The most interesting AI implementations

AI implementations in businesses are a crucial part of changes across various industries in 2023. The growing computing capabilities, widespread availability of data, and advancements in machine learning empower companies to introduce advanced AI solutions. These changes are reshaping how businesses operate by streamlining processes, cutting costs, and enhancing customer satisfaction.

AI implementations – table of contents:

In this article, we’ll take a closer look at some interesting AI implementations in businesses. From Stripe’s use of machine learning to detect fraud, to Swiggy’s personalized order experiences, to GitHub Copilot providing real-time support for developers. These examples showcase how artificial intelligence is fueling the creation of innovative services, transforming customer experiences, increasing conversion rates, and streamlining internal processes for businesses and institutions. Read on.

AI implementations in Stripe

Stripe (https://stripe.com/) is an advanced payment platform that enables businesses and institutions to process transactions both online and in traditional retail settings. It provides integrated solutions for payment management, invoicing, financial process automation, and the creation of subscription and loyalty programs. The application of modern technologies, including machine learning, allows Stripe to optimize conversions and minimize the risk of fraud. In 2023, Stripe is recognized as one of the most groundbreaking solutions in the field of online payment systems.

However, why is Stripe Radar one of the most interesting AI implementations in companies in 2023? Stripe Radar utilizes advanced AI techniques for fast and precise fraud detection, making it one of the most innovative solutions in the online payment industry this year. Its main advantages include:

Speed and accuracy. Radar Stripe evaluates over 1000 transaction details in less than 100 milliseconds, accurately blocking risky transactions. It achieves a precision where only 0.1% of valid payments are mistakenly rejected.
Advanced ML models. Stripe moved from basic machine learning to advanced neural networks, greatly improving the model’s performance.
Innovative architecture. The latest architecture has enabled faster model training and better scalability, allowing for quicker prototyping and implementation of new ideas.

Source: Stripe (https://stripe.com/)

„Complete the Look”, or the AI implementation from Walmart

Walmart, the American supermarket giant, offers a diverse range of products, including groceries, clothing, cosmetics, electronics, and more. As one of the leading retailers globally, Walmart employs over 2.3 million people worldwide. However, it has also been expanding its presence in the field of artificial intelligence recently.

Walmart’s newly launched “Complete the Look” (CTL) module is an innovative product recommendation system in the fashion and home decor category. But why is CTL considered one of the most interesting AI implementations in 2023?

Personalization of style. CTL generates comprehensive, stylized outfits around the product selected by the customer, making it easier to discover and choose suitable wardrobe pieces.
Increased confidence and conversion. Presenting personalized outfits enhances customers’ trust in their choices and encourages them to make purchases.
Time savings. The system lets you put together whole outfits quickly instead of searching for individual items, making shopping more efficient.
Advanced algorithms. CTL uses various algorithms, including appearance generation and coverage expansion, to scale and customize recommendations for users.

This innovation solves the problem of information overload and choice among a vast number of available options, offering customers an easy and enjoyable shopping experience that is both inspiring and stylistically coherent.

Source: Medium (https://medium.com/walmartglobaltech/personalized-complete-the-look-model-ea093aba0b73)

Uber — ETR Forecasting

Uber is a mobile app for booking car rides, but its AI innovations don’t stop at transportation. Take airport parking lots, for instance. With the introduction of ETR Forecasting (Estimated Time to Request), Uber has rolled out a cutting-edge system to predict waiting times for drivers at airports. Using sophisticated AI models, it anticipates demand and queue lengths, factoring in queue fluctuations and external elements like flight delays. This system provides drivers with info about expected waiting times, helping them manage their time more effectively and plan their locations better.

Why is this innovative solution noteworthy? Mainly because of:

Solved problem. Uber’s ETR Forecasting tackles the challenge of having too few or too many drivers at airports, impacting both passengers and drivers. Shortages mean passengers wait longer, and surpluses waste drivers’ time as they wait.
Innovation. The forecasting system informs drivers about the expected waiting time for requests, allowing them to better manage their time and positioning.
Application of AI. It uses advanced AI models to predict demand and queue lengths, considering queue dynamics and external factors such as flight delays.
Impact on the industry. It is one of the most interesting AI implementations in 2023 because it optimizes resource allocation in real-time, enhancing the efficiency and user experience of Uber services at airports.

Source: Uber (https://www.uber.com/en-GB/blog/demand-and-etr-forecasting-at-airports/)

AI implementations in Pinterest

Pinterest (https://pinterest.com/) probably doesn’t need an introduction. This American social media platform allows users to browse and share photos, GIFs, and videos covering various topics such as fashion, culinary, interior design, and more. Users can create their own boards with visual content they find interesting and explore other people’s boards for inspiration.

The platform relies on advertisements, and in 2023, it embraced AI to move from traditional reactive methods to more proactive ones in preventing advertisers from leaving. This stands out as one of the notable AI implementations in 2023 because:

It tackles the challenge of advertisers leaving the Pinterest platform. Traditionally, this issue was dealt with only after advertisers had already left, making it challenging to win them back. Thanks to Machine Learning (ML), it now allows for early detection of potential churn, empowering the team to take proactive measures.
he Pinterest team has created a Machine Learning (ML) model that predicts the likelihood of advertiser churn in the next 14 days. It uses a set of advertiser features to make this prediction. The sales team utilizes this information to prioritize actions aimed at preventing churn.
Preliminary experiments have shown that this approach can achieve a 24% reduction in churn in the test group compared to the control group. This indicates the effectiveness of a proactive approach in preventing churn.

Stitch Fix, or headlines and product descriptions generated by AI

Stitch Fix (https://www.stitchfix.com/) is an innovative platform that allows users to order clothing through a mobile app. Users can fill out a survey specifying their style preferences, size, and budget. Afterward, they receive five individually curated clothing suggestions from a pool of over 1000 different brands and styles.

Stitch Fix utilizes advanced artificial intelligence algorithms to generate attention-grabbing ad headlines and detailed product descriptions. This makes the process of creating marketing content and product descriptions less time-consuming and costly while ensuring uniqueness and consistency with the brand image.

The company employs the “expert-in-the-loop” method, combining AI creativity with human supervision, ensuring high quality and efficiency. With AI, Stitch Fix can generate product descriptions for hundreds of thousands of styles, addressing the scale and complexity challenges in e-commerce. Continuous improvement of algorithms alongside expert knowledge allows for a constant enhancement of the quality of generated content.

This innovation addresses the time-consuming and costly process of creating marketing content for e-commerce and product descriptions while ensuring uniqueness and alignment with the brand’s style.

Swiggy

Swiggy (https://www.swiggy.com/) is an Indian online food delivery service that allows users to order meals from local restaurants. The Swiggy app provides the option to choose dishes from favorite restaurants, track orders in real-time, and utilize additional features such as personalized recommendations and a set minimum order value.

Swiggy used AI to personalize food orders in their app, solving the “paradox of choice.” Here are the key points:

Paradox of choice. Swiggy noticed that customers struggle to decide what to order due to too many options. This phenomenon, known as the “paradox of choice,” leads to customer dissatisfaction.
Customizing orders. The company introduced a recommendation system for “baskets” with a limited number of dishes tailored to customer preferences to make their selection easier and enhance the user experience.
AI application. Swiggy utilizes AI to analyze historical order data, customer dietary preferences, and product seasonality to generate personalized recommendations.

This innovation tackles the problem of too many choices, enhancing both customer satisfaction and platform efficiency. It’s one of the most interesting uses of AI in businesses in 2023 because of its effectiveness and complexity.

AI implementations in FoodPanda

Foodpanda (https://www.foodpanda.my/) is an online food delivery service that allows users to order meals from local restaurants through the app. The Foodpanda app allows users to order their favorite dishes, track the status of their orders in real-time, and access various features, including personalized recommendations and a set minimum order value.

Foodpanda uses A/B testing to improve its menu and boost conversion rates. The innovation involves updating the B version of the menu through advanced scheduling and automation. Here are the key points:

Test automation. Using Apache Airflow to automate the menu update process.
Scalability. Optimizing the process enables faster testing, even across multiple countries.
Efficiency. Reducing the execution time from 9 hours to about 3.75 hours and lowering the error rate to 2.2%
Further improvements. Further improvements are planned, such as dynamic pagination configuration and separate DAGs for different countries.

Thanks to these innovations, Foodpanda addresses the issue of slow and ineffective menu updates, which is crucial for maintaining competitiveness and increasing user satisfaction.

Source: FoodPanda (https://www.foodpanda.my/)

Zillow

Zillow (https://www.zillow.com/) is an online real estate platform that facilitates searching, buying, renting, and selling properties for its users. The platform features millions of property listings, enabling users to compare values and connect with local industry experts. Zillow utilizes advanced technologies, including machine learning, to provide accurate property valuations and streamline the processes of buying, selling, or renting.

The innovative approach of Zillow to property valuation through the “Neural Zestimate” tool positions the company as a strong player among AI innovators. Here are key points that illustrate why “Neural Zestimate” is one of the most interesting AI implementations:

Quick response to market changes. Thanks to “Neural Zestimate,” Zillow can rapidly respond to shifts in the real estate market, providing up-to-date valuations on a national scale,
Simplified updates.The new system makes it much easier to update and maintain valuation models, improving their accuracy,
Local and seasonal trends .”Neural Zestimate” effectively incorporates local information and seasonal market changes into the learning process, enabling more accurate estimations of home values.
Valuation range. “Neural Zestimate” uses quantile regression to generate price ranges, providing a better understanding of the potential property value and reducing uncertainty in the estimation.

This innovation addresses the issue of the lack of current and accurate property valuations, which are crucial for both sellers and buyers in the dynamic real estate market.

GitHub Copilot with AI implementations

The list of innovative AI implementations couldn’t be complete without GitHub Copilot – an AI-powered coding tool that utilizes large language models (LLM) from OpenAI. GitHub Copilot is a breakthrough in code generation, enabling real-time code suggestions within the IDE environment.

Thanks to the collaboration with OpenAI, the creators of ChatGPT, and continuous improvements to LLM models, Copilot is becoming increasingly accurate and tailored to users’ needs. This tool enhances the productivity of programmers by automating parts of the coding process and providing instant suggestions.

GitHub (https://github.com/) plans to expand Copilot’s capabilities by adding voice support and integrating it with other elements of the platform. However, even now, it addresses the challenge of long and complex coding processes by offering intelligent suggestions that speed up work and assist in solving programming issues. This makes it one of the most interesting AI implementations in companies in 2023.

AI implementations – summary

The examples presented in this article are just the tip of the iceberg when it comes to AI innovations in modern business. In 2023, more companies are turning to AI technologies to work smarter, understand customers better, and stay current with industry trends. The increasing use of AI is bringing a fresh perspective, transforming the experiences of both employees and customers. According to Gartner, by 2025, 80% of businesses will be adopting at least one AI-based solution, signaling a positive trend for the business world.

If you like our content, join our busy bees community on Facebook, Twitter, LinkedIn, Instagram, YouTube, Pinterest, TikTok.

Author: Robert Whitney
JavaScript expert and instructor who coaches IT departments. His main goal is to up-level team productivity by teaching others how to effectively cooperate while coding.
View all posts