did this was all the way back in 2013! Over 10 years ago, they used AI to see whether making The House of Cards was a good idea and how big the audience for the series would be.

Now let’s look at what do you need for this to happen. There are four components that make up an AI factory.

First, it’s data on your users and how they behave. So in the case of Netflix, it’s data on what kinds of genres people like to watch. Btw, I heard that codifying genres is really hard, which shows the importance of having good clean data, but that is another lesson.

The second component of an AI factory is algorithms: you need machine learning engineers to work with domain experts to create algorithms that use the data you collect. And there is plenty of room for non-technical people to participate here, so don’t just leave algorithm development to engineers alone.

You probably know about the first two components of the AI factory already. So let’s make it interesting and add the other two:

The third component of an AI factory is an experimentation platform. Basically, this is the bit where you test whether your algorithm is good or a load of rubbish. For example, you could feed in data about an existing series that Netflix has already made into the system and see what the algorithm predicts. If the prediction the algorithm makes is close to what actually happened, then it’s pretty good. If the prediction is totally off, then it’s not good.

I’m actually on the Advisory Board of an AI company that does this called Riviter. Riviter uses AI to predict consumer trends, like the colours that people will want to wear in 3 years time. They’ve been around for about 10 years, so they can actually prove that their predictions are correct. They can look at what people are buying from Zara today, and show the prediction they made 3 years ago, to tell how close their prediction is to what actually happened. And they get it right. It’s actually quite insane. I actually had a really good interview with the CEO of Riviter on this show called Advisory Boards: why join them & why have them. It’s episode 170 and I linked it in the show notes, or just scroll down a bit from this one.

And now, what’s the fourth component of the AI factory, I hear you cry? Well, wait no more. I shall tell you.

It is Software infrastructure - basically it means that the data you collect all needs to be in one bucket, as opposed to in lots of different buckets.

This sounds simple, but it’s actually where some of the most heated issues arise, because people get really protective about sharing data. Like marketing doesn’t want to share their full data set with finance, because they hate the finance people and don’t trust them.

This is an actual thing that happens, which I think is interesting because what stands in the way of the AI revolution is not technology, but just plain old human behaviour. We have our petty squabbles and our fiefdoms, and our human incentives, and it’s usually these human factors that are the hardest to work with, not the tech.

I do recommend that you read Competing in the Age of AI, but it is not a light bedtime read.

I’ve pasted the link to the book in the show notes. But, if that’s just not going to happen in your life, I have two ideas for you:

1. listen to my interview with Professor Iansiti. It’s episode 9: and it’s called The Business of AI with Harvard Business School Professor Marco Iansiti. Listening to that episode that’s the easiest thing to do, since you are already listening to this podcast.
2. And to go deeper, I cover the concepts from the book in the Tech for Business Leaders and Tech for Non-Technical Founders courses. You can get it on the website on demand at techfornontechies.co or you can get in touch if you want a tailored version for your organisation.

Well done for listening today. I know this was quite an information packed lesson today, with lots of definitions and terms. So maybe you need an ice cream after listening to this. I find ice cream to be a helpful aid to intellectual endeavours.

So well done for investing in your education and in your future today. I am really proud of you.

And if you found this lesson valuable, please do leave this show a rating and a review. It helps spread this work to other people, so more people can learn here, for free! And also, reading your reviews really does make my day.

So be lovely, and leave a rating and a review on Apple, or just a rating on Spotify.

Have a wonderful day, and I shall be back in your delightful smart ears next week.

Ciao!