How agentic commerce is reshaping customer journeys
In July 2025, Barcode® and Shopify hosted a closed-door roundtable with senior technology executives from global commerce brands. The following insight offers a glimpse into the conversation we had about AI, infrastructure, and the future of commerce platforms.
Product discovery no longer begins with a keyword and ends with a click.
The entire customer journey is being rewritten by machines that think, browse, and buy for us, and it’s happening fast and often invisibly.
Today, more shopping journeys are being influenced by generative engines, smart assistants, and AI-powered tools that browse, recommend, and even transact on behalf of users. These tools offer suggestions, make decisions, and bypass traditional paths to purchase.
This marks the beginning of agentic commerce: a model where AI serves as the customer’s co-pilot (and sometimes, their representative).
“We’re moving toward a zero-click world. But here’s the truth: customers never wanted to click in the first place.
”Bridget Fahrland, VP of Applied AI at Barcode
As the landscape changes, most brands are still designing for an outdated reality: one where people search, land on a homepage, browse, filter, click, evaluate, and make a decision. But what if that entire loop is compressed to a single conversational moment?
What is agentic commerce, and why now?
Agentic commerce refers to shopping experiences where AI-powered agents or tools (ChatGPT, voice assistants, or embedded commerce copilots) act on behalf of users. These agents might:
- Browse your site
- Parse product details
- Compare prices
- Complete a transaction
And they might do all of this without the user ever laying eyes on your PDP.
These behaviors are already emerging. Consumers are using AI to summarize reviews, explore collections via natural language, and even create carts via chat.
This shift is being driven by:
- The rise of LLM-powered search experiences (e.g., Google SGE, Perplexity)
- Integration of commerce tools into AI platforms
- Consumer fatigue with clunky, multi-click buying flows
…But commerce isn’t designed for agents
Today’s ecommerce experiences were built for humans. You optimize for metrics like time on site, click-through rates, and even pages per session.
But agents don’t care about any of those things.
“We’ve been setting up experiences where there’s search and browse and click and click and click. But nobody liked that.
”Bridget Fahrland
An agent doesn’t want to admire your UI. It wants fast access to structured, accurate, and well-tagged data. And that means rethinking not just design, but how you prepare and present content under the hood.
What makes a site AI-browsable?
So what does it mean to design for agentic commerce? Here are some practical shifts brands can start exploring:
1. Structured product data
Ensure key information (dimensions, ingredients, shipping, stock) is accessible, consistent, and machine-readable. Schema markup is a must.
2. Semantic clarity over visual tricks
Avoid hiding vital information behind carousels, tabs, or image-only formats. AI tools can miss what a human might explore visually.
3. Modular frontends
Composable, headless architecture makes it easier to expose content to multiple surfaces, including AI agents.
4. Taxonomy & tags that make sense
Your internal categorization matters. Poor naming conventions and inconsistent tagging can confuse automated agents.
5. Experimentation infrastructure
You don’t need to rebuild from scratch. Test an AI-optimized PDP, create a chatbot plugin, or trial a guided shopping assistant in one vertical.
Don’t panic. Experiment
Agentic commerce doesn’t mean every customer disappears behind a chatbot. It means some of them will interact with your brand through new channels, and you need to be ready.
This is less about replatforming and more about creating optionality. About testing ways to serve both the human eye and the machine parser.
Start small. Run controlled pilots. Observe how agents behave. Learn what metadata they rely on. You don’t need to predict the future perfectly — you just need to be able to adapt as it unfolds.
“You don’t need to predict the future perfectly — you just need to be able to adapt as it unfolds.
”
Conclusion: Design for humans, prepare for machines
Agentic commerce isn’t a gimmick. It’s a natural progression of the digital shopping journey, and one that’s finally removing friction customers never wanted in the first place.
The brands that succeed in this next era will be those that serve both audiences:
- Humans who crave convenience, clarity, and care
- Machines that crave structure, speed, and precision
Design for both.
