Breaking News

Marketplaces Powered By Data-Centric AI

Marketplaces Powered By Data-Centric AI

Gayatri Iyengar is Head of Engineering for Dasher Growth platforms And Last Mile Logistics at DoorDash.

The world of e-commerce and marketplaces is undergoing a remarkable transformation.

E-commerce platforms are described by their participants—buyers, sellers and the foundational fulfillment platform (a.k.a. the network itself that enables the exchange of products, information and services). These marketplaces have been innovative in their own right. They have evolved from connecting people and services to people and algorithms.

Marketplaces like Amazon connect the right buyers to the right sellers and products via algorithms. The next wave of evolution in these networks is connecting people to “personalized” content and serving as a co-pilot for their needs. I will present my thoughts on the next wave of AI-first marketplaces and the opportunities, challenges and innovations that will redefine the future.

The Rise Of AI In Marketplaces

Historically, marketplaces and matching algorithms focused on core functions, leaving limited room for AI integration. Many marketplaces underutilize the data they collect, do not have clean data engineering pipelines and try retrofitting AI into existing systems. The rise of AI and the importance of data is challenging old assumptions and sparking a new race to build AI-first marketplaces and networks.

Effective AI models require large, high-quality training datasets and expertise in AI, which not all e-commerce companies possess. Also, these marketplaces face day-to-day challenges and prioritize immediate features over AI investment. Some of this could also be due to the lack of information and the benefits that data-centric marketplaces can provide.

Consider a traditional vacation rental marketplace, where users search for rentals on marketplaces like Airbnb. In the current model, users input their preferences and sift through listings to find the ideal accommodation. The list of suggestions is generated via the algorithm.

However, I believe this model will very soon be replaced by the network understanding more about the consumer and going beyond what’s typically shared in search queries or profiles. It will delve into their travel history, specific preferences and even past experiences to truly understand their ideal vacation. This will then become a personalized vacation planner, recommending not only the perfect property but also suggesting local activities, dining options and even travel routes. On the other side, property owners can benefit from an AI assistant that understands their unique property features and helps optimize pricing and marketing strategies.

In this scenario, AI enhances the user experience by collecting nuanced data and connects both sides of the marketplace—travelers and property owners—by offering personalized services, making vacations more fulfilling and property management more efficient.

Data-Centric Marketplaces Advantages

Data-centric AI revolves around the idea that data is not just a resource but the very core of AI systems. Harnessing the power of data-centric AI in marketplaces offers a multitude of advantages. Let’s delve into the key advantages here.

• Personalization And Recommendation Engines: Data-centric AI allows marketplaces to create highly personalized experiences for users. These systems can analyze user behavior, preferences and historical data to suggest products or services that are not only relevant but also anticipate future needs. For instance, think of a marketplace that knows your fashion taste better than you do, offering suggestions that align with your unique style.

• Real-Time Optimization And Improved Efficiency: Matching algorithms are the brain of fulfillment, and by using data-centric AI, these algorithms continuously update and refine their recommendations in real time. For ride-sharing services like DoorDash, this means that the closest available dasher who meets the user’s criteria (such as vehicle type and rating) can be dispatched promptly. This real-time optimization ensures minimal waiting times and maximizes user satisfaction.

• Improved User Experience Via Co-Pilots: One of the most interesting new opportunities with AI is going to be “co-pilots.” Users receive more relevant and tailored recommendations, which increases the chances of successful transactions.

• Adaptability: These algorithms can adapt to changing circumstances and user behavior. If a user’s preferences evolve over time, the system can adjust its recommendations accordingly, ensuring that the user continues to receive relevant matches.

At its core, data-centric AI can revolutionize matchmaking, making it highly personalized and adaptable. However, as we embrace data-centric AI in marketplaces, we must address crucial challenges, including data quality, privacy, security, governance and bias. Yet, these challenges have sparked innovations.

• Federated Learning enables AI models to learn from decentralized data while preserving user privacy. It combines insights from various devices to improve AI without compromising privacy.

• Transfer Learning allows AI models to apply knowledge from one task to another, accelerating learning.

• Generative Adversarial Networks (GANs) create realistic content, such as images or videos, by learning from each other.

• AI Operations (AIOps) combines AI and machine learning to manage IT systems efficiently, predicting and resolving issues.

These challenges are also opportunities, propelling marketplaces toward efficiency, personalization and security.

The innovations in data-centric AI are reshaping marketplaces as we know them. As we move forward into this exciting era, it’s clear that data-centric AI is a fundamental paradigm shift that has the potential to redefine ecosystems of exchange, making them more efficient, personalized and secure than ever before.

Embracing this change while upholding the values of privacy and fairness is the path to unlocking the full potential of this remarkable fusion of technology and data in our marketplaces.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?