The Birth of Goal-Based Shopping

From search boxes to life goals - how agentic commerce rewires commerce

- The really exciting part of agentic commerce is that it enables entirely new shopping behaviors — not just better versions of old ones.

- Goal-based shopping is one of them: instead of searching for products, you describe a life goal ("plan my kid's birthday party") and the AI decomposes it into a complete basket.

- This shifts commerce from demand capture (customer knows what they want) to demand creation (agent surfaces things you didn't know you needed).

- It also blurs the line between planning and buying — whoever helps you plan, captures the transaction.

- The early battle is between Walmart/Amazon (control fulfillment, building AI) and ChatGPT/Google (control AI, building commerce). Who closes their gap first wins.

Every new technology shift brings two types of value. The first is doing old things better - the same thing you were already doing, just faster or cheaper. The second is enabling net-new behaviors - things that simply weren’t possible before.

Agentic commerce also delivers both - the “old things better” side is about better search, faster product comparison, less cognitive load when choosing between similar options. These translate into stronger commerce metrics - higher conversion rates, higher AOV, fewer returns. Walmart reported that customers using Sparky, their AI shopping assistant, have an average order value 35% higher than non-Sparky customers. That's a meaningful lift from what we think is essentially a better way to do what shoppers were already doing.

The net-new side is where it gets interesting. And the one we find most exciting is what we call goal-based shopping.

Products vs. goals

Today, the majority of commerce starts with a product in mind. You know what you need - or at least you think you do. You go to a retailer, type a keyword, get a list of results, compare, buy. There are obviously other entry points to commerce as well (e.g. ads), but the dominant form is still the search box. And the search box requires you to already know, at least vaguely, what you’re looking for.

But that’s not how people actually think. People don’t think in products - they think in goals:

• “I’m going camping this weekend”

• “My daughter turns 5 next Saturday”

• “I want to look like I stepped out of a magazine for this wedding”

• “We’re planning our honeymoon”

• “I need to make my skin look younger”

• “We’re hosting Thanksgiving for the first time”

In a search-based world, each of these goals requires the shopper to manually decompose the goal into individual product searches. Planning a camping trip means separately searching for a tent, sleeping bag, cooler, headlamp, bug spray, camp stove, and a dozen other things - assuming you even know what you need.

Goal-based shopping flips this. You state the goal, and the AI decomposes it into products.

The conversation is the commerce

What makes this work is the AI conversational layer. A search box gives you one shot - type a keyword, get results. A conversation gives you depth via rich descriptions and an interactive process.

Take the camping example - A good AI application/agent doesn’t just hear “camping” and dump a list of tents. It should ask:

• How many people?

• Where? (altitude and weather matter)

• How many nights?

• Driving in or hiking to the site?

• What gear do you already have?

By the end of a 60-second exchange, the agent has a far richer understanding of your intent than any keyword search could capture. And then it builds a complete kit - including things you forgot you’d need, like a headlamp or a first aid kit.

Now consider a birthday party:

“My daughter turns 5 next Saturday, 12 kids, outdoor, unicorn theme, budget $200.”

The agent can generate the full list: decorations, plates, cups, napkins, a piñata, favor bags, cake supplies, a banner, candles, games. Instead of a 12 separate product searches, you compress it to one conversation.

From demand capture to demand creation

This is the structural shift that matters most - traditional e-commerce is a demand capture machine. The customer already knows what they want, and the retailer’s job is to make sure they find it and buy it here instead of somewhere else.

Goal-based shopping is demand creation. The agent surfaces products the shopper didn’t know they needed - because they were thinking about a life event, not a product category. You can’t search for what you don’t know to search for.

Where planning ends and commerce begins

The other thing that’s quietly radical: goal-based shopping blurs the line between planning and buying. When someone says “I’m planning a ski trip,” that’s partly a commerce problem (gear, clothing, snacks for the drive) and partly an operations problem (lodging, lift tickets, route planning, weather check). Traditional retailers only see the commerce half. An agentic system can serve both - and the planning layer feeds the commerce layer naturally.

The agent becomes a planner first, a store second. And because it helped you plan, it earns the right to help you buy.

What we think will happen is that there will be more symbiosis between planning and commerce. The tools people use to plan their lives - trip planners, meal planners, event organizers, fitness apps - will increasingly have commerce built in. And the commerce tools people use to shop will increasingly understand the life context behind the purchase. The wall between “figuring out what I need” and “buying what I need” collapses. Whoever helps you plan captures the transaction.

What this means for shoppers and retailers

A few implications worth noting:

For shoppers: the barrier to complex purchases drops dramatically. Things that were previously too much work to plan properly - a themed party, a full home renovation shopping list, a week of meals with dietary constraints - become simple conversations.

For retailers: product data needs to be organized around use cases, not just categories. If an agent is going to recommend your camping stove for a weekend trip, your product metadata needs to say “works for car camping, boils water in 3 minutes, serves 2–4 people” - not just “2-burner propane stove, 20,000 BTU”. This creates a compounding loop: trust leads to richer intent, richer intent leads to better personalization, and better personalization drives higher AOV and deeper trust. It's a returns-to-scale effect - and it means the winner here won't just win by a little.

Where will goal-based shopping happen?

If you believe goal-based shopping is going to be big, the natural next question is: where will it live?

For this to work, you need two things - a wide enough product catalog to actually solve life goals, and a place where consumers feel enough trust to hand over that level of intent. The early contenders are splitting along a clear line: Walmart and Amazon on the retailer side, and ChatGPT and Google Gemini on the AI platform side.

The trade-offs are interesting because Walmart and Amazon control the full stack, with inventory, fulfillment, returns, and customer service operating inside one integrated system. That structure makes it much easier to create a high-trust, end-to-end experience. You tell Sparky to build your camping trip basket and Walmart delivers it to your door tomorrow, and if something is wrong you know exactly who to call. It is a closed model, and closed models tend to deliver consistent, reliable experiences.

OpenAI and Google are playing a fundamentally different game. They’re not selling anything - they’re building platforms where merchants plug in via protocols like ACP and UCP. The AI layer is potentially stronger: deeper memory about users across contexts, better personalization, richer conversations. But the commerce layer is fragmented. When your birthday party order comes from five different vendors with five different fulfillment timelines, and one of them doesn’t deliver - who do you ask for a refund? Stitching together a seamless shopping experience across a decentralized merchant network is a very different problem than running your own warehouse.

So the trade-off is essentially: retailers have the logistics but are building the AI. AI platforms have the intelligence but are building the logistics. The retailers start with trust and fulfillment, and need to catch up on conversational depth. The AI platforms start with the best intent-understanding layer ever built, and need to solve the messy reality of actually getting products to people’s doors.

Which side closes their gap faster? That’s probably a trillion(s?) dollar question. Get the 🍿 ready.

Closing thoughts

We're in the first inning of agentic commerce, and the building blocks are taking shape across the AI platforms. Agentic commerce will be around both optimizing the existing user journey, but also about building new exciting journeys through leveraging AI in new ways.

Goal-based shopping will be one of those ways. It is built for how people actually think - in life events, not product categories. The retailers and platforms that figure this out first won't just capture more transactions. They'll be present at the moment intent is formed, which is a fundamentally better place to be.

Hope you enjoyed reading!

If you want to read more in-depth analysis about agentic commerce, don’t forget to subscribe below!