Agentic AI is ushering in the biggest paradigm shift in commerce since the late-90s internet and eCommerce boom.
This article defines the new concept of “agentic commerce” and explores how self-driving software entities are inserting themselves into the retail customer journey.
We’ll dive into the critical opportunities and risks this presents for merchants and detail the new direct and channel agentic touchpoints emerging today. In a follow-up post, we’ll analyze how four key areas of the customer journey are fundamentally changing.
What is Agentic Commerce?
The term “agentic commerce” seemingly came out of nowhere. While its use is growing, that use is not totally consistent.
The term “agentic” is derived from the word “agency”, the ability to make decisions and act independently.
An “agent” is something or someone that has agency, so an “AI agent” is a software system that is given the agency to solve a problem without a human needing to explicitly guide it. It uses modern AI technology to do so.
The practice of building and deploying AI agents is “agentic AI”.
Agentic commerce is agentic AI applied to commerce. It means that between the merchant (retailer, brand, distributor, etc.) and the buyer, one or both parties have included AI agents to facilitate some or all of their side of the relationship.
The buyer may be using AI agents to help find and purchase goods, possibly autonomously. The seller may be using AI to facilitate that sale or provide post-purchase support, also possibly enabling AI to do so autonomously.
Conceptually, commerce hasn’t changed: it’s people or companies buying goods and services from other people or companies. Agentic commerce just adds new parties into the mix: self-driving software entities called AI agents.
Opportunities and Risks of Agentic Commerce
While the basics of commerce haven’t changed, the introduction of agentic AI into the process does create both new opportunities and new risks for a merchant.
Opportunities
Agentic commerce brings about the following opportunities for a merchant, each of which has the potential to be value generative:
- AI agents can improve margins by reducing mistakes, automating tasks that relieve the pressure to scale headcount, and streamlining clunky business processes.
- New strategies and new channels for commerce open a brief window to create first-mover advantage for merchants who are successful early.
- Those new AI channels also create a direct path to acquire high-intent buyers that potentially didn’t exist before. The new modality for shopping may actually create new shoppers.
- AI is consultative and conversational, so it opens possibilities to digitize commerce for premium-positioned products that traditionally required a high-touch sale.
Risks
On the other hand, agentic commerce brings about the following risks for a merchant, each of which has the potential to be value corrosive:
- By handing off some customer touchpoints to emerging third-party platforms (e.g., ChatGPT, Gemini), you risk losing complete ownership of the customer relationship.
- You also lose control of customer touchpoints, some of which could become important as consumers adopt AI for shopping. This introduces platform risk: the platform can change the rules at any time.
- Agentic AI is brand new. There are no well-worn best practices for the shopping experience, which means best-practice customer experience is a moving target.
- Whether introducing direct or channel AI agents into the mix, there are new costs to consider, especially if you are taking on inference costs (the cost of using an AI model).
New Agentic Touchpoints
Even with the above definition, “agentic commerce” is an abstract concept. In this section, we’ll discuss more concretely which agentic touchpoints are feasible. However, it’s important to keep in mind that this space is moving fast. What is state-of-the-art today may be obsolete or significantly changed tomorrow.
You may have noticed we categorize agentic touchpoints as “channel” or” direct”. We categorize this way because the risk-reward profile for each is a bit different.
Channel Agentic Touchpoints
Channel touchpoints happen on someone else’s platform. So, while they open new revenue opportunities, they cede control to a third party. Today, this mostly means extending your customer experience to agentic channels like AI assistant platforms or AI search platforms.
The agentic channels most worth considering today are:
- ChatGPT
- Gemini
- Google AI Mode (powered by Gemini)
- Microsoft Copilot
- Perplexity
Aside from taking general action for organic reach (mainly AEO/GEO strategy), you should also just keep an eye on the following AI players. They don’t have the same level of dedicated commerce focus today, but there are hints that any of them may get there soon:
- Anthropic Claude
- xAI & Grok
- Meta (especially WhatsApp)
You should also consider a near future where new kinds of channels emerge. For example, many believe agent-first social networks are in our future, especially after Moltbook’s viral weekend. Could this be a new revenue opportunity unlike any before it?
Merchants should consider each agentic channel individually, because they offer different levels of control, charge different fees, and offer different customer acquisition opportunities. At least for now, some of these large platform channels are being very deliberate about approving merchants to sell through their platforms. Factor into your plan that you may be on a very long waitlist.
Direct Agentic Touchpoints
Direct touchpoints leverage agentic AI on your owned web properties and applications. This gives you much more control, but it does come with cost and overhead to manage them.
The main way to build a direct agentic touchpoint is with a chat interface. Buyers are generally accustomed to the idea of chat, either with a live (human) support agent or with a traditional scripted bot.
The chat form factor works well for building an agentic user experience that ideally provides value more akin to a live support agent, with a cost structure more akin to a bot (i.e., you don’t have to staff people).
Consider platforms like Algolia, which enable you to build agentic direct search experiences. You can also use products, such as LangGraph, to create a more bespoke agentic experience. It’s more work, but it gives you maximum flexibility to try something creative.
As AI capabilities evolve, we expect new form factors and new design paradigms to emerge, expanding your direct touchpoint options from agentic chats. There’s a lot of room for innovation in this space.
Part 2 guides you through how agentic commerce changes the customer journey and what you and your business can do about it.