Consumers are already shopping with AI. Winning agentic commerce takes more than a sales bot—it takes a system that knows the full relationship. Gladly CEO Charlie Besecker explains why.
Brands spent the last few years deploying AI to handle post-purchase support—deflect the return, close the ticket, clear the queue. That work made sense. There were tangible cost savings.
And yet, according to Forrester, global customer service (CX) scores have fallen every year since 2021, hitting their lowest point in nearly a decade. The AI everyone deployed didn’t make the customer experience better. It made it cheaper.
Meanwhile, the buying journey has shifted. In the first three months of 2026, AI traffic to U.S. retail sites rose 393% compared to a year earlier—39% of consumers now use AI for online shopping. They’re having those conversations with ChatGPT, Perplexity, or Gemini—not with the brand. And when they do land on a brand’s site, they find the same experience that was there a decade ago. Product grids. Filter sidebars. A chat widget in the corner that opens with, “Hi, how can I help you today?”
Consumers feel that mismatch immediately. If a brand can’t match that on its own surfaces, the conversation—and the decision—moves somewhere else.
The instinct will be to layer in a sales AI. A pre-purchase bot to intercept the question before the customer leaves.
That’s the wrong fix.
A sales AI that only sees the pre-purchase moment knows the session. That’s it. It doesn’t know what the customer bought last time, what they returned, or what they asked about six months ago. And customers can tell.
Real usefulness requires knowing the customer—not just the moment.
That context doesn’t live in the pre-sale. It lives across the entire relationship. Customers have never experienced their relationships with a brand as two separate things: a sales relationship and a service relationship. That fragmentation is a brand-side problem, a vestige of how organizations were built. AI is finally the thing that lets brands align to how customers actually think.
Pre-sale and post-sale aren’t separate problems solved by separate tools. They’re one journey. Retailers that figure that out will build on it. The ones that don’t will spend years patching together systems that were never meant to talk to each other.
What we announced at Gladly Connect Live
Three things: agentic commerce capabilities that extend the Gladly platform into the full customer journey—including where consumers are already shopping on ChatGPT, Perplexity, and Gemini—enhancements to the AI Performance Loop, and a new model for human and AI collaboration.
- Before the purchase. Gladly engages proactively, wherever the consumer is. On a brand’s own site, AI is no longer a widget in the corner. It’s woven through the experience itself—answering questions, making recommendations, guiding the consumer across pages, and carrying a conversation that spans the entire site. The website stops behaving like a static catalog. It starts behaving like a store, with AI as the knowledgeable associate on the floor.
- During discovery and consideration. Gladly helps consumers find what they’re looking for, even when they can’t quite articulate it: personalized product comparisons, specific answers to specific questions, or AI-summarized reviews that surface what shoppers like this one said—not a generic star rating.
- At the point of decision. Gladly handles the questions that kill conversion in the final moments: fit, shipping, and policies. It surfaces the next product or upgrade in a way that reads as a suggestion from someone who knows the customer. Not a recommendation engine—a conversation.
- After the purchase. This is the moment Gladly has always owned—now working in concert with everything that came before. A return handled with full context becomes an exchange. A subscription question becomes an upgrade. A cancellation becomes a save. Every post-purchase conversation is a chance to grow both the relationship and the revenue. Because the AI never lost the thread.
- Showing up where consumers are already looking. Many buying journeys now run through ChatGPT, Perplexity, Gemini, and other AI surfaces no brand controls. Gladly announced capabilities that let retailers show up in those conversations with their own information—product details, policies, content, availability—rather than whatever an external AI happens to surface on its own. Be present where consumers are looking, without losing control of the story.
We also announced the operational layer most vendors skip
AI running at consumer scale is not a set-and-forget product decision. Retailers that succeed with it treat it as a discipline—continuously training it, testing it, monitoring it, and understanding what’s happening inside millions of conversations. At Gladly Connect Live, Gladly announced enhancements across the full AI Performance Loop: autobuilding AI agents from a brand’s own internal documentation, testing changes at scale before they go live, automated monitoring that flags issues as they emerge, and an Insights Agent that lets teams ask plain-language questions of their conversation data and get answers instantly. AI that isn’t continuously improving is quietly getting worse.
And human and AI working together—the right way. Not every conversation should run on full automation. An expensive purchase. A high-emotion service moment. A judgment call that needs a person. Gladly announced enhancements to how team members and AI collaborate on a single conversation. Instead of AI stepping aside when it gets stuck, AI continues to run the conversation and brings a team member in at specific moments—an approval, an exception, a judgment call—then carries the conversation forward. Pure automation scales its mistakes. The loop between AI and your team is what actually gets better over time.
Why it matters who builds this
Three things separate AI that works from AI that doesn’t.
- Category depth. Generic AI won’t get retail right. Product nuance, brand voice, seasonal rhythms, return patterns, and the expectations of consumers in high-consideration categories—that takes years. Gladly has spent over a decade doing exactly that.
- An architecture anchored to the customer, not the ticket. Most platforms organize around sessions that lose context the moment they close. Every conversation on the Gladly platform is anchored to a person—purchase history, prior conversations, preferences, channel context, all of it available every time. Without that, every conversation starts from zero.
- CX teams in control. How the AI behaves, what it knows, when it brings a person in—that belongs with the people closest to the customer, not engineering. AI that lives in an engineering backlog drifts. AI that lives with the CX team improves.
At the end of any given conversation, a consumer doesn’t remember whether they were talking to a sales AI or a support AI. They remember whether the brand knew them.
That’s what Gladly is built for.
Learn more at gladly.ai.
Note: This content was created by Gladly.