In Part 1, we covered the basics of agentic commerce, including the risks and opportunities it brings about. Now armed with a more sophisticated understanding of what agentic commerce is, what are you supposed to do?
In this second half, we’ll cover four ways the customer journey is changing meaningfully. Understanding these changes can help you plan for the future and, specifically, decide where it makes the most sense to invest in agentic commerce. We’ll also provide immediate, actionable advice on data, governance, and learning to help your organization prepare for this fast-moving future.
New Implications for the Customer Journey
The following are some new implications that you should understand when evaluating how agentic commerce fits into your company’s strategy.
Awareness: Top of the Funnel Content & AEO/GEO
Using top-of-the (marketing) funnel content to attract new potential buyers to your products has been a long-used and often successful way to grow customers. Top-of-funnel content aims to attract an audience that ideally overlaps strongly with your buyer personas but isn’t necessarily ready to buy yet. Some brands produce thought leadership content, others entertainment, and others educational content. There are many content strategies to consider, but the throughline is that this content is not intended to convert consumers into customers (yet).
Pairing top-of-the-funnel content creation with search engine optimization (SEO), among other marketing channels, has been a peanut butter-and-jelly strategy for a long time. Top-of-the-funnel content tends to lend itself to long tail keywords, where winning the SEO contest is less of a gauntlet. Long form, information-rich content tends to be treated as high-authority by Google’s ranking algorithms.
Agentic commerce throws a bit of a wrench in that strategy, though, because:
- Segments of your customer base may be turning to ChatGPT (or other agentic channels) for learning or information, and…
- The way agentic channels crawl, catalog, and intent-match content for users is pretty different from how traditional search engines do so
The basic advice here is to implement a strategy for Answer Engine Optimization (AEO), also sometimes called Generative Engine Optimization (GEO). Why two names? The market hasn’t decided which one to use yet. AEO/GEO involves updating your content strategy to optimize for consumption by AI engines like ChatGPT and increase the likelihood of being cited as an authoritative source when a user’s query is answered in chat.
AEO/GEO isn’t a replacement for SEO. You still need that traditional SEO strategy in place, and some of those traditional SEO strategies apply going forward (e.g., creating quality content). Consider AEO/GEO an extension of what you’re already doing, but focused on the new technical realities of how AI agents consume content.
Consideration: Query Intent and Purchase Behavior
It’s been a pretty long time since Google gave you explicit analytics about which keywords drove which traffic to your website. But there are a multitude of strategies that helped brands get mostly there (e.g., creating landing pages highly targeted to certain keyword groups).
Despite its imperfection, there has always been a clear strategy to target keywords that match the user’s intent with the problem your products solve. Sell TVs? Target queries for information about TVs. The relationship was pretty direct.
Query intent matching becomes significantly more complex when you start to consider external agentic channels. The “query” that leads someone to your site isn’t just a brief 3-10 word phrase typed into a box anymore. It might be an entire conversation! And, that conversation may have weaved through all kinds of irrelevant topics.
Agentic channels are also a black box. Search engines like Google are mostly the same, but they at least provide some guidance on how their algorithms work. Likewise, many people are scientifically studying them to gain a better understanding. For AI engines like ChatGPT, it’s not a total guessing game, but it’s close to one. The platforms give you basically nothing to go on. (Remember this when implementing the above recommended AEO/GEO strategy.)
What this means is that you may start to build success by attracting new customers through an agentic channel like ChatGPT or Gemini, but how or why they found you through that channel is far less clear. The information is less available to you than in traditional search, and even if it were provided, it would be far more complex.
Purchase: New Modalities for Browsing & Shopping
It’s not just that agentic commerce opens up new channels for marketing or commerce with new rules. AI is arguably the biggest tech paradigm change since the Internet in the late 90’s, and with that, fundamentally new ways to browse and shop are emerging.
This is a space where everyone, from established enterprises to individual analysts, is still navigating the learning curve. “Everyone” includes your customer. They are fiddling and fumbling around with agentic AI while you, the merchant, are trying to figure out ways to leverage it for revenue or cost-cutting.
Even when you get past the nascency, agentic AI is a very different way of working with a computer. It’s not visual, point-and-click like most people are used to. That was the modality that eventually created the few common patterns we use in eCommerce web design today. Instead, agentic AI is highly textual and oriented around a conversation.
Think back to the first time you used an AI assistant like ChatGPT. You probably sat in front of the empty text box and thought, “Okay, what do I do here?” While many people are beyond that point now, we’re all, broadly speaking, still spiritually there. To prepare for this, you have to be ready to experiment and explore. You have to be ready for some initiatives to fail. You have to be ready to kill initiatives that aren’t working.
Quite frankly, you need your digital and tech teams to start operating more like startups than longstanding legacy IT teams. This introduces risk (which should be managed) and costs money (not all of which will net a return), but it’s the only way to navigate the shifting sands. Sound challenging? Now, consider how complicated it becomes when the commerce modality extends to agents in addition to people. Virtually no one is sending off AI agents to search and transact completely on their own, despite what you’ll see people claiming online. That future is probably coming fast, though. Now is the time to prepare.
Post-Purchase: Changes to Customer Support
Customer support is hard. You mostly scale customer support by hiring more people. Hiring more people eventually means hiring managers; more people who cost more. Because it’s largely people-based, it’s hard to build repeatable processes to ensure high quality. Furthermore, the people who are best at customer support are that way because they possess skills that make them attractive to higher-impact, higher-paying jobs. So, you’re basically always churning out your best people.
Agentic commerce changes customer support in three meaningful ways:
- It introduces the opportunity to weave AI agents into the customer support team, so scaling isn’t as human-intensive.
- It creates opportunities to build systems that allow everyone to operate like your best “natural” support person.
- It introduces new complexities that the support team needs to be ready to handle as customer queries come in.
The first two of these are about margin improvement and scaling internal operations. This is one of the earliest successful applications of agentic AI (not just for retail), because the ROI math is easy. The third requires you to update training and processes for customer support. Support people need to now understand that an AI agent may have played some (eventually all) of a role in purchasing a product. The AI agent has much of the information you’d traditionally ask your customer for; however, its ability to share that information remains to be determined.
Customer support must be aware of every point on the customer journey that incorporates agentic AI, even when it isn’t obvious to the customer. They need to be aware of what could go wrong, how to identify it, and what to do about it. This is a whole new category of exceptions to potentially manage.
The Path Forward: Data, Governance, & Learning
We’ve covered a lot of ground, but the central takeaway is this: the internet of the late 90s fundamentally changed how we browse and shop, and agentic AI is arguably the next paradigm shift of that magnitude.
Commerce is still about buyers and sellers, but the self-driving software entities (AI agents) are inserting themselves into every part of the customer journey, from discovery (AEO/GEO) to transaction to post-purchase support.
Ignoring this change is a high-risk strategy. Embracing it and accepting some risk is a necessity for merchants looking to improve margins and acquire high-intent buyers. This is an ambiguous and fast-moving space, but waiting for certainty is no longer an option.
As you begin to build your own strategy for agentic commerce, here are three immediate pieces of advice to prepare your organization for the coming shift:
- Get Your Data AI-Ready. AI is built on data. To leverage agents for direct experiences or to integrate effectively with external channels, your data must be clean, organized, and readily accessible. Start the process of getting your data in order now.
- Establish Agile Governance. The pace of AI adoption and its inherent risks demand a governance framework that isn’t rigid. You need to enable your teams to experiment quickly and to kill initiatives that aren’t working, all while responsibly managing the new risks. Governance must be an enabler of speed, not a source of arbitrary friction.
- Build a Structured Learning Culture. This space is moving faster than most are used to, even in tech. Make a conscious, structured effort to stay on top of what is happening across the various agentic channels and new modalities. Rapid learning is the only way to navigate the moving target that is the agentic customer experience.