Industry
Insights

Winning the agentic customer.

For thirty years, we've optimized product experiences for human shoppers. Now a second customer has arrived: AI agents that discover, evaluate, and increasingly transact on shoppers' behalf. AI will influence a trillion dollars or more in commerce in the coming years.

AI won't replace how we shop today any more than ATMs replaced bank branches or ecommerce replaced the store. Brick-and-mortar still represents 80% of retail spend. But AI becomes another interface through which consumers discover, evaluate and buy. It's a demanding one. If an agent can't understand a product because the data is incomplete, inconsistent, or out of date, it won't discover it, recommend it, or buy it. Machines don't read bullets. They read narratives and they mine ratings and reviews to infer whether a product matches shopper intent.

What separates retailers generating real ROI from those stuck in AI experimentation mode is their data foundation: governed, centralized product data as the single source of truth, modern infrastructure feeding every channel and AI application, and treating product data as core commerce infrastructure rather than a back-office chore. The retailers who build that foundation will show up ready at every moment of choice, on every shelf, for humans and machines alike.

Look for more solutions around inventory truth, because it is the prerequisite for the other outcomes. Labor substitution, loss reduction, uptime, fulfillment accuracy, dynamic pricing, personalization, and AI automation all depend on knowing what inventory exists, where it is, whether it is sellable, and whether the system record matches operational reality.

Inventory truth directly affects revenue, margin, customer experience, and labor. If inventory data is wrong, retailers oversell online, underperform in-store, misallocate labor, lose sales through out-of-stocks, and make AI recommendations on flawed data. For executives, inventory truth is a board-level issue because it links operational execution to omnichannel growth, shrink control, working capital, and customer loyalty.

The biggest priority for US retailers planning for 2027 is delivering on the expectations of the hyper-intentional shopper. Burdened by compounding inflation fatigue, these consumers are extracting value at every turn, demanding more from their discretionary purchases, and exerting strict control over expenses by using AI tools to curate choices and shop agentically.

Maintaining price discipline on functional essentials where shoppers are prone to trade down must be balanced with offering choices that satisfy emotional needs and prove a brand is “worth it.” That sense of worth will increasingly come from products and services that deliver an undeniable advantage in quality while providing comfort, alignment with personal values, or a micro-indulgence, a critical emotional release given that a 2026 Capgemini Retail Trends study found that 71% of consumers say small indulgences help them cope with financial stress.

We see this at play today, but with increased adoption of LLMs for discovery, retailers will need to meet these functional and emotional needs while continuing to refine their algorithmic relevancy and ensure that their omnichannel experience is seamless and intuitive. In this respect, 2027 looks like prioritizing the fundamentals of commerce as practiced for thousands of years, but now with technologies that were once the stuff of science fiction.

Over the next two years, retail media’s greatest constraint won’t be measurement, it will be operational complexity. As networks expand across thousands of in-store screens, ecommerce placements, and mobile experiences, teams will face a constant flood of decisions that even the best dashboards cannot manage at scale. Every missed decision has an immediate cost: lost revenue, wasted inventory, and a degraded customer experience.

The next generation of platforms won’t simply report what happened. They will continuously observe their own operations, apply reasoning, and resolve issues before anyone files a ticket. Operational intelligence will become as fundamental to retail media’s future as measurement is today.

Retailers’ 2027 priority: build the AI-ready discovery foundation.

The biggest focus for retailers planning 2027 should be moving from “better search” to intent-led, AI-ready discovery. One thing is clear: shopper behavior is changing faster than most retail infrastructure. While customers are still just typing keywords into a search box, many are now expressing needs in natural language, expecting relevant outcomes instantly, and increasingly discovering products through AI tools before they ever land on a retailer’s site.

That makes 2027 less about adding another shiny AI feature and more about fixing the foundation: product data, content structure, retrieval, relevance signals, availability, merchandising logic, analytics, and personalization. Otherwise, retailers will end up bolting conversational interfaces onto systems built for yesterday’s browse-and-filter world. That is not transformation. That is lipstick on a legacy database.

The urgency is obvious. 52% of retailers still rely on out-of-the-box third-party search, while only 8% are actively preparing infrastructure for AI agents to interact with or navigate their sites. Investment is also scattered across recommendations, personalized results, semantic search, agentic search, merchandising tools, and content-led discovery, creating a risk of doing a little of everything and mastering nothing.

The winners in 2027 will be the retailers that pick a clear lane: capture intent better, structure data for AI, unify search, browse, recommendations, and content, and give teams the tools to test, optimize, and move quickly. In short, stop treating discovery as a website feature. Treat it as the operating system for commerce.

Made before shoppers ever visit a retailer. Our College 2026 research suggests that many students are already using AI to compare products and narrow their choices before they ever visit a retailer's website or store. We also found that traditional factors like brand recognition matter less than many retailers assume. Instead, students are much more likely to buy products they believe are popular with their peers.

For retailers, that means influencing the decision earlier in the process. Rather than focusing only on driving traffic to a website or running standalone digital campaigns, they'll need to build credibility where students and Gen Z are discovering products and reinforce that message across the places and communities students already trust.

Winning consumer decisions—not just consumer attention.

Retailers are entering 2027 as multiple structural forces reshape how consumers make purchase decisions. AI is changing how shoppers discover and evaluate products. GLP-1 adoption is influencing category demand and basket composition. Value-conscious consumers are planning earlier and purchasing more intentionally, while rising expectations around convenience continue to redefine the shopping experience.

These shifts are happening in real time, making traditional planning and post-campaign measurement increasingly insufficient.

The old marketing model—plan, launch, measure weeks later—is too slow for today's consumer. Brands need to know whether a protein-forward assortment is driving incremental purchases, whether a value message is changing basket behavior, or whether a convenience-focused campaign is increasing conversion—not after the holiday season, but while there's still time to optimize.

That's why the next competitive advantage isn't simply more data—it's Outcome Intelligence: the ability to connect media investment to real shopper behavior, continuously optimize against actual business outcomes, and make smarter decisions while campaigns are still in flight.

The retailers that outperform in 2027 won't be the ones with the biggest budgets or the most sophisticated algorithms. They'll be the ones that can recognize changing consumer decisions first—and confidently turn those insights into action before the competition does.

Realize that the biggest retail media opportunity does not sit inside any single retailer's walled garden. It sits in the store itself, activated across retailers. Roughly 95% of CPG purchases happen in-store, yet in-store captures only about 1% of retail media spend. Over the next 12 months that gap starts to close, driven by cross-retailer networks that combine store-level sales data, geotargeting and closed-loop measurement, so brands can plan, activate and prove campaigns at the shelf with the same precision they expect online.

Look for ways to reduce their AI costs through efficient diffusion of intelligence throughout their business. The increasing cost and uncertainty around the largest AI models will drive them to diversify and choose the right model for the right task, especially in high-volume, latency-sensitive applications that are core to their business.

The e-commerce experience in particular will undergo a renaissance for both merchants and buyers. AI will drive more accurate and predictive insights; more relevant recommendations and results; improved protection against supply chain disruptions and transaction/payments fraud; diversity in buying channels with the rise of AI agents -- and it will all be powered by efficient, tailored AI models running at scale.

This is a natural part of the AI maturity curve: start with the biggest and most powerful capabilities at the frontier, build organizational muscle and learn what is possible, and then optimize over time for real-world constraints and value generation. We are reaching the inflection point where simply using AI is not enough–leveraging AI with meaningful, measurable ROI will be the trend for 2027.

…retailers realizing that a single BNPL provider is a growth ceiling, not a safe choice.

Every BNPL brings its own audience to checkout: its own customer base, credit profile, and spending habits. A retailer running just one provider is only ever tapping into that one pool of shoppers. You wouldn't drop Mastercard because you already take Visa, yet retailers still treat BNPL like a one-and-done decision. In 2027, the retailers pulling ahead will be the ones who stop asking "which BNPL should we offer" and start asking "which BNPLs." Diversifying payment options at checkout won't be an afterthought anymore. It'll be treated as a growth lever, the same way retailers already treat ad channel diversification. The ones who get this right will out-acquire the ones who don't, simply because they're opening the door to multiple ready-to-spend audiences instead of one.

The biggest shift facing marketing in 2027 will be the collapse of the marketing funnel as the operating model for how organizations plan, measure, and invest.

Consumers no longer move neatly from awareness to consideration to purchase. Their journey zigzags across social, retail media, search, creators, AI assistants, and marketplaces—sometimes collapsing into a single moment of inspiration and purchase.

The real challenge isn't that the funnel has collapsed. It's that most organizations haven't.

Marketing teams are still organized around channels instead of customer outcomes. Budgets sit in rigid P&Ls that make cross-channel optimization difficult. Success is measured by platform-specific metrics rather than business impact.

The next generation of marketing organizations will look fundamentally different. Teams will optimize customer journeys, not individual channels. Budgets will become more fluid, moving to the investments driving the greatest business outcomes. And measurement will shift from last-click attribution and channel KPIs to connected signals that reveal how every marketing investment contributes to growth.

This isn't just a technology evolution—it's an organizational one.

That's why we built Pacvue Prism. Not as another dashboard, but as the intelligence layer that connects every channel, every signal, and every outcome into a brand's unique Agentic Commerce Grid—giving marketers a complete view of how commerce actually works today.

To find the dollars to fund new assortment, brand building, advertising, shipping speed and cross-channel orchestration. The customers expectations on quality, value and immediacy keep rising. The brands and retailers that work together to maniacally remove wasted activity, extra costs and supply chain expenses will win market share.

Rebooting their POS systems.

We have seen physical brick-and-mortar stores continue to be the primary engine for retail growth–our research with RSR shows 85% of retailers view their stores as their primary growth channel. However, achieving this growth requires a massive shift away from rigid, legacy technology toward agile, associate-friendly solutions.

Physical store-led growth can only materialize if the technology powering the store can keep up. A third of retailers told RSR their Point of Service (POS) is actively holding them back, and less than half said it supports an innovative or differentiated experience.

The checkout or the “cash wrap” area has always been the most uninspired element of any store visit. This is why we will see retailers move to overhaul what has been a dead and lifeless “last mile” of the shopping experience.

Retailers in 2027 will move to combine the power of digital marketing and payments by strategically placing interactive customer-facing screens at checkout, to be a meaningful and personalized touchpoint, up-leveling the shopping experience while streamlining payments and driving incremental sales.

Everything after the buy button, because it's about to become the only part of the customer journey they fully control.

I've spent 14 years in the part of commerce most people ignore: post-purchase. It's always been treated as a cost center. I think 2027 is the year that flips.

Agents are taking over the front of the funnel. Shoppers will discover, compare, and buy inside a conversation the brand never sees. Your search strategy, your ads, your storefront, all of it matters less when an agent is doing the shopping. What do you still own? Everything after the order. And that's no consolation prize. US retailers will take back about $850 billion in returns this year.

The industry hasn't caught up to the obvious problem: agents can think, but they can't act. Refunding a customer or rerouting a shipment means touching carrier and warehouse systems that took over a decade to wire together, and plenty still don't have a real API. On the newest benchmarks, the best models finish fewer than 4% of real operational workflows end to end. Intelligence was never the bottleneck. Access is.

My early call for 2027: the agentic commerce race won't be won by whoever has the smartest model. Models are the easy part now. It'll be won by whoever owns the data and the pipes that let an agent finish the job. In commerce, that battle will be won after the buy button, not before it.

As we look ahead to 2027, I wanted to share my executive perspective on the U.S. logistics industry, drawing from current indicators, freight forecasts, and insights from our clients. As we navigate these changes and seize the new opportunities they present.

Carrier Market Outlook:

The freight market has shifted significantly. After years of excess truck capacity, industry forecasts indicate we are entering a carrier-favorable environment with a tighter truckload market by 2027. Key trends include:–Declining carrier capacity–Improved spot rates–Strengthening contract pricing as capacity exits the market.

In summary, these changes highlight the growing importance of strong carrier relationships over purely transactional procurement. Managed transportation providers, like ODW, become increasingly valuable as market complexity rises. Many shippers who moved transportation in-house during the soft market are now returning to managed transportation partners as capacity tightens. This evolution presents exciting possibilities for us all.

Warehouse Automation Trends:

Labor availability remains the most significant structural challenge across U.S. distribution networks. Internal ODW materials emphasize continued investment in warehouse automation, labor management systems, robotics, inventory optimization, and fulfillment productivity. By 2027, we expect:–Autonomous Mobile Robots (AMRs) to become common–AI-enabled labor planning to be standard practice–Vision systems, automated unloaders, and inventory technology tools, including ODW’s investment in Dexory, to see broader deployment.

In summary, automation is moving from a competitive advantage to a requirement, enabling your business to stay competitive and productive.

Finally make good on one of marketing's longest-standing promises: true 1:1 personalization. Not because marketers suddenly have more resources, but because AI is fundamentally changing what's possible.

Retailers will move beyond broad audience segments to create experiences that reflect individual context, intent, and behavior at a scale that simply wasn't achievable before. Those that combine AI with continuous experimentation will learn faster, optimize every interaction, and deliver relevance without adding operational complexity. The winners won't be those creating the most content, they'll be the ones delivering the most relevant experiences.

Start testing what more human, judgment based and opinionated service looks like. An example being, “meal architects” in grocery sector.

If allowed, AI-enabled technology solutions will continue driving retail toward a homogenized experience; every store starts to feel the same, so it stops being a reason to choose one over another. The differentiation shifts to something AI can never be: a human with real opinions.

That’s where meal architects come in–staff who still use data (tied to your loyalty account) to know you cook for a family of six or that you just went dairy-free, but who bring their own judgment and taste to the conversation.

The winners won’t be the grocers with the best AI, they’ll be the ones who used AI to make room for people worth listening to; people with real knowledge, opinions, and education, out front and approachable.

See the demand for shopping agents take on new life. These agents had a slow ramp up–retailers weren’t exactly sure how to implement them, and customers didn’t trust they weren’t just old chatbots in new wrappers. There’s no doubt that’s changed. Tools like Gemini and ChatGPT have retrained customers on how to shop, and conversation has become the expectation. Retailers now have to adapt if they want to meet customers where they are. While we’ve already seen early adoption of agents at retail giants like Amazon and Walmart, expect to see an industry-wide shift in 2027. We’ll see shopping agents really take off across retail, and further, we’ll begin to see the convergence of shopping agents with search engines. More ecommerce search bars will serve as the conversation’s starting point, allowing customers to discover products, receive guidance, and even handle customer service inquiries through a single point of entry.

Building a trusted content infrastructure: authentic UGC, always on.

AI will make content nearly infinite and increasingly difficult to trust. As synthetic product imagery, automated reviews and AI-generated campaigns flood every channel, the scarce asset will no longer be content itself. It will be credible proof from real customers.

Winning retailers will build the infrastructure to continuously capture authentic customer experiences, verify their provenance, secure the rights to use them and turn them into measurable assets across commerce, retail media and AI-powered discovery.

This is bigger than influencer marketing or UGC. Customer content will become a strategic data layer: helping retailers understand why people buy, giving AI systems stronger signals and giving shoppers confidence that what they see is real.

In 2027, the brands that win won’t be those producing the most content. They’ll be the ones consumers and machines can verify, trust, and act upon.

In 2027, point AI solutions will start consolidating towards a unified AI platform with persistent memory.

Brands have spent years adding a new tool every time a new problem appeared on their storefront. It made sense when each tool was solving its own problem, but I don’t think this stack will continue as storefronts become more agentic.

A beauty quiz may know the customer’s skin type, while the support system has the order history. But neither system carries that context into the customer’s next visit, so the customer has to start again even though the brand already knows them.

The reason for consolidation is not only that brands are tired of managing multiple vendors. As AI starts doing more of the shopping journey, memory becomes much more important. An agent helping the shopper over time needs to know what happened earlier, and it cannot do that well when the context is sitting inside separate tools.

Over time, I think many of these separate functions will come together into one commerce agent for each brand.

This will make things difficult for single-function commerce tools. Shopify and Salesforce will continue building more agentic capabilities into their platforms, while the more complete commerce agents take over a larger part of the customer journey.

I won’t be surprised if we see quite a few acquisitions in this category in 2027, while many smaller tools quietly shut down. Most likely, the tools that own the customer memory will survive. It will be much harder for a tool that only solves one small part of the journey.

That AI becomes part of the shopping process itself. More consumers will rely on AI assistants to research products, compare options, and make purchasing decisions before they ever visit a retailer's website. That means product content will need to be structured, complete, and understandable for AI systems as well. Brands that prepare for AI-assisted shopping today will be much better positioned for tomorrow.

Transition from physical storefronts to sophisticated online experiences powered by real-time, cross-product personalization.

To win, brands must shift from passive, backward-looking 'likeliness to buy' models to capturing live, in-session signals to 'incite to buy', in the moment of highest intent. Personalization is no longer about guessing what a customer might do next week; it’s about responding dynamically to what they are doing right now.

Crucially, this will be paired with an urgent demand for privacy-first trust. Consumers are fatigued by 'creepy' AI that attempts to mimic human empathy. In 2027, winning brands will reject synthetic human replication and instead deploy AI that prioritizes high-utility, context-aware relevance. By delivering helpful, respectful, and boundary-aware experiences, retailers will drive authentic conversions and build pipeline while protecting consumer trust. The future of retail AI isn't about synthetic intimacy, it is about being exceptionally useful, replacing guesswork with real-time intent to boost conversions without invasive surveillance.

Over the next 12 months, I expect retailers to continue relying on retail consolidation networks to improve delivery performance and reduce costs. The shift will be toward 3PLs that can consolidate freight for a broader range of retailers; not just major chains like Walmart and Target but also middle-tier retailers. As customer requirements become more complex, providers with more extensive consolidation networks and greater retail coverage will be best positioned to help shippers improve efficiency, reduce chargebacks, and simplify their supply chains.

How to manage increasing pricing pressure, while offering unique, differentiated and personalized products for buyers.

Shifting from reactive noise-matching to agentic, strategy-driven decision intelligence.

Over the past few years, the retail sector rushed to adopt real-time data feeds and basic AI automation. However, this led many brands into a dangerous trap: reacting to every competitor price shift, algorithm fluctuation, or market blip—effectively waging a race to the bottom that erodes margins without driving true customer loyalty.

By 2027, the market will penalize blanket automation that operates as a black box. The next phase is agentic pricing: AI agents that don't just recommend but execute, filtering signal from noise, aligning prices across digital channels, marketplaces, and stores, and acting autonomously within guardrails your commercial team defines. But autonomy without transparency is a liability. Winners will demand "Glass Box" AI, where teams understand why every decision is made and retain strategic governance.

The retailers who thrive in 2027 won't be the fastest to match a discount. They'll be the ones who elevate their teams from manual execution to strategic oversight, letting agents handle the how, while humans own the why.

Automate your promo campaign production, and get it 100% right the first time. One of the key insights of our recent Brandwidth research into Creative Operations in Grocery Retail: 72% of respondents recently experienced a promo failure (error in price, visual, discount, product information), and only 11% are confident their infrastructure and workflows are sufficiently robust and will pre-empt the next error. Main bottlenecks: endless Q&A processes and unavailable media and data. Nevertheless, the demand for more creative output will only grow in 2027. Therefore, building a strong foundation for automated, AI-powered promo production needs to be priority #1.

Creating stores that are built to adapt. Consumer expectations continue to evolve, and retailers need the flexibility to respond in real time. That applies to pricing, inventory, promotions, labor challenges, and changing shopping behaviors.

According to Pricer's global consumer research of more than 5,000 shoppers, trust is built through consistency. Shoppers expect prices to be accurate, products to be available, and information at the shelf to match what they see online. When those expectations are not met, even small moments of friction can influence purchasing decisions and long term loyalty.

The retailers that will be most successful in 2027 will not necessarily be those with the most technology. They will be the ones investing in connected store infrastructure that helps associates work more efficiently, enables faster execution, and gives stores the agility to respond to change without adding complexity. A smarter shelf is not just about displaying information. It creates a real time connection between the retailer, the store, and the shopper.

As retail continues to evolve, adaptability will become a competitive advantage. The ability to make changes instantly, empower store teams with better tools, and deliver a consistent experience across every location will separate retailers that are simply responding to change from those that are proactively ready for whatever comes next.

A consumer who reroutes spend faster than retailers can re-plan. This year's gas spike showed the pattern: dollars rotated before they disappeared. Fuel spend shifted into warehouse clubs, value formats picked up share from full-price channels, and every income tier put a bigger share of the wallet into gas. None of that was a pullback. It was reallocation, and it happened in weeks.

If gas prices ease, history suggests the recovery will be just as ordered: in 2022, sporting goods, off-price, and quick service re-accelerated first, and that's the sequence to watch for. If they don't, expect the rotation into value to deepen instead.

Either way, the 2027 challenge is speed. Consumers now reroute the wallet in weeks while most retail planning still runs in quarters. Closing that gap is the difference between reading the rotation and being on the wrong side of it.

Treating delivery as a decision system, not a siloed logistics function. Every decision should first be driven by customer behavior and what actually impacts conversion. Surfacing choice, such as BOPIS, same-day, next-day, and standard shipping, at the right price is table stakes. The real differentiator is knowing which node to source from (store, DC, 3PL) and which fulfillment channel to use (in-house fleet, third-party gig, big-and-bulky, regional carrier), all based on real-time performance, profitability, and flexibility, not fixed routing rules. Retailers who master that will turn delivery from a cost center into a conversion lever.

Making their entire product catalog instantly discoverable to both traditional search engines and emerging AI search bots. As consumer behavior shifts from "blue link" queries to conversational discovery on platforms like ChatGPT, Gemini, and Copilot, the traditional SEO playbook is no longer enough. We are firmly in the era of GEO (Generative Engine Optimization). To win in 2027, retailers must ensure their product data, long-tail landing pages, and internal link structures are highly structured, lightning-fast, and optimized for AI crawlers to parse. Crucially, executing this cannot be held back by developer queue lines. Retailers must bypass the IT bottleneck to dynamically deploy programmatic search strategies at scale and on autopilot. The brands that automate this discoverability will capture high-intent traffic across all search channels; those relying on manual, slow, IT-dependent deployments will simply become invisible to the next generation of shoppers.

Shift to personalized protection, and it will replace one-size-fits-all protection bundles. The businesses that win in 2027 will treat protection the same way they treat pricing and product recommendations: as a real-time decision informed by transaction data, purchase behavior, and inferred risk, not a generic add-on shown to every customer. Protection will become another personalized surface in the buying journey instead of an afterthought at checkout. The gap between platforms that personalize protection and those that don't will show up directly in the numbers: attach rates, average order value, and retention. Protection will become a measurable growth input because the businesses offering the right coverage to the right customer convert more of the customers already in their funnel.

The end of mass-market discovery.

The mass market was never a law of nature. It was a workaround. Shelf space was scarce. Media was expensive. So brands collapsed the messy diversity of human needs into one brand, one message, one mass market.

That’s all changed. Ecommerce broke the shelf. Social media broke broadcast. AI is breaking the last constraint: Algorithmic search.

Shoppers no longer squeeze what they want into blunt keywords like “best laundry detergent.” They ask AI: “Is there a plastic-free laundry sheet that shrinks my environmental footprint and still gets my kids’ clothes clean?” “What cereal will actually keep my 7-year-old full until lunch?”

AI can now answer all those messy questions. But it can only recommend what it can understand. If your product page says “powerful, eco-friendly clean,” no model knows your sheets are plastic-free or tough on grass stains. If the story isn’t written down, the product doesn’t exist.

In 2027, winning brands will stop treating product content like copy and start treating it like infrastructure, covering every real reason people buy.

The USP is a broadcast-era relic. Intent coverage is what will replace it.

Stop treating personalization as a nice-to-have and start treating it as the baseline for winning attention. Grocery and consumer brands are competing against everything in a shopper's feed, and audiences decide in seconds whether an ad is relevant to them. Generic video–no matter how well-produced–blends into the noise.

The retailers and brands that pull ahead in 2027 will be the ones who solve the 'scale vs. relevance' tradeoff: turning one creative concept into thousands of personalized variations using real shopper data, assembled in real time at the moment of impression, rather than relying on static, one-size-fits-all creative. We're already seeing this shift show up in the numbers–unified, data-driven personalization is outperforming decoupled, generic approaches by wide margins in head-to-head tests.

The winners won't be the brands with the biggest budgets–they'll be the ones who can activate their existing creative and data to deliver relevance at scale, across CTV, social, and programmatic, without sacrificing speed or brand control. In grocery specifically, that means dynamic, real-time offers and pricing that reflect what's actually happening on shelves and in shopper carts–not a generic spot everyone sees the same way.

Building a predictable retention system, especially before adding more AI.

Across the retail accounts we look at each month, the pattern repeats. Fifty or sixty active flows, pages of segments nobody can explain, post-purchase journeys still recommending products that are out of stock, and repeat revenue sitting in the low single digits. Every agency added its own flows and tools. Almost nobody stepped back and decided how the journey from first order to second is actually supposed to work.

Acquisition is usually coherent. There is a budget, an owner, weekly reporting. After the first order, email holds one version of the customer, support holds another, and SMS, the site and the stores each hold fragments.

By 2027, using AI will not be a differentiator. Most retailers will have similar tools. The advantage will come from whether those systems are all working from the same version of the customer. If the data stays fragmented, the AI will be fragmented too. AI will make a strong retention system much more valuable, but it will also make a weak one much more obvious.

The collapse of the wall between front office and back office. Retail has always treated them as separate worlds: the storefront sells, the ERP counts. AI shopping agents end that separation, because what they interact with is not your website's design–it's your operational truth. Live inventory, accurate product data, real pricing, actual delivery promises: agents query it, compare it across merchants, and transact against it directly. In that world, your ERP effectively becomes your storefront, and every operational weakness becomes a lost sale you never even saw. This is the argument for unification that retail has postponed for a decade. Point solutions made sense when a human clicked through your site and forgave small inconsistencies. Machine buyers forgive nothing. By 2027, the retailers winning AI-driven commerce will be the ones who rebuilt on a single connected system where commerce, inventory, orders, and financials share one brain–because that brain is what the agents are talking to.

Your fastest-growing segment in 2027 won't have a pulse.

In 2004, about 2% of US retail happened online. Most merchants considered it to be "interesting, not urgent." That 2% is now nearly 1/5th of all US retail, $1.2 trillion a year. E-commerce readiness was the defining trait of the winners of this era.

Today, AI agents sit at roughly 2% of online commerce. And this time, the curve is steeper: Adobe measured a 4,700% year-over-year jump in AI-driven traffic to US retail sites; those visits convert 31% better than average. Bain and Morgan Stanley both forecast that 10-25% of US e-commerce will be agent-led by 2030.

E-commerce took 20 years to run that arc. Agents will do it in 5. And once again, the winners will be those best prepared for the coming shift.  For every merchant, the playbook has to change now.

Building confidence in the information behind every capital decision.

Every portfolio strategy eventually meets site-level reality. If the information behind your stores is incomplete, outdated, or inconsistent, even great strategies turn into change orders, delays, and difficult tradeoffs.

I think we're entering an era where an accurate understanding of the physical portfolio becomes a competitive advantage. The organizations that can make capital decisions based on what's actually true across hundreds or thousands of locations will simply move faster and execute with greater confidence than those still rediscovering every store with every initiative.

Curious whether others are seeing the same shift as they think about priorities for 2027.

How retailers react to the newest layer of commerce visibility. Most retailers didn’t realize that agentic shopping would become such a core pillar of the ecommerce industry–it was a “nice to have” for a moment before it became a make-or-break competitive differentiator for even the most established retailers to stay relevant. I expect 2027 to be the year where the AI winners continue to rise, while many others are left behind.

Maintaining a presence across global marketplaces, ad networks, retailer websites, social media and AI platforms is one thing, but doing so with accurate, high-quality product and inventory information is another–this is now a non-negotiable with AI. Over half (58%) of shoppers say their trust in a brand decreases when AI gives them the wrong product information. Loyalty often stems from the first touchpoint a shopper has with a brand. In the next 6-12 months, that touchpoint will increasingly be AI-powered, whether on a retailer’s website, directly through LLM or a Google search. When executed correctly, retailers can give shoppers access to products wherever they begin their shopping journey–setting the retailer up to crush peak sales seasons and every moment in between.

Higher grocery prices and GLP1's are causing the biggest contraction in grocery we've seen in years (2% since February). One way to grow revenue that most grocers overlook is EBT. It represents over $100B in grocery spend each year, and it's one of the best ways to build trust with cost-conscious shoppers.

1 in 2 Americans, over 150M people, will be eligible for EBT at some point in their lives. And for many families, the EBT checkout is the most frustrating part of grocery shopping. Legacy processors are down 4–5% of the time, which means that 1 in 20 EBT transactions gets declined for no reason.

Cost-conscious shoppers remember how they were treated when they were struggling, and that's when long-term loyalty is earned.

That Grocery Retail will finally feel the pain (and challenges) that most other industries have faced over the last 20 years. Grocery will finally reframe their eCommerce operations approach with AI forcing acceleration from customer, pricing, merchandising, and operations. eComm grocery is now big enough and growing fast enough to force many to break things down back to first principles. Grocery Retail needs to start from the beginning–not the middle–because the middle is full of baked-in assumptions. Building for the future on the old foundation is the wrong recipe. The bookstore, the shoe store, the department store all had their days of reckoning. The complexity of the cold chain is the main reason why Grocery hasn't shifted as quickly as other industries... but in 2027, we are here.

Rebuilding their teams around AI rather than replacing their teams with it. Having walked the Shoptalk floor for years, I've watched the technology stack across fulfillment, payments, and CX mature at extraordinary speed. The next leap isn't more tools; it's people who know how to run them. We're already seeing offshore teams use AI to triage inquiries and surface real-time insights, cutting wait times dramatically while lifting satisfaction. In 2027, the competitive edge will belong to retailers who pair AI with skilled global talent that's genuinely integrated into the brand, not bolted on beside it.

Retailers will move beyond using AI simply to create marketing content and start using it to manage the entire customer journey. Creating emails or texts faster is valuable, but the bigger opportunity is autonomously and iteratively crafting the right message, channel, timing, and sequence for every customer.

For years, marketers have known personalization drives better results, but creating unique experiences for millions of customers wasn't realistic. AI now makes that level of personalization possible at scale and in real-time.

We'll also see AI agents fundamentally change customer relationship management (CRM). Instead of marketers manually building journeys and making campaign adjustments, AI will handle more and more of those tactical decisions in real time based on customer behavior and business goals. Marketers won't become less important, but their role will evolve. They'll spend less time managing execution and more time setting strategy, establishing brand guardrails, auditing outputs, and measuring business impact.

The retailers that pull ahead won’t be the ones that use general tools like Claude or ChatGPT to run an analysis or create content faster. The leaders will use purpose-built systems to personalize every stage of the customer journey 24/7, creating more relevant experiences that customers actually value while driving measurable business results.

In 2027, success won't be measured by how much AI a brand uses–everyone will use it one way or another. It'll be measured by who experiences a step change in revenue from personalizing every CRM touchpoint, and who doesn’t.

Orchestration.

Retailers are not short on technology. They are investing in AI, data platforms, cloud applications, store systems, and new customer experiences. But adding more technology does not automatically make retail easier to run. Sometimes it just creates more moving parts.

That matters more as consumer AI and agentic commerce start to shape how people shop. AI can help build a basket, check availability, recommend a product, or tell a customer an order will be ready at a certain time. But customers do not care how smart the AI is if the store cannot follow through. If AI says an item is available, the item needs to be there. If it says an order is ready, the order needs to be ready.

That takes trusted data, reliable systems, accurate inventory, enabled associates, and workflows that turn signals into action. Retailers should refocus pilots from proving a use case works to proving the business can scale the process, data, systems, and store execution behind it.

The retailers that win will be the ones that can connect the promise made to the customer with the action required to keep it.

Choosing to treat AI not as a branding exercise, but as a rigorous operational discipline, will help companies earn that trust.

Discover that agentic commerce doesn't fully control its own destiny. Ecommerce, like every industry that doesn't own the underlying models, runs on rented intelligence, and the rent is set by forces outside retail entirely: compute supply, model access, and competition for tokens from industries willing to pay far more for them. Agentic experiences will be shaped by what inference costs, not by what's technically possible. Expect agents that economize, leaning on cheap retrieval over expensive reasoning, and rewarding the brands who make themselves easy to understand. The retailers who win won't be the ones with the most ambitious AI roadmap. They'll be the ones who assume the agent is on a budget and make themselves legible anyway.

My prediction: the retail industry will keep getting squeezed! private label winning, baskets shrinking thanks to GLP-1s, and shoppers ditching big-box for small and fast. It’s everything all at once!

A visit must be earned, and the experience has to justify leaving the house. Retailtainment and immersive retail sampling experiences will continue to drive consumer engagement. Captivating experiences, rather than simple consumption, become the draw, as surface-level engagement isn’t enough. Consumers will continue to value experiences and discovery in-store, even as e commerce grows, and retailers must make physical locations work harder as destinations for innovative experiences and connection.

Making every operational promise (be it in-stock availability, delivery date, return performance) verifiable by machines, not just convincing to humans, because AI agents don't buy the pitch–they buy the proof.

What we expect product data to do, and how we expect organizations to manage and orchestrate it.

For most of the last decade, product information management meant getting accurate data into the right format and to the right places. That's table stakes now.

By 2027, the companies setting the pace won't just have clean product data–they'll have product data that is ready to be consumed by AI: structured, contextual, traceable, and governed well enough that an AI system can act on it, possibly without a human in the loop.

That shift changes the role of the platforms that manage product information.

A system of record answers the question: do we have this data in a consistent format?

A system of work answers the question: is this product ready–launch-ready, marketplace-ready, AI-ready–right now, across every channel it needs to reach and can we do launch faster?

By 2027, the industry's most capable operators will be running product content operations with people and agents, not just storing product data.

For retailers and manufacturers, this matters because their partners are being asked to meet higher data standards across a growing number of channels and AI-driven touchpoints. The organizations that help their partners get there–and build the internal infrastructure to keep pace themselves–will move faster, reduce friction in supplier relationships, and convert more buyers.

Start marketing supply chain resilience directly to customers through in-stock guarantees, transparent delivery windows, and "we'll deliver no matter what's happening in the world" messaging, much like free two-day shipping became table stakes a decade ago. Consumers have become sharply more attuned to stockouts and delivery misses after years of visible disruption, and loyalty is starting to hinge on reliability as much as price or assortment. But only retailers with real-time visibility into their own order-to-delivery processes will be able to make that promise credibly—you can't market a guarantee you can't actually see yourself keeping.

See the impact of GLP-1 medications become impossible to ignore in grocery sales data, while many retailers continue searching for explanations in the wrong places.

Category shifts are already underway, but purchase data alone won't explain what's happening. Nearly four in ten grocery shoppers say they are taking, have taken, or are considering GLP-1 medications. Among those consumers, 63% are buying fewer sugary snacks, 59% fewer salty snacks, and half are purchasing fewer sugary beverages, while 61% are buying more high-protein foods and 60% are choosing more small-portion products.

The retailers that understand these changes won't simply react by cutting space in declining categories. They'll talk to shoppers to understand how these medications are reshaping eating habits, meal planning, and shopping missions, then adjust assortments, merchandising, and private label accordingly.

The biggest mistake retailers can make over the next year is assuming this is just another short-term diet trend. It's a behavioral shift with measurable category implications. The companies that combine shopper insight with sales data will identify the opportunity early. Everyone else will spend the next 12 months explaining unexpected volume declines instead of adapting to the consumers driving them.

The Empowered Consumer: How AI Agents Are Reshaping the Digital Marketplace.

The landscape of digital commerce is undergoing a quiet revolution. Increasingly, customers are delegating the initial legwork of shopping to AI-powered bots. The digital agents are defined by patience and scalability that human shoppers simply cannot match. These assistants can analyze 10 to 15 different shopping sites in the time it takes a consumer to brew a cup of coffee, aggregating data and comparing options before a final decision is ever made.

Importantly, this shift does not diminish human agency; rather, it amplifies it. The consumer remains firmly in command, setting the parameters and making the final choice, but they now do so armed with a depth of pre-processed intelligence previously unattainable. Behind the scenes, many of these sophisticated agents rely on the Brave Search API to navigate the web. As one of the few remaining independent search indexes, Brave provides the unbiased, privacy-focused foundation these bots need to determine which merchants deserve a visit, ensuring that the path to purchase is driven by data integrity rather than paid placement.

The single-channel grip finally breaking. The value-driven shopper already moves across every banner and screen: researching at home, passing screens on the way, starting at one store and finishing at another. Their promotions do not move with them. Each deal sits locked inside one retailer's ecosystem, so a brand can promote hard and still miss the shopper at the store they walk into. Off-site is now 62.6% of US retail media display spend, growing nearly 30% year over year (eMarketer, June 2026). In 2027 the promotion follows the shopper, not the store. The brands that win will stop buying one retailer's audience at a time and follow the shopper across all of them, rendering the live, local deal wherever they are. Retail-agnostic stops being a feature and becomes the brand's default. Measurement is table stakes now, so the edge is proving it fast, accurately, and independently, in time to act, not the ROAS a platform marks for itself. Off-site is no longer the experiment; it is the engine. The defining question of 2027 is who follows the shopper across every channel, and who is still buying one banner at a time.

… the shift from AI agents that recommend a purchase to agents that actually complete one.

In my conversations with B2B suppliers and buyers this year, that distinction comes up again and again. An agent can afford to be approximately right when it is suggesting a product. When it is placing a real order, approximately right is not good enough. It has to get the price, the approval, and the source right every time. When a person makes that mistake, we usually catch it. When an agent makes it, the order goes through and no one notices until it has become a problem.

That is why I do not believe 2027 will belong to whoever builds the most impressive agent. It will belong to the B2B companies that invested in the infrastructure underneath it, the governance that keeps an agent from buying without authorization, pricing off stale data, or purchasing from a competitor when it should have come to you. Get that foundation right and agents can open up real growth. Get it wrong and you are losing revenue, and hard-earned trust, faster than any team can respond.

Everyone will be watching what the agents can do. The more important question is what is governing them.

Protecting margin with the same intelligence they use to drive revenue.

Retail has invested heavily in understanding the customer, optimizing pricing and forecasting demand. The next opportunity is applying AI behind the scenes, where fragmented vendor data, complex contracts and millions of transactions can allow small errors to quietly become significant losses.

In 2027, leading retailers will move beyond periodic audits and use agentic AI to continuously examine purchasing and payment activity, identify anomalies before money leaves the business and trace each issue to its root cause. This will not replace finance teams. It will give them the visibility and speed to focus on the exceptions that matter.

As consumers remain value-conscious and operating costs stay under pressure, profitable growth will depend not only on selling more, but on keeping more of the revenue already earned.

Ensuring their investments in AI come with human oversight, so that they can accelerate value while avoiding risk to operations.

As AI takes on more autonomous actions across pricing, inventory, fulfillment and customer service, the real risk is not that it makes mistakes (after all people do, too), but that the same mistake can be repeated instantly across thousands of products, orders, or locations before anyone notices. An AI agent that misreads a demand signal could slash prices across an entire category overnight, or an automated fulfillment system could reroute stock away from stores that actually need it. These decisions made in seconds by the AI can have long-lasting repercussions.

Retailers should therefore judge AI not only by what it can automate, but by whether they have the safeguards needed to detect, contain, and correct its decisions when something goes wrong. In practice, this can mean keeping a human in the loop for high-value or high-risk decisions, for example flagging anything above a set threshold for approval rather than letting it execute automatically, or building in rollback capability, so bad decisions can be reversed before they compound.

The most resilient retailers won’t be those whose AI agents never err, but those who can spot and isolate mistakes before any damage is done to their reputation, financial standing, customer trust, or compliance. The retailers with effective guardrails in place will be the ones able to move fast with confidence.

Closing the gap between how fast they can build and how well they can trust what they've built.

As agentic AI enters every part of the ecommerce development process, from AI coding assistants speeding up development to an uptick in personalization, generative content, and AI-powered features, the surface area for bugs and quality issues is only increasing, creating a validation gap. Across organizations, we're seeing the number of experiments and feature rollouts climb while QA processes become the biggest bottleneck.

For brands that need to scale for the future of AI in development and AI on the storefront, building a validation harness that can "QA" almost anything on your site (merchandising use cases, accessibility testing, functional validation on every release, and Jira tickets) is going to be the biggest moat and differentiating factor. A brand releasing changes 5X faster than its competitors will be more competitive, experiment more, and win more. Validation systems are one of the biggest investments brands will have to make.

Turning data from a reporting function into a revenue-generating operating capability.

Most retailers still treat data as an input to strategy: something analysts study and hand up to decision-makers. That model is too slow and disconnected to matter by 2027. The retailers who successfully monetize data will close the loop, letting signal flow into models, models into decisions, decisions into action, and results back into the model, continuously and with minimal human handoff.

That means three shifts. Architecture must be decision-ready, not just storage-ready: unified identity resolution and real-time processing vs. a warehouse built for quarterly reporting. Decision ownership must sit inside commercial teams, not above them in a separate analytics function whose insights arrive too late to act on. Activation must move past segment-and-blast defaults toward real orchestration logic.

By 2027, this same infrastructure will also need to serve AI agents deciding on customers' behalf, raising the bar on structure further. Retailers building decisioning capability now will be ready for that. Retailers still treating data as reporting will not.

From the conversations I'm having with brands, they don't just want another dashboard, they want clarity. If a product isn't converting, they want to know what's driving it. Is it the imagery? The product title? Mobile legibility? Missing content? Retailers that can help brands quickly identify the root cause of underperformance and point them toward meaningful next steps will become much more valuable partners.

Adapting to a shopper whose first move is asking AI. We're already seeing it in the data: AI-driven traffic to retail is up nearly 400% year over year.

That doesn't mean stores are closing or search is dead. It means the discovery moment has shifted. Shoppers are increasingly asking AI before they search, browse, or buy. Brands with authentic reviews, rich product content, and strong loyalty signals are the ones AI recommends.

This isn't a crisis, it's a new acquisition channel. The brands investing in trust today will have a meaningful head start because the way AI evaluates brands isn't all that different from the way people do.

Brands are realizing that physical retail remains the most natural destination to influence shopper behavior. This will result in national media budgets shifting further down the Marketing funnel to where shoppers make actual purchase decisions. Retailers will also rebalance in-store tactics intended to monetize shopper attention and extend their RMN strategies into the physical store. The outcome is that legacy silos—between Brand vs Shopper on the advertiser side and Merchandising vs Media on the retailer side—will continue to diverge. Companies that embrace how to optimize this divergence will succeed while companies that fail to adapt will struggle.

That retailers will accelerate technology and infrastructure investments to launch and elevate third-party marketplaces as a new growth driver. Many retailers will start with a highly curated and invitation-only seller recruitment approach that focuses on brand owners and local delivery experiences, and then expand to self-serve models that introduce large-scale, cross-border sellers and assortments to their marketplaces.

AI and automation will continue to revolutionize seller onboarding, merchandising strategy, and most importantly, customer acquisition. The real advantage, however, will belong to retailers with frictionless operational systems. This includes multi-currency payouts, compliance automation, and trusted seller networks that enable local shopping experiences at global scale.

As retailers build their 2027 tech stacks, prioritizing curated supply, seller empowerment, and the ability to monetize advertising alongside marketplace gross merchandise value will be crucial. Operational excellence will be the ultimate competitive moat.

Closing the visibility gap between their online and physical operations.

Retailers have spent a decade perfecting digital analytics–they know exactly what happens on their websites and apps. Walk into their stores, and that precision disappears. They can’t tell you which display drives engagement, which staff interactions convert customers, or where operational compliance actually breaks down in real-time.

This gap isn’t just a reporting problem–it’s a competitive problem. As AI adoption accelerates, retailers who win in 2027 will be the ones who finally unify operational intelligence across digital and physical channels. One source of truth for customer behavior, staff performance, and compliance–whether that happens online or in-store.

The technology exists now. What’s missing is organizational will. Most retailers treat in-store operations as a separate data universe. By 2027, that separation will be a liability.

That’s the competitive edge in 2027.

Transforming the physical store into an intelligent engagement platform.

For years, retailers have invested heavily in digital commerce while the in-store experience has evolved much more slowly. But shoppers no longer judge stores against other stores, they compare every experience to the best digital experiences they have every day.

The goal isn't to make stores look like websites. It's to bring the right digital capabilities into the physical shopping journey: making it more seamless, transparent, rewarding, guiding, personalized and connected, while preserving the immediacy, discovery and human interaction that make physical retail unique.

The retailers that will lead in 2027 will stop managing digital and physical as separate channels. Instead, they'll build one connected commerce platform that powers shopper engagement, retail media and real-time store intelligence across the entire journey. The store will evolve from a place where transactions happen into a platform that continuously creates value for shoppers, retailers and brands.