Global marketplaces such as Uber, Amazon, and Airbnb are collectively used by at least half a billion users every month and are some of the most successful business models of the past two decades. They are also some of the earliest and most voracious adopters of Artificial Intelligence technologies.
The early adoption of AI could arguably have been one of the key factors they owe their success to. Today these global giants are leaning on these technologies to unlock even greater value from it than they did in the past.
The Focus On Predicting Customer Behaviour
Airbnb recently made headlines for using AI to reduce party house reservations. Naba Banerjee, is in charge of Airbnb’s worldwide ban on parties. She uses AI to identify high-risk reservations by analysing various factors such as reservation details, user demographics, location proximity to assess the risk of a party. Each reservation is then assigned a party risk score, which determines approval or rejection.
Airbnb is further enhancing its customer experience through AI, according to Airbnb’s CEO Brian Chesky, who claims “next May you are going to see a whole new Airbnb” including an AI augmented conversational search experience, and AI powered customer experience using chatbot.
Uber believes that every department should be driven by AI and democratised by the usage of machine language (ML) within the company through its ML-as-a-Service platform, Michelangelo, that enables teams to develop models at scale. In April, it applied for a patent on technology that predicts customer behaviour.
Uber’s latest ridematching algorithm forecasts customer’s potential ride requests by feeding context data found in Uber user profiles such as ride history or location that predicts users habit. This algorithm ascertains the confidence of a potential ride request and decides whether to match a driver to the ride to relocate to a nearby location to reduce pick-up wait times.
So if you regularly request Uber to go to work on Wednesday, Uber can anticipate that ride and have a driver ready to pick you up. Reduced wait times increase the efficiency of the app, earning potential of the driver, and improve the rider’s customer experience.
Amazon is increasingly using advanced machine learning algorithms to better predict which items customers in various parts of the country will want and when they will want them, and we work with our vendors and selling partners to store those products closer to customers. These efforts of “regionalisation” along with a recently announced AI-powered robotics system “Sequoia” will increase the percentage of same-day and next-day fulfilment.
Predictions And ChatGPT-Style Interactions
Beyond the use cases of predicting customer behaviour, according to recent job postings, Amazon is working on building a ChatGPT style search to its online store, expected to make product search “an interactive conversational experience that helps you find answers to product questions, perform product comparisons, receive personalised product suggestions, and so much more”.
While there is no launch date yet announced for the chat-style search experience, Amazon has already launched products that use generative AI to improve seller experience by improving listing creation and management experience.
These platform enhancements are a testament to increasing accuracy of marketplaces to predict customer behaviour and in turn make business decisions that affect wait times, warehouse stock of goods, pricing and much more. The higher the accuracy of prediction, the more effective these decisions are, and this in turn creates a network effect of improved efficiency at scale.
The pace of similar enhancements are expected to accelerate in the next few years as AI enhanced developer tools improve productivity and significantly reduce the timeline to productionalise effective models.
AI & Electricity: Economic Parallels
Andrew Ng, AI expert and co-founder of Coursera and Google Brain, says AI is the new electricity. He reiterates that prompting will revolutionise AI development from reducing time to build and deploy a commercial grade model from around 6 months to a few days.
So far the ability to develop large scale models that have high prediction accuracy was limited to the bigger tech giants, given the resources required to build and maintain these models. Today, with low code tools and prompt-based AI generation, this technology can now be availed by smaller businesses in any industry, thereby democratising its usage throughout the globe.
This is the paradigm shift that will allow smaller teams to customise AI solutions tailored to their specific needs using their internal unique data.
By 2030, PwC research estimates that global GDP is expected to gain approximately $16 trillion (a 14 percent increase) as a result of accelerating development and uptake of AI. This will be driven by productivity gains from businesses augmenting the labour force with AI tech, automating processes and increasing consumer demand due to personalisation and/or higher quality AI enhanced products.
Sneha Mathew is a Strategy and Operations expert in the tech industry leading Scaled Partnerships Strategy at Google Play. Views are personal, and do not represent the stance of this publication.
Discover the latest Business News, Sensex, and Nifty updates. Obtain Personal Finance insights, tax queries, and expert opinions on Moneycontrol or download the Moneycontrol App to stay updated!
