Agentic Commerce: 70% of Online Buys by 2028

Recent projections indicate that agentic commerce, where AI autonomously handles purchasing decisions, will influence over 70% of online transactions by 2028, a staggering shift from current human-driven models. This isn’t just about automated reordering. It’s a fundamental reimagining of how goods and services are acquired, driven by intelligent agents making choices on behalf of users. But what does this mean for the average consumer, and how do we prepare for a future where our digital assistants are also our purchasing agents?

The 70% Influence Mark: Delegating Trust and Control

That 70% figure, reported by a recent industry outlook from Gartner, isn’t just a number. It represents a deep delegation of trust. Consumers are increasingly comfortable allowing AI to manage tasks once considered deeply personal, including financial decisions. My experience in digital product development suggests this comfort stems from AI’s ability to process vast amounts of data and identify patterns that humans often miss, leading to theoretically “better” purchasing outcomes. For instance, an AI buyer could monitor supply chain fluctuations in real-time, predict price drops, and execute a purchase for household staples precisely when they are cheapest, factoring in delivery times and personal consumption rates. This level of granular optimization is simply beyond human capacity for everyday items.

The challenge, however, lies in establishing the initial parameters for these agents. How do you instill your unique brand preferences, ethical considerations (e.g., avoiding certain companies or products based on labor practices), and a nuanced understanding of “value” beyond just price? It requires a deliberate, iterative process of defining rules and providing feedback. I’ve seen early adopters struggle here, treating their AI as a static tool rather than a dynamic extension of their own decision-making. The 70% influence doesn’t mean AI takes over completely. It means AI is the primary filter and executor, with human oversight still essential, particularly in the setup and refinement phases.

The Rise of the AI Buyer: Beyond Simple Automation

The concept of an AI buyer goes far beyond the “subscribe and save” models of the past. These agents are designed to understand context, anticipate needs, and negotiate. Consider a scenario where your smart home system detects a declining stock of a specific air filter. Instead of merely reordering the exact same item, an advanced AI buyer might cross-reference current air quality reports, review alternative filter types offering better allergen protection, check for compatible models from different manufacturers, and even factor in your recent energy consumption data to suggest a more efficient HVAC component. This complex decision-making matrix is driven by sophisticated algorithms and access to real-time market data. According to a report on autonomous systems by the Institute of Electrical and Electronics Engineers (IEEE), the core innovation is the agent’s ability to act on its own initiative within predefined boundaries, making choices that were once exclusively human territory.

My professional take is that this demands a new kind of literacy from users: AI governance literacy. You’re not just buying a product. You’re programming a proxy. This involves setting not only budget constraints but also preference weights. Do you prioritize sustainability over speed? Local businesses over global conglomerates? These aren’t binary choices, and your AI buyer needs to understand the sliding scale of your values. Without this granular input, the AI might make “optimal” decisions that are technically efficient but deeply unsatisfying or even contrary to your personal ethos. The “set it and forget it” mentality is a recipe for buyer’s remorse when dealing with these sophisticated agents.

Data Privacy: The Invisible Cost of Convenience

A Pew Research Center study from 2023 revealed that 68% of Americans are concerned about how companies use their personal data. With agentic commerce, this concern intensifies. For an AI buyer to function effectively, it requires access to an unprecedented amount of personal information: financial records, purchase history, browsing habits, location data, and even biometric information if integrated with smart home devices. This data allows the AI to build a complete profile of your needs, preferences, and spending patterns. The convenience is undeniable, but the privacy implications are substantial. Each interaction, each purchase, each browsing session contributes to a continually refining digital twin that knows your consumption habits better than you do.

The conventional wisdom often suggests that users simply need to “read the terms and conditions.” I disagree. That’s a passive approach to an active threat. The reality is that the complexity of these terms, coupled with the dynamic nature of AI, makes genuine informed consent challenging. What’s needed is transparency from platforms about exactly what data their AI agents access, how it’s stored, and with whom it’s shared. More importantly, users need intuitive dashboards to monitor their agent’s data usage and clear mechanisms to revoke access or delete specific data points. The onus shouldn’t solely be on the user to decipher legal jargon. Platforms deploying agentic commerce solutions must build privacy-by-design into their core architecture, offering granular control over data sharing and usage, not just an all-or-nothing toggle.

The Feedback Loop: Refining Your Digital Purchasing Proxy

One of the most overlooked aspects of successfully integrating an AI buyer into your life is the continuous feedback loop. Early data from companies piloting agentic systems, such as the SAP Concur Labs, indicates that user satisfaction significantly correlates with the frequency and quality of feedback provided to the AI. It’s not enough to set initial preferences. You must actively engage with your agent’s decisions. Think of it like training a new employee: you give initial instructions, observe their work, provide corrections, and offer praise for good performance. An AI agent learns from these interactions. If it buys a product you dislike, explicitly tell it why. If it finds a deal you love, reinforce that positively. This iterative refinement is how the agent aligns itself with your evolving preferences and values.

My professional advice here is to schedule regular “check-ins” with your AI buyer. This might be a weekly review of its proposed purchases or a monthly audit of its spending patterns. Many platforms are developing interfaces that allow users to easily review agent activity, highlight discrepancies, and adjust parameters. This active participation is what differentiates successful adoption from frustrating experiences. Without it, you’re essentially letting a sophisticated algorithm make significant financial decisions without proper guidance, which, while efficient on paper, rarely translates to personal contentment. The notion that AI will simply “figure it out” ignores the inherent subjectivity in human purchasing behavior.

Disagreement: The Illusion of “Optimal” Purchasing

Conventional wisdom often champions AI for its ability to find the “optimal” purchase: the lowest price, the fastest delivery, the highest-rated product. I strongly disagree with the premise that such a singular “optimal” exists for all consumers, or even for a single consumer across all contexts. Human purchasing decisions are rarely purely rational. They are influenced by emotion, brand loyalty, impulse, ethical considerations, and even the simple joy of discovery. An AI agent, by its nature, is designed to maximize efficiency based on quantifiable metrics. It might consistently choose a generic brand over a beloved artisan product if the price difference is significant, even if the artisan product brings you more satisfaction. It might prioritize a faster delivery from a less ethical supplier over a slightly slower, more sustainable option, simply because “speed” is weighted higher in its algorithm.

The illusion of optimality can lead to a sterile, predictable consumption experience that strips away the very aspects of shopping many people enjoy. My perspective is that the true value of agentic commerce isn’t in finding a universal “best,” but in freeing up cognitive load for routine purchases while still allowing for human intervention and delight in more meaningful ones. Users need to understand that “optimal” in an AI context often means “most efficient according to its programming,” which may not align with their personal definition of a good purchase. We must teach our agents to understand our subjective “optimal,” even if it means sometimes paying a little more or waiting a little longer for something that brings genuine joy or aligns with deeper values. This nuance is the frontier of effective human-AI collaboration in commerce.

The move towards agentic commerce is undeniable, transforming how we interact with markets. Understanding how to delegate trust, govern AI decisions, and maintain privacy will be key to harnessing its power effectively and ensuring these intelligent agents truly serve our needs, not just market efficiencies. For businesses, focusing on AI supply chain integration and strong AI governance will be important for success in this evolving field.