How Amazon uses AI to drive sales, well everything

Amazon uses AI

Jeff Bezos, Amazon CEO, is the wealthiest person in the world. Was it DI (Divine Intervention) or how Amazon uses AI (Artificial Intelligence Intervention) to drive its sales to stratospheric levels?

Ya shouldnt a asked!

That Amazon uses AI to drive sales is no secret. What that AI is and does is a very well-kept secret. So much so that Jeff would have to kill you if he told you. It is the secret sauce of the behemoth’s business with the smiley face logo.

Let’s put it simply. Amazon uses AI on your data to decide what to sell. It then orders those goods (or decides to make a house brand). AI then decides which distribution warehouses should get what on what it thinks will sell and where. It then laser focuses advertising to you to sell those goods. And the rinse-and-repeat cycle begins.

It is fair that Amazon knows more about you than your mum? Amazon has gigabytes of highly personally identifiable information (PII) on you. It seems that users are all too happy to give that over in return for ‘perceived’ benefits. All that does is give it more ammunition to sell to you.


Bernie Brode continues his series on how Big Tech uses AI to dominate the world. Bernie is a US expert on AI. He has spent a lifetime delving into the inner workings of cryptography. Now the confluence of nanotechnology and artificial intelligence. He is eternally curious about where that ‘intersection’ will lead us.

BTW – his first article How Facebook uses AI to manipulate you shot to the top of our weekend reading charts. He writes:

How Amazon uses AI to drive sales, well everything

Let’s start with a simple explanation of AI. It is a computer algorithm to make decisions or draw conclusions that mimic what a human could do. Only a hell of a lot faster and with Everest’s of data.

Its basis is in machine learning – giving a computer a problem and basic instruction on how to solve it. By sheer processing speed, it arrives faster at the right answer. And its gets better as it processes more data.

A decade ago, Amazon had a problem. It was severely lagging behind Big Tech when it came to AI.

Why? It was fairly late to the AI party. Its roots as an online book and later general consumer goods retailer meant it was focusing on ERP. Enterprise resource planning means getting its back-of-house systems in order. Simply put if you order something the very process-driven ERP system should work flawlessly.

At that time, visionaries at Google, Facebook, Apple, and Microsoft were well ahead. They had grander visions of AI and what it could do for humanity. Three were focusing on AI and operating systems. Facebook – well it has always focused on extracting and selling your data to make money.

Bezos began to focus on how to use AI to extract more sales and turn customers into clients. Note the term clients infers an ongoing relationship – not a one-off encounter. And that is what happens when you go down the Amazon rabbit-hole. Amazon gradually learnt to love and trust AI’s answers.

Today the company is a world leader in deploying AI, machine learning, and deep learning in the consumer space.

Amazon AI is smarter than us – Houston we have a problem

Joe and Jane Average don’t sit over the breakfast table, contemplating how AI is influencing pretty well everything they do. They don’t equate why Jane’s new Facebook photo of puppy ‘Fluffy” starts an advertising tsunami for pet grooming and pet bling.

Joe, on the other hand, gets messages about pet care, insurance, pet food and meds. And their friends start getting subtle pressure to buy Fluffy a gift.

Well, Amazon, Facebook, Google and a whole lot more AIs talk to each other. They merrily swap titbits of information for a fee of course. If you go online, you can trust AIs to collate and use this information against you.

Amazon’s primary aim is to build a hugely comprehensive consumer profile on you. Why? So it’s AI can predict what you need before you know you need it. Buying extra information is simply a marketing cost.

Amazon uses AI
Gee, I did not know I needed that!

The ‘Averages’ are not even vaguely aware of Amazon’s AI ‘recommendation engine’. It is only the conspiracy theorists and tech-savvy types that know it is more ominous.

Enter the Flywheel

The Flywheel (known as the Virtuous Cycle Model) was a 2001 Bezos marketing strategy. Essentially it was to use customer experiences to drive more traffic to Amazon. It was also about cutting inefficiencies in the marketing/sales/order/distribution process.

But the task was too big for normal ERP process improvement. Amazon had to invent its own AI tools to get anywhere.

Since then, AI has infiltrated every aspect of its business. Why?

Simply put, Amazon ‘team/department leaders’ must use AI to transform their part of Amazon. To meet almost impossible goals for cost control and efficiency.

Read the caption – Thousands of employees focused on AI

AI became the grease for any squeaky wheel. It was for incredibly diverse uses such as products to stock, where to warehouse them, how to market, how to apply surge or opportunity charging, delivery, cost reduction and so much more.

Amazon has Independent ‘Dev’ teams – ‘AI Islands’ to develop the code, rather than use a monolithic, centralised structure. But it does not stop there.

Those AI Islands then connect to become the Amazon nation. These Islands collaborate to see where else their AI project could benefit the empire. Hence, the new term that Amazon execs call the ‘AI flywheel.’

A flywheel requires less and less energy to keep spinning

AI/machine-learning innovations in one part of the company fuel the efforts of other AI teams. In turn, these can build new products or services for the company or its clients.

For example, the Alexa voice platform started life as Amazon’s way to harness voice commands in its operation.

Alexa is not a person! She would like to be

The Danger in how Amazon uses AI

Prime Example: Ask a voice assistant how much manicures cost?

Google Assistant replies “On average, a basic manicure will cost you about $20. However, special types of manicures like acrylic, gel, shellac, and no-chip range from about $20 to $50 in price. I can recommend some local salons if you wish.”

Alexa replies, “The top search result for a manicure is Beurer Electric Manicure and Pedicure Kit. It’s $59 on Amazon. Want to buy it?”

Alexa is an integral part of Amazon’s AI ‘recommendation engine’ to sell more Amazon products. Before Alexa answers, it has already consulted your profile, knows where you live and what you can afford. It is essentially using your data against you to sell more.

This has made Amazon the world’s leading one-stop, online shopping destination and Jeff Bezos, the world’s richest man. Oh, and Alexa has ambitions to run your home, life, and everything.

It has also raised significant security concerns.

Alexa sits inside your home network as a trusted device. Security analysts say there is nothing to prevent the ‘monolithic and opportunistic thief’ (Amazon) from using our data against us.

Instead of AI helping us, it can shape our emotions, curate what we hear/see and exploit our desire for convenience. “Want to buy it?”

These concerns are more acute during the Covid-19′ lockdown’ as we shop online more. It has triggered renewed anti-trust concerns over Amazon’s almost total dominance of the online retail market. Note that Amazon pays few taxes in Australia

The Future

It is unimaginable that Amazon will lose its lead in AI research. In fact, is has monetised its AI expertise by offering off-the-shelf machine learning and AI tools to AWS developers.

These are a few of its CX (Customer Experience) AIs

The AWS platform offers easy ways for developers to take advantage of its house AI. Automated computer vision, recommendation engines, language translation and much more – some of the fruits of Amazon’s research. And its AWS platform is a dominant public cloud provider. It accounts for nearly 50% of this market.

Hmmm – I am sure this is part of the Bezos vision too. Amazon’s enviable lead in AI technologies (for its use), paid for by third-party users!. That’s a pretty sweet deal if you can find it.

Amazon uses AI

The Bottom Line

AI systems are the heart of Amazon’s enormous success. AI is in all aspects of Amazon’s operations, from profiling potential customers to automating shipment warehouses. And while the “flywheel” model has allowed this kind of propagation, it also comes with dangers.

Why? Because the ‘flywheel’ relies on your data passing between various parts of the company. That openness means your data finds itself in many different AI systems in completely unrelated parts of the firm.

The Amazon transformation – from an online book retailer to an AI-powered behemoth – is a marketing success story.

However, this success is also worrying. Amazon has a concerningly high share of the online retail market. Extending this dominance via AI can only lead to more acute privacy concerns and high-profile data leaks. 

I have yet to hear an astute politician or consumer say, “I like what Amazon is doing.”

As you Aussies like to say, “It would not pass the pub test”.

For consumers worried about their privacy – which is an increasing number of us – this is a real problem.

GadgetGuy’s take – Amazon uses AI to do it virtually better

Bernie is an AI expert. His challenge is to make you aware of how AI is fundamental to the success of FAANG. In every case, AI uses your data and extreme manipulation in the guise of helping you. In this case, it is about emptying your wallet.

Fundamentally it is about power and the application of AI to your data to screw you. Yes, screw you because the ‘Power Asymmetry’ has shifted. Remember when the customer was king? Well, we need to retake that position by restricting the use of our data!

Though few would claim AI hasn’t been a boon to Australian society in a variety of ways. A recent joint seminar hosted by Institute of Public Administration Australia (IPAA) and the Australian Council of Learned Academies (ACOLA) posted this question to Australia’s Chief Scientist, Dr Alan Finkel: “What will it take for you to trust artificial intelligence?”

The answer. Regulation with teeth to protect our privacy. You can read more about our take on Digital Privacy here.