Commerce isn’t getting “more digital” so much as it’s getting more intermediated. Consumers still want confidence, convenience, and a fair deal, but the path from intent to purchase is being routed through systems that don’t belong to the brand.
AI shopping assistants compress discovery into a conversation. Retail media networks decide what gets seen and what gets funded. Analytics increasingly determines whether growth is “real” or just expensive motion. That’s why the strategic question isn’t simply “Which channel should we scale?” It’s “Who gets to sit between us and demand?”
The National Retail Federation’s NRF 2026 conference in January was a tell: the industry is no longer debating whether AI will influence shopping; it’s negotiating where recommendations and checkout happen and who owns that interface. When the interface moves, everything downstream moves with it: pricing tolerance, return behavior, service expectations, and the economics of what you can afford to promise.
The first mistake leaders make in a re-intermediated market is treating the new front door as only a marketing problem. If customers can ask an assistant for “the best option under $150 that ships fast and is easy to return,” your product page isn’t the persuasive unit anymore; your data, policies, and reliability are.
From Sites to Answers
Google’s move to bring buy-buttons into Gemini is a representative expression of the new shape of commerce. If a shopper can complete a transaction without leaving the AI interface, your carefully optimized landing page becomes less central than your eligibility to be chosen. The competition shifts upstream. It’s no longer just about winning the click. It’s about winning the recommendation.
That recommendation won’t be driven by brand affinity or creative flair. It will be produced by systems that evaluate structured inputs: product data quality, real-time availability, pricing clarity, delivery confidence, and policy logic that doesn’t collapse when a customer asks an edge-case question. This is the operational reality of agent mediated commerce. If removing your brand name causes the system to lose confidence, you’re invisible regardless of how strong your site experience may be.
Returns And Shipping Become One Profit System
Returns are the clearest proof point. Digital Commerce 360’s returns reporting and Vogue’s look at fashion’s returns reckoning point to the same reality: free, frictionless returns were an acquisition subsidy that made sense when growth was cheap and abuse was tolerable; they’re a structural liability when fraud is rising, restocking is costly and margin is thin.
The shift isn’t just “charge a fee.” It’s a move from universal generosity to governed flexibility; segmenting who gets what experience and investing in prevention upstream so fewer purchases become returns in the first place.
Fraud accelerates that shift because it turns policy into adversarial ops. CBS’ reporting on retailers using AI to combat return fraud captures how quickly “routine” reverse logistics becomes a risk system. Once you accept that a meaningful share of returns can be fraudulent, the economic logic changes: you can’t fix it with better copy. You need detection, triage, and consequences that don’t collapse trust for everyone else.
Returns Policy Clarity And Shipping Cost Volatility
The same AI that is starting to guide discovery is also being deployed to police the back end, flagging high-risk returns and forecasting return rates while customers increasingly use chat tools to find policy answers. That makes clarity and consistency non-negotiable even as consumers remain price-aware. Retail Brew’s CX study coverage suggests shoppers will abandon after a single bad experience at rates comparable to, or higher than, abandoning over price increases.
Translation: every friction you introduce (e.g. fees, tighter windows, stricter thresholds) must be paired with fewer surprises and a more uniform experience across channels. If your marketing promise says “easy returns” but your operational reality says “fees and scrutiny,” you don’t just lose a sale, you teach the next buyer (and the next recommendation engine) to downgrade you.
Meanwhile, costs you don’t control are tightening the boundaries of what “great experience” can mean. SupplyChainBrain’s breakdown of the FedEx and UPS 2026 general rate increases (GRIs) makes the point that the headline number isn’t the real story. Minimum charges, accessorials and off-cycle adjustments compound into volatility that shows up as margin leakage in every “free shipping over $X” promise.
In practice, shipping and returns are converging into one system design problem: you’re engineering a set of promises that must hold up under cost swings and abuse without feeling punitive to good customers.
Measurement Becomes The Battleground
If you could measure perfectly, you could manage that tradeoff cleanly. But measurement is the other place intermediation bites, and it’s why retail media is having its maturity moment.
Street Fight frames 2026 as the year experimentation ends and incrementality becomes non-negotiable. Marketing Dive’s coverage of IAB’s Project Eidos makes the infrastructure case: the industry is pushing toward interoperable standards for outcomes, attribution, and incrementality because the current patchwork can’t keep up with cross-channel reality. When every network offers its own closed loop, brands and retailers end up grading their own homework.
That breaks once retail media becomes a core commercial operating system. Then regulation arrives to make the whole stack auditable.
Governance Makes Automation Auditable
Mayer Brown’s summary of California’s CCPA regulatory updates, especially around automated decision-making, cybersecurity audits, and risk assessments, signals where governance is headed: personalization and AI decisioning are no longer just product features; they’re compliance surfaces.
If you’re using automation to decide who sees which offers, which experiences get throttled, or which returns get flagged, you may need to explain those decisions and certify controls. This pushes commerce operators toward privacy-by-design measurement, tighter vendor oversight, and decision logs that were never part of the ecommerce playbook.
Put the signals together and a pattern emerges: the “growth stack” is turning into a governed system.
Agents and retail media compress discovery and raise the stakes of being machine-readable. Returns and shipping volatility force experience promises to be margin-aware. Fraud turns policies into risk management. Measurement standards become the battlefield for budget credibility. And regulation makes automation accountable.
Even resale fits: Digital Commerce 360’s reporting on online resale growth highlights secondhand becoming a primary option, changing the reference price in the customer’s head and introducing a competitor that monetizes the lifecycle of your product, not just the first sale.
The strategic move for 2026 isn’t to chase every new surface or tighten every policy at once. It’s to decide what you want to be true in an intermediated market: what you can prove, what you can afford, and what you won’t compromise. If your data can be trusted by machines, your economics can withstand volatility, your measurement can stand up to skepticism, and your automation can survive governance, you’ll be resilient no matter which interface becomes the front door next.
The Net Effect
For CEOs
Intermediation is a margin and dependency problem: the interface that wins attention will increasingly dictate cost-to-acquire, cost-to-serve, and the rules of attribution. Treat returns, shipping, and retail media as one system that either protects profitability or quietly taxes it.
For CMOs
The promise must match the reality more tightly than before, because both consumers and recommendation engines learn from friction and disappointment. Push for fewer surprises, clearer policy communication, and measurement that can prove incrementality across the surfaces where discovery is shifting.
For CDOs
Data readiness is now commercial readiness: machine-readable product truth, policy clarity, and decision logs matter as much as site speed. Build governance that can withstand audits while still enabling automation in returns, personalization, and performance measurement.
References
Why agentic commerce stole the spotlight at NRF — Retail Brew
Ecommerce Trends: Why AI choices, charges, friction matter for returns — Digital Commerce 360
Customers are more likely to forgive high prices over a bad customer experience: study — Retail Brew
IAB seeks to standardize interoperable media measurement — Marketing Dive
Updates to the CCPA Regulations: What Businesses Need to Know Now About Automated Decision-Making, Cybersecurity Audits and Risk Assessments — Mayer Brown
Comparing the 2026 FedEx and UPS General Rate Increases — SupplyChainBrain