

AI assistants are changing how customers discover and choose products, while shopping journeys continue to span online and physical stores. To prepare for 2027, businesses need to adapt their presence to these new realities and ensure their data and operations can support a unified, frictionless customer experience.
We've selected 25 eCommerce and unified commerce trends, along with practical examples to help you assess their relevance to your business. Retailers, DTC brands, B2B manufacturers, and distributors will find insights to guide their investments in 2027, as well as five emerging signals to monitor through 2030.
Three major shifts are converging: AI assistants are playing a growing role in product discovery, expectations for seamless experiences across channels continue to rise, and profitability requires closer coordination between marketing, customer experience, and operations.
According to Novatize's 2026 eCommerce Trends Report, based on a Léger survey, 52% of Canadians aged 18 and older say they have already used an AI tool while shopping online, compared with 39% in 2025, a 13-percentage-point increase. Additionally, 63% say they are willing to use these tools. Adobe has also reported growing traffic from AI assistants to U.S. retail websites. These figures illustrate an ongoing transition, but they do not mean fully autonomous transactions have become the norm.
For more context on these trends in the Canadian market, explore Novatize's 2026 eCommerce Trends Report. The Léger survey was conducted from May 8 to 10, 2026, among 1,525 Canadians aged 18 and older. The results were weighted to represent the Canadian population.

Agentic commerce refers to experiences in which an AI agent can assist with, and sometimes execute, stages of the purchasing process: defining a need, comparing options, checking availability, or initiating a transaction.
It's important to distinguish between AI-generated recommendations and purchases actually completed by an authorized agent. In 2027, the first is already playing an important role, while the second remains unevenly adopted across industries and markets.
Practical example: For an outdoor retailer, document a tent's weight, capacity, rain resistance, and intended uses. A request such as "I'm looking for a lightweight tent for two people on a three-day hiking trip" should be matched to these criteria without the system having to make assumptions.
To learn more, read our article on agentic commerce, UCP, and their practical implications for merchants.
Generative Engine Optimization (GEO) aims to improve the visibility and accuracy of a company's information in AI-generated responses. It complements SEO rather than replacing it. Detailed product pages, well-structured buying guides, and clear answers to frequently asked questions can help AI assistants understand your offerings and recommend them when they match a customer's needs.
The emergence of advertising offerings within AI models, including those recently announced by ChatGPT, introduces another opportunity. These ads may be tailored to the context of a conversation, clearly identified as sponsored, and displayed separately from organic responses. Purchasing an ad placement does not influence the assistant's organic recommendations.
For merchants, this channel is worth evaluating based on market availability, purchase intent, and measurable results.
Practical example: For a furniture retailer, test a question such as "What sofa should I choose for a small living room with a cat?" Check whether your products appear and whether the information presented is accurate, then improve the relevant product pages and buying guides. At the same time, when advertising placements are available in your market, test an ad related to this need, with a relevant landing page, and measure the traffic and sales it generates.
Related reading: Our SEO vs. GSO article explains how GEO complements your eCommerce search optimization strategy.
Are your products appearing in AI assistant responses? Explore our GEO and product discoverability services to identify opportunities for improvement.
Product catalogs must be understandable to search engines, commerce platforms, and AI assistants. Vague titles and marketing descriptions are no longer enough.
Specifications, dimensions, materials, compatibility, certifications, variants, and usage limitations all have direct value. For an industrial distributor, a well-documented part number can make the difference between a relevant recommendation and an incomplete answer. Product Information Management (PIM) and data governance are becoming important drivers of commercial performance.
This requirement reflects a measurable expectation. According to the 2026 Novatize–Léger report, 44% of Canadians would be willing to use AI to compare products or brands, or understand the differences between similar products. This figure represents willingness to use AI, not actual usage. For merchants, it highlights the importance of documenting the criteria that help customers compare options, including dimensions, compatibility, applications, and technical specifications.
Practical example: For a range of replacement parts, add compatible models, dimensions, and verified equivalent part numbers. A product page stating "compatible with models A and B, 25 mm diameter" is more helpful than a generic description such as "durable and versatile part."
The emergence of standards such as the Universal Commerce Protocol (UCP), introduced by Google and industry participants in 2026, aims to facilitate exchanges between shopping interfaces, merchants, and payment services.
Their impact will depend on actual adoption, available integrations, and security controls. For businesses, the immediate priority isn't to implement every new protocol, but to avoid building catalogs and architectures that cannot connect to emerging technologies.
Practical example: Before connecting an AI assistant to your online store, test a specific scenario: an item goes out of stock between the recommendation and payment. Your system should be able to recheck availability and request confirmation if the price or delivery option changes.
The value of AI agents also extends behind the scenes. They can support catalog enrichment, request classification, response preparation, anomaly detection, and data reconciliation.
Unlike a traditional conversational assistant, an operational AI agent can perform multiple actions within a defined workflow. The potential efficiency gains are significant when data is reliable and exceptions are properly managed. However, poorly controlled autonomy can also multiply errors.
Practical example: For a small selection of products, have an AI agent extract dimensions and materials from supplier documentation. The agent suggests fields to complete, and a team member validates the information before publication. Keep missing values marked as unavailable rather than allowing the tool to invent them.
Is a manual process slowing down your team? Explore custom commerce AI agents and the tasks they can support across your operations.
AI should be managed both as an acquisition channel and as an operational capability.
On the acquisition side, teams can monitor product visibility, citation quality, sessions referred by AI assistants, and associated sales, while accounting for attribution limitations.
On the operational side, businesses need to track completed tasks, corrections, costs, and permissions. A clear policy governing shared data and human oversight helps build trust.
Practical example: For an assistant answering delivery-related questions, review a sample of responses to verify that it uses the correct postal code, inventory levels, and applicable delivery times. Track accurate answers, escalations to human agents, and cost per resolved request to determine whether a broader rollout is justified.
To explore the role of integrations in AI initiatives, download the MIT Technology Review Insights report presented with Novatize and Celigo: Bridging the Gap Between AI and Operations Through Intelligent Integration.

Consumers no longer search exclusively by product name. They describe problems, upload images, or ask complete questions, such as "What equipment is suitable for this application?"
On-site search engines that understand intent, synonyms, and product attributes can reduce unsuccessful searches.
This trend also affects B2B commerce, where buyers may search for a part using a partial reference number or a photo.
Practical example: In a hardware store, connect a search such as "fix a leak under my sink" to relevant categories and guides, then to compatible products based on the details provided. For a parts distributor, also enable searches using partial part numbers, with compatibility confirmed before purchase.
Rather than showing every visitor the same storefront or relying on static customer segments, merchants can adjust recommendations and content based on context: customer status, purchase history, local inventory, seasonality, or expressed needs.
Useful personalization should remain explainable and respect customer consent. Delivering a relevant experience doesn't necessarily require complete individual-level personalization. A few context-based rules can already improve product discovery.
The 2026 Novatize–Léger report shows that 41% of Canadians say they have already used personalized shopping recommendations and advice based on their profile or behaviour, while 58% say they would be willing to do so.
These findings support a practical priority for 2027: improving the usefulness of recommendations and measuring their contribution to the purchasing journey.
Practical example: Someone purchasing a coffee machine may receive a recommendation for compatible filters, while an existing owner could instead see their usual supplies. Test these scenarios separately and determine whether the suggestions make purchasing easier or simply add unnecessary distractions.
A purchase can begin with an AI assistant, continue on a smartphone, be confirmed in-store, and be completed online.
To prevent disruptions, businesses need to preserve essential information across touchpoints: viewed products, availability, shopping carts, customer accounts, and delivery options.
A seamless experience doesn't depend on a single tool, but on coordination between customer touchpoints.
Practical example: Allow customers to save products for a renovation project on their phones, then access the same list in-store or through their accounts. With the customer's consent, store associates can continue the conversation using the same product references and availability information.
As technologies multiply, the quality of the shopping journey remains essential: speed, accessibility, navigation, filters, product information, shopping carts, and checkout.
A complex website doesn't become more effective simply by adding a conversational AI agent. Merchants benefit from prioritizing improvements that eliminate real friction, particularly on mobile and in purchasing journeys involving multiple decision-makers.
Practical example: On mobile, display shipping costs and delivery options before the final checkout step. For a product available in multiple sizes, clearly show quantities and unit prices so customers can compare options without calculating prices elsewhere.
Where do your customers experience the most friction? Discover our conversion rate optimization approach to identify which improvements to test first.
Customers want to know whether a product meets their needs, when it will arrive, and what happens if something goes wrong.
Size guides, comparison tools, compatibility information, realistic delivery dates, order tracking, and proactive support can reduce uncertainty.
In B2B, actual product availability and contractual terms play a similar role. The promise made before purchase must align with the experience delivered afterward.
Customer reviews also contribute to trust. According to the 2026 Novatize–Léger report, 30% of Canadians would be willing to use AI to summarize product reviews.
Making reviews accessible and sufficiently detailed is therefore useful both for customers and for the tools helping them evaluate purchases.
For more insights, explore the role of customer reviews in agentic commerce and omnichannel growth.
Practical example: For furniture delivered to a customer's home, display packaged dimensions, delivery conditions, and return policies near the purchase button. For a technical component, clearly identify compatible products and exclusions. Address the concerns that might cause hesitation before purchasing that specific item.
To explore how shopping journeys are evolving, watch our webinar, "What Will the Online Shopping Experience Look Like in 2027?", presented with Les Affaires and Shopify.

Unified commerce connects data, systems, and processes to deliver a consistent experience online, in-store, and across assisted sales channels.
Customers don't necessarily distinguish between a POS, ERP, CRM, or eCommerce platform. They see one business.
Problems arise when prices, promotions, inventory, or order statuses differ between channels. Integration quality is also becoming a prerequisite for reliable AI deployment.
Practical example: When a promotion applies online and in-store, ensure the same price is recognized in both locations and clearly communicate any exclusions. Customers reserving a product should also have their reservation visible in the systems used by store associates.
An Order Management System (OMS) can identify the best fulfillment or shipping location based on inventory, capacity, delivery times, and costs.
The objective isn't simply to deliver faster, but to meet customer expectations while maintaining profitability.
Fulfillment rules can coordinate warehouses, stores, and logistics partners, with exceptions based on order type.
Practical example: For an order containing three items, compare shipping everything in one package from a warehouse with sending two separate shipments from stores. Configure your OMS to account for total cost and the promised delivery date. The closest location won't always be the best option.
Real-time or near-real-time inventory visibility helps businesses display reliable availability, support in-store pickup, and reduce cancellations.
This challenge goes beyond inventory tracking technology. Reservations, adjustments, returns, and synchronization delays all need to be managed.
Across a retail network, inventory accuracy directly affects the ability to deliver on local fulfillment promises.
Practical example: Display "available today at our Quebec City store" only when units are genuinely accessible. When a customer reserves the last unit, remove it from the inventory available to other shoppers to avoid promising the same product twice.
Physical stores are increasingly places for product discovery, advice, service, pickup, returns, and sometimes order fulfillment.
Effective strategies account for staff capacity, store location, and the in-store experience.
Omnichannel operations shouldn't transfer invisible costs to store teams. Businesses need appropriate processes and shared performance indicators.
Practical example: For in-store pickup, send confirmation when the order is ready, including opening hours, the appropriate entrance, and pickup instructions. Also allow associates to retrieve online purchases when processing returns, according to applicable policies.
Managing a network of retail stores? Explore our retail and unified commerce services to connect your online customer experience with in-store operations.
Demand forecasting, replenishment, and margin analysis are increasingly interconnected.
Sales growth can hide high delivery costs, returns, or excess inventory. Predictive models can help identify potential issues, but their usefulness depends on clean data and human oversight.
Merchants can begin by improving visibility into the actual contribution of each order.
Practical example: Compare two orders of equal value: one ships in a single package, while the other requires multiple shipments and a return. Include these costs in your analysis before offering the same free shipping conditions to both. You can also test a minimum order threshold or consolidated shipping option.

Business customers want to access their pricing, product availability, purchase history, and documents without waiting for a sales representative to respond.
A B2B customer portal becomes an extension of the sales function, particularly for recurring orders.
Sales teams remain essential for complex requests, but straightforward tasks are better handled through self-service.
Practical example: Give business customers access to invoices, negotiated pricing, and order tracking through their portal. Sales representatives can then focus on new requirements or technical questions instead of resending documents.
A single business account may include several buyers, budget managers, and approvers.
The transactional experience must accommodate contractual pricing, catalogs, quotes, purchase limits, and permissions.
This type of personalization is less about marketing recommendations and more about enforcing commercial rules. Proper configuration reduces errors and encourages adoption of digital channels.
Practical example: For a customer with multiple branches, allow each buyer to build a cart while requiring approval above an agreed amount. Managers can view their team's orders, while each user accesses the products and prices authorized for their account.
Do your business rules go beyond your platform's standard capabilities? Discover SPINE, Novatize's Shopify B2B Accelerator, including ERP-connected pricing, user roles, and approval workflows.
For parts, consumables, and frequently purchased products, ordering speed matters more than an inspiring storefront.
Shopping lists, quick ordering by SKU, file imports, history-based recommendations, and availability alerts can reduce the number of steps required.
However, automation shouldn't eliminate necessary validations when a product or order quantity changes.
Practical example: For a business that regularly purchases the same gloves and consumables, provide a saved shopping list containing its usual product references and sizes. The buyer adjusts quantities, reviews current pricing and availability, and confirms the purchase without rebuilding the cart each time.
B2B purchasing processes increasingly connect with procurement environments, including ERP systems, procurement solutions, PunchOut catalogs, EDI, and APIs.
The more effectively pricing, inventory, order, and invoicing data flows between systems, the less manual re-entry teams need to perform.
Future purchasing agents could rely on this same connectivity, making it essential to strengthen existing integrations first.
Practical example: A customer builds a cart in your PunchOut catalog and retrieves it in their procurement system for approval. Once approved, the order and its reference number flow into your systems according to the configured integration. The invoice then includes the reference required by the customer's accounting team.
To learn more about connecting ERP systems with the purchasing journey, watch our webinar, "Extend Your ERP Capabilities with Shopify B2B".
In B2B, buyers move between channels and interact with multiple stakeholders.
According to McKinsey, B2B decision-makers use an average of ten channels throughout their purchasing journey.
As a result, traditional sales and eCommerce can no longer be treated as competing approaches. CRM data, portal interactions, quotes, and sales follow-ups should contribute to a shared understanding of each account.
Practical example: When an account requests an online quote, make the quote and related interactions accessible to the assigned sales representative. If the customer subsequently places an order, adjust marketing communications to avoid sending a follow-up about an offer they've already accepted.
Related reading: Our analysis of how to unify B2B marketing, eCommerce, and sales, based on insights gathered from B2B leaders in 2026.

A growing share of information searches takes place in environments that answer questions directly.
Success can no longer be measured solely by search engine rankings or session volume.
Authoritative content, brand recognition, structured information, and trust signals are becoming more important. However, businesses still need to distinguish between visibility, qualified traffic, and attributable sales.
Practical example: For a category such as work boots, publish a guide comparing selection criteria based on different applications, supported by verified product specifications. Then monitor how AI assistants answer related questions and the traffic they refer to your website, alongside measurable sales.
As customer acquisition costs remain under pressure, it makes sense to focus more on second purchases, reactivation, and post-purchase needs.
Useful communications, reliable service, and benefits tailored to actual customer behaviour can improve retention.
Customer Lifetime Value (CLV) is a valuable metric, provided it is interpreted alongside margins and purchase frequency.
Practical example: For a skincare brand, test a replenishment reminder based on the product's estimated usage period, with an option to adjust reminder frequency. Exclude customers who recently reordered and measure whether the reminder generates profitable repeat purchases.
Consumers compare prices, but they also consider delivery times, quality, convenience, service, and trust.
In its 2026 Global Retail Industry Outlook, Deloitte highlights the importance of these non-price factors in perceived value.
For merchants, this means communicating more clearly what makes an offering worthwhile and avoiding excessive dependence on discounts.
Practical example: For a coat priced above the market average, explain what justifies the difference: materials, repairability, warranty, and service, when these benefits genuinely apply. Make this information visible on the product page rather than relying exclusively on a promotional code.
Complementary services, relevant subscriptions, replacement parts, maintenance, and trade-in programs can extend the customer relationship.
In B2B, the customer portal can become an entry point for services and spare parts ordering. In B2C, post-purchase content and related offers can encourage repeat purchases when they address genuine needs.
Growth then depends as much on service quality as on acquiring new customers.
Practical example: After selling equipment, recommend compatible parts, consumables, or appropriate maintenance services. Clearly explain frequency and conditions, then allow customers to choose a reminder or recurring option rather than enrolling them automatically.

These developments are less certain or less widely adopted than the previous 25 trends. They can inform market monitoring and experimentation, but shouldn't be considered guaranteed predictions or equally important priorities for every merchant.
Over the medium term, some consumers may delegate routine purchases to AI agents based on predefined budgets, preferences, and rules.
The main challenge won't simply be whether technology can compare offers, but how to ensure control, consent, security, and error management.
Scenario to watch: One potential application is allowing an AI agent to replenish consumables within a set budget, with confirmation required if the brand or price changes. The important signal will be whether customers repeatedly use these features and can easily cancel or correct orders.
Voice, visual, or connected-device interactions could make it possible to place orders without opening a traditional online store.
Websites would remain useful, but might no longer be the visible interface for every transaction.
Reliable identities, catalogs, and purchase confirmations will be essential.
Scenario to watch: Imagine someone in a workshop verbally requesting a replacement part while keeping their hands free. Watch for experiments in which systems can unambiguously confirm the part number, quantity, and price, as well as evidence of repeat usage beyond an initial trial.
Products could be accompanied by verifiable data about their origin, composition, repairs, and end-of-life management.
Depending on the industry and regulatory environment, this information could simplify maintenance, resale, or compliance processes.
Data collection costs and interoperability remain obstacles.
Scenario to watch: For a piece of furniture, a digital product passport could include materials, care instructions, and available replacement parts. Monitor projects where this information is actually used to repair or resell products, along with requirements applicable to your product category and markets.
Digital twins create virtual representations of store networks, warehouses, or supply chains to test scenarios.
Combined with predictive capabilities, they could help anticipate the effects of supply disruptions, volume changes, or new facility deployments.
Their value will depend on model accuracy and maintenance costs.
Scenario to watch: A retailer could simulate adding in-store pickup across several locations before expanding the service throughout its network. Initial results to examine would include impacts on staff workload and fulfillment times, followed by differences between the simulation and a limited real-world pilot.
As commerce becomes more automated, some customers may increasingly value straightforward experiences and accessible human support.
Low-tech doesn't necessarily mean abandoning AI. A business can automate internal tasks while offering simple communications, expert guidance, and service without unnecessary technological barriers.
Gartner reported in 2026 that 87% of surveyed customers consider access to a human essential when a company uses AI for customer service.
This is a concrete indication of the tension between efficiency and human connection.
Scenario to watch: Test direct access to a person for complex purchases or a short, downloadable buying guide for customers who prefer that format. Observe whether these options simplify decisions and reduce abandonment before introducing another automated interface.
Launching 30 initiatives at once is neither necessary nor realistic.
A useful first step is to evaluate opportunities based on customer impact, feasibility, data maturity, and potential profitability.
For most organizations, high-quality product data, a seamless customer experience, and reliable information across channels provide particularly valuable foundations.
Next, select one or two use cases to test with straightforward performance indicators.
For a retailer, this might mean improving inventory reliability and order orchestration. For a DTC brand, priorities could include product content, GEO, and customer retention. For a manufacturer or distributor, the focus might be B2B self-service, technical product data, and ERP integrations.
The goal isn't to adopt every new technology, but to build an organization capable of evolving alongside customer expectations.
Manufacturers and distributors can explore our analysis of digital investment priorities to connect these trends with budget decisions.
Commerce in 2027 will be shaped by AI, but also by very practical expectations: finding the right product, accessing reliable information, purchasing without friction, and receiving the service promised.
Businesses that invest in their data, integrations, and customer journeys will be better positioned to take advantage of future innovations.
Over the longer term, maintaining meaningful human interactions could become just as strategically important as automation.
At Novatize, we help retailers, manufacturers, and distributors evolve their eCommerce and unified commerce experiences, from strategy and technology implementation to continuous optimization.

Would you like to discuss this with a Novatize specialist?

Pierre-Olivier Brassard
Sources
Novatize — 2026 eCommerce Trends Report, Léger survey of 1,525 Canadian adults, conducted May 8–10, 2026. AI adoption: pp. 5 and 11; personalized recommendations: p. 7; motivations for use: pp. 18 and 19. Percentages related to motivations are based on the general population, with multiple responses allowed.
Novatize — 25 eCommerce Trends to Watch in 2026
McKinsey (October 28, 2026) — Who Wins When AI Does the Shopping?
Adobe (October 29, 2026) — AI Traffic Now Generates Higher Revenue per Visit
Google Developers (January 11, 2026) — Universal Commerce Protocol
McKinsey (May 28, 2026) — Global B2B Pulse
Deloitte (January 8, 2026) — Global Retail Industry Outlook
Gartner (August 4, 2026) — Access to a Human Agent in AI Customer Service





