Next.js + Vercel SFCC Composable Storefront

There is a version of the AI conversation that has become painfully familiar in enterprise commerce. Teams say they are "using AI." Developers mention Claude Code. Someone demos a copy generator. A few hours get saved here and there. Meanwhile, CIOs and CTOs look at the org chart, look at the budget, and ask the only question that matters: if AI is real, why are the benefits mostly showing up as a better day for staff instead of a better business for the enterprise?

That is the right question. And it is exactly why Storefront Next on Salesforce matters.

The argument for Storefront Next is not that it gives your team a shinier storefront framework. The argument is that Salesforce is finally putting an AI-first commerce architecture on the table, one that can change how the stack is assembled, how work is divided, how much manual tuning is required, and where automation can compound over time.

For technical leaders, that distinction matters. AI that helps an individual contributor write code faster is useful. AI that changes the architecture so the enterprise can reduce complexity, automate repetitive commerce work, and create a cleaner path to measurable operating leverage is strategic.

The real problem: AI has been too personal, not operational

Most enterprise teams are seeing narrow AI gains. Developers use copilots. Analysts summarize documents faster. Merchandisers generate a few drafts instead of writing from scratch. Those are legitimate improvements, but they are local improvements. They help individuals. They do not automatically rewire the system the business runs on.

That is why so many executives are underwhelmed. If the only visible effect of AI is that smart people can do the same job a little faster, then the enterprise has not captured much value. The headcount line stays the same. Delivery models stay the same. Support models stay the same. Vendor sprawl stays the same. The organization ends up paying for AI, talking about AI, and still operating largely the same way.

That is not transformation. That is productivity theater.

Storefront Next changes the argument

Storefront Next is a full-stack React framework for B2C Commerce that combines server-side rendering with client-side navigation in a managed Salesforce architecture. Salesforce is positioning it as an AI-first storefront and has tied that positioning directly to faster implementation, AI-powered developer tooling, and simpler launch patterns.

That matters because architecture determines how far AI can go.

An inconsistent, over-customized commerce stack is hard for humans to maintain and hard for AI tools to reason about. A more opinionated, convention-driven stack is easier to scaffold, easier to extend, easier to support, and easier to automate. That means Storefront Next is not just another frontend option. It is a better substrate for AI-enabled delivery and AI-enabled operations.

This is the part technical executives should care about. The value is not only in what AI does today. The value is in choosing a platform shape that will allow more of tomorrow's build, support, merchandising, and optimization work to be delegated to AI systems with less friction.

Cimulate is the bigger story than most teams realize

Salesforce's acquisition of Cimulate is one of the clearest signals that this is not a superficial AI story. Cimulate combines real shopper behavior with simulated shopper journeys to infer intent and improve search, discovery, and recommendations in ways that go beyond traditional keyword and rules-based systems.

That should get the attention of any CIO or CTO who has watched teams spend years tuning search synonyms, patching low-conversion discovery flows, and layering manual merchandising rules on top of brittle relevance models. Cimulate's approach is built around understanding intent through both real and synthetic commerce behavior, which means the engine gets stronger not only from what happened, but from what likely could happen across millions of modeled sessions.

That is enterprise value.

Why? Because discovery is one of the most labor-intensive and under-optimized layers in commerce. Teams pour human time into relevance tuning, catalog shaping, landing page curation, campaign setup, and recommendation logic. If Salesforce can productize Cimulate inside the Storefront Next and Agentforce Commerce motion, that opens the door to automating a category of work that has historically required constant expert intervention.

That is the shift skeptical leaders should focus on. This is not AI as a sidecar. This is AI moving toward ownership of an economically meaningful layer of commerce operations.

The architecture is the real signal

Put those two moves together, the AI-first storefront and an intent-aware discovery engine, and the pattern is clear. Salesforce is not bolting AI onto the edges of commerce. It is reshaping the substrate so that more of the work can eventually be automated rather than hand-tuned.

For technical leaders, that reframes the evaluation. The question is no longer whether AI is useful. It is whether your architecture is built to let AI compound. Choosing a platform shape that AI systems can reason about, extend, and operate is the first real decision, and it is the one that determines how much value everything after it can capture.

There's a version of every Salesforce conference that feels like a product catalogue in keynote form, with big themes, bigger promises, and a gap between the stage narrative and the actual code shipping. Connections 2026 in Chicago was not that event. For the first time, the AI story felt less like aspiration and had a whiff of execution to it. The question wasn't whether Salesforce was committed to the agentic future. The question was whether the audience had to energy to go back home and actually do something.

For those of us in commerce, the headline under the headline was this: Storefront Next went GA. And that matters more than most people outside the SFCC ecosystem will appreciate.

Storefront Next: The Foundation Underneath Everything

Salesforce had been telegraphing Storefront Next for months as a next-generation framework built on React Server Components, React Router 7, and TypeScript, with tighter Business Manager integration and managed hosting via MRT. It was always positioned as more than a PWA Kit replacement. It's a platform reset. And at Connections 2026, that reset became real.

Why does anyone need a platform reset right now? What Storefront Next actually represents is a codebase architecture that makes AI tooling work earlier in the project lifecycle. When Salesforce talks about AI-assisted development, they're not talking about a chatbot strapped to your IDE. They're talking about a framework that's tightly coupled enough, standards-based enough, that AI can understand, extend, and generate within it with genuine coherence. SFRA sites can be super fast. They can have CMS and Search plugged in to great effect. But they are not AI-friendly. Storefront Next is. That's not incidental. It's the point.

Our own team has been deep in this since last year: running a PoC, building out a starter, asking hard questions about how you extend cleanly without drifting from upstream. The early read is that the Commerce Apps extension model represents a genuine improvement in how integrations get built. The Page Designer investment, which Salesforce is clearly serious about, couples tightly here too, shifting how content experiences get delivered in ways we're still fully digesting.

At Connections, this was the foundation everything else was built on. The AI-assisted merchant workflows, the agentic campaign management, the Guided Shopping experiences, all of it flows more naturally through a storefront architecture that wasn't designed in 2015. Storefront Next is the plumbing, and the plumbing finally got replaced.

AI in Practice, Not Just in Pitch

What struck most observers at Connections 2026 was not the ambition (that's been consistent for two years), but the specificity. The sessions felt less like vision decks and more like implementation guides. One analyst noted that if Connections 2025 was about AI assistance, 2026 was about AI autonomy:Salesforce leaning hard into the Agentic Enterprise, where agents don't just generate content or summarize data, but take action across the marketing and commerce stack. Sure, I would have thought last year. This year they came with case studies. Maybe it’s trade show slop? But there was energy behind it.

I agree with that framing. Jon Reed at Diginomica covered the Generative Engine Optimization thread, which took up significant floor time, optimising content and products for AI discovery, showing up across the internet, not just brand-owned channels. It was a smart session and Reed covered it thoroughly. Though I'd gently push back on the implication that GEO should be the primary commerce headline of CNX 2026. For the platform builders in the room, the commerce infrastructure story (and specifically the Storefront Next-plus-Agentforce architecture) was the more lasting signal. GEO is a vital question to ask: how will people find us if Google aren’t sending them and they don’t actually need a website to buy stuff. But the agentic commerce platform is the foundational tool that allows a lot of those folks in Chicago to start answering that question. 

The Slack Moment

Let's talk about Slack, because the square footage alone told a story.

Salesforce gave Slack a keynote presence and show floor space at Connections 2026 that made one thing unmistakably clear: this is not a collaboration tool they acquired and manage at arm's length. Slack is being positioned, actively and visibly, as the agentic OS for the enterprise. Core Salesforce apps including Agentforce Sales, IT and HR Service, and Tableau Next now surface directly in Slack conversations. The vision is a single conversational workspace where humans and AI agents collaborate in the same interface, with CRM and all other data embedded and actionable in real time via 3p apps. I’m completely on board with this idea. Humans and Agents need to connect somewhere (for now!). The messaging platform is best place for it to happen. And Slack ought to win here. It helps too that even in Microsoft shops we work with the dev team uses Slack. So this may be a parallel sale to Teams rather than a replacement. The sale will be made in IT and with their endorsement to the rest of the business. Teams will just be the meetings part. 

If you've been watching Slack's positioning carefully, this isn't surprising. Salesforce has been calling it the "agentic enterprise operating system" for months. But there's a difference between a booth at Agentforce World Tour and a show floor centerpiece in Chicago. At Connections, it was at the middle of everything. For enterprise buyers evaluating whether Salesforce is serious about this architecture, that kind of investment in physical presence and keynote time is not accidental. It's a declaration.

For SFCC practitioners, the relevance is direct: the merchant workflow future that our 64labs product team have been sketching out (where an Easter campaign gets planned, built, measured, and adjusted across multiple teams without six FTEs and four weeks of elapsed time) lives in something like this. Slack is the winner.

‍

The Contentful Question: Wait, or Don't?

Then there's Contentful. Salesforce signed a definitive agreement to acquire Contentful on June 1, 2026. The analysis is largely positive: Contentful closes the CMS slot in Headless 360 that had previously been left open, and the integration story for Agentforce deployments becomes considerably cleaner when your content management layer is owned by the same platform.

It may be a bit of a nothing burger for Commerce Cloud. A kneejerk question for any commerce buyer evaluating a new SFCC build right now is: should I wait? If Contentful becomes native to the Salesforce platform, won't the integration be smoother in twelve months than today?

Maybe. But Commerce is not Salesforce’s priority. The deal won’t close until September. The integration into Agentforce beyond commerce will take priority. Contentful already works in Storefront Next if you need it to. The integration question used to be a stronger argument for hesitation than it is now. The development complexity calculus has shifted materially with Storefront Next. The extension model is cleaner. AI-assisted development with a standards-based codebase means implementation velocity is meaningfully faster than it was eighteen months ago. The carrying cost of waiting (in roadmap delay, in continued technical debt on SFRA, in missing the Agentforce capabilities that are already GA) is real.

So the answer isn't wait. The answer is to plan intelligently: choose Contentful if you love it for what it is. But don’t delay Storefront Next or imagine that there isn't still a good reason to pick Contentstack or Amplience or Builder et al. There are. So the best time to be implementing Contentful in SFCC is probably two years from now. I’m not sure the purchase changes much in ecommerce until 2028. 

The Through-Line

What Connections 2026 made clear is that Salesforce has arrived at the point they've been building toward for several years. The AI isn't decorative anymore though some presentations still very much need the disclaimer slide. Storefront Next provides the foundation for AI to actually function in commerce. Slack is the interface layer they're betting on at enterprise scale. And the Contentful acquisition, whatever its timeline, closes the last obvious gap in Salesforce’s “composable monolith” story.

The enthusiasm in Chicago wasn't the performative kind you sometimes sense at these events, the kind where everyone's excited because everyone else is excited. It was the quieter, more durable kind: practitioners who came in skeptical, left with viable case studies and a clearer sense of what the roadmap actually holds. That's the version of a Salesforce conference that goes beyond nabbing cuddly toys for your kids and actually having something to think about on the plane home.

‍

For years, Salesforce Commerce Cloud has had a cost problem; and it was never just licensing.

The real cost lived in the ecosystem around it: custom templating languages, hard-to-find specialists, slow iteration cycles, and a persistent dependency on external partners for even modest changes. Teams didn’t just pay for SFCC, they paid to operate it.

Storefront Next changes that equation in a fundamental way.

By moving to a modern, standard frontend stack - React, composable architecture, familiar tooling - you eliminate one of the biggest hidden taxes in the system: specialization. No more hunting for niche ISML experts. No more translating between modern frontend thinking and legacy templating constraints. You can hire from the broader market, use proven patterns, and move at the speed your team is actually capable of.

That alone lowers total cost of ownership. But it’s just the starting point.

What really changes is how companies choose to operate.

Two Paths To Tread

Once the barrier of “expensive to change” disappears, companies will likely split into two operational camps.

The first group will optimize for steady state. They want the site to run cleanly, predictably, and with minimal intervention. Their goal is a near-zero backlog, a very small team, and a system that largely takes care of itself. Updates are incremental. Automation handles the routine. Human involvement is reserved for exceptions. 

For these teams, Storefront Next is about cost control. It allows them to finally run SFCC themselves as a stable, low-maintenance platform instead of an ongoing project.

The second group sees the opposite opportunity. If the cost of change drops, why not do more?

More experimentation. More landing pages. More personalization. More campaigns. More integration across systems. The same team, or an even smaller one, can now produce significantly more output because they are no longer constrained by platform friction.

This is where Storefront Next can be more about growth than cost.

And of course in practice, most companies should do both.

Stabilize, Then Expand

This is the pattern we recommend for dealing with the confusing world of AI. Don’t be carried away in the hurly burly of innovation. Go up the learning curve in stages. Iterate and learn over time. Keep moving forward but don’t go faster than your organization (probably still quite slow) can handle.

First, get to a solid steady state. Reduce the noise around support. Eliminate the backlog. Automate the obvious operational work. Bring the cost of “keeping the lights on” as close to zero as possible. This is a business process question as well as a development and marketing one. Do what you do better and quicker and more automatically. Stop doing things that humans require but machines don’t - about 50% of the meetings you do each day? 

Then take that recovered time and reinvest it.

Not into rebuilding what you already have or adding fancy so called AI-native tools - but into extracting more value from the stack you’re already paying for.

That means connecting the pieces of your stack that have historically operated in silos:

CMS feeding landing page generation
Personalization driving dynamic content decisions
Search influencing merchandising and discovery
Campaign tooling tied directly into experience delivery

Individually, these systems are powerful. Connected, they become multipliers. (You already knew all this but there wasn’t time or mental energy to actually achieve the achievable)

And once connected, all these newly exploited existing systems become candidates for automation. Without spending an extra penny on an AI enablement platform. 

Automation Before AI

There’s a misconception that AI transformation starts with buying an AI product.

It doesn’t. It mustn’t.

It starts with getting your stack into a state where it can actually support automation. Clean data flows. Defined workflows. Clear system boundaries. Reliable integrations.

If you automate your current workflows and your current workflows are designed around human frailties then you will just have automated mediocrity. What a waste. Get the most out of what you already have. Organize yourselves to be efficient, and let automation take care of as much of the resulting work burden as “humanly” possible.

Only then does AI have something to work with.

What happens next is not a switch you flip, it’s a gradual process. You automate pieces of the workflow. Then you introduce AI to optimize, generate, or orchestrate within those workflows. Over time, more of what was previously human-gated becomes system-driven.

But none of that works if your frontend is slow, rigid, or dependent on scarce skills.

This is where Storefront Next quietly becomes strategic.

Why SFCC Becomes More Valuable

SFCC has always been flexible. That was never the issue.

The issue was that most teams couldn’t fully use that flexibility. The cost of implementation, the friction of the frontend, and the dependency on specialized skills meant that much of the platform’s capability remained theoretical.

Storefront Next changes that.

When teams can actually build, iterate, and integrate without friction, SFCC’s flexibility turns into a real advantage. It becomes the center of a composable, connected, and increasingly automated commerce stack.

And importantly, it becomes a platform that can evolve.

Not through large, expensive replatforming efforts, but through continuous, incremental improvement.

That’s the shift.

Lower cost of ownership is the entry point. Operational leverage is the outcome. And long-term adaptability is the real payoff.

If you’re already on SFCC, the decision is straightforward. The question isn’t whether Storefront Next is worth it - it’s whether you want to keep paying for constraints that no longer need to exist.