AI is becoming a more practical part of ecommerce, influencing everything from product discovery and customer service to marketing, merchandising and internal operations.
The opportunity isn't simply to automate as much as possible. Retailers need to decide where AI genuinely improves the customer experience or removes unnecessary work, and where human judgement, creativity and expertise remain more valuable.
That distinction is becoming increasingly important. Adobe's 2026 research found that 49% of customers would use AI for personalised product recommendations and 44% for instant customer service, while 70% said personalised offers and recommendations should still feel human rather than robotic.
The most effective approach is therefore less about AI versus people and more about using each where they add the most value.
When to automate in ecommerce
AI and automation are particularly useful for tasks that are repetitive, high-volume or dependent on analysing large amounts of data. Automating these processes can free teams to spend more time on work that requires judgement and creativity.
McKinsey found that 90% of the retail executives it surveyed had begun experimenting with generative AI, with customer service and internal value chains among the areas seeing the most experimentation.
Customer support with AI
Routine customer service queries are a natural starting point for automation. AI assistants can answer common questions, provide order updates, explain delivery options and direct customers towards relevant information.
The important distinction is that automation shouldn't mean removing human support altogether. Customers should have a straightforward route to a person when a question becomes more complicated or requires discretion.
Adobe found that 66% of consumers expect quick, efficient support through automated systems such as chatbots or virtual assistants, but only 33% feel brands currently deliver this effectively.
Product recommendations and discovery
AI can analyse browsing behaviour, previous purchases, product attributes and other signals to make product discovery more relevant.
This can be particularly useful for large catalogues, where manually determining which products to surface for every customer isn't practical.
AI-powered recommendations can also help shoppers compare products, discover complementary items and narrow down a large selection.
Marketing and email automation
AI can support marketing teams by identifying customer segments, suggesting content, analysing campaign performance and triggering communications based on customer behaviour.
Automation is particularly useful when the same process needs to happen repeatedly. An abandoned basket, for example, can trigger a predefined sequence without a marketer having to manually monitor every customer journey.
The role of AI can go further than simply sending emails. Retailers are increasingly using it to personalise offers and content across different customer touchpoints. Adobe's retail research found that personalised offers and promotions based on customer data were a priority for 51% of retailers surveyed.
Inventory and operational forecasting
AI can also support the operational side of ecommerce. Demand forecasting can identify purchasing patterns and help retailers anticipate changes in demand, while automated processes can help flag low stock or recommend when replenishment may be required.
This is particularly useful for retailers managing large catalogues or multiple sales channels, where manually monitoring every product becomes increasingly difficult.
However, AI-generated forecasts should support operational decision-making rather than automatically determine every purchasing decision. Unexpected events, supplier issues and changes in customer behaviour can all affect demand.
When to keep the human touch
Not every customer interaction benefits from automation. The more complex, sensitive or commercially important the interaction, the more valuable human judgement can become.
Trust is particularly important as AI becomes more visible in the shopping journey. Adobe found that 87% of consumers expect retailers to handle their personal data responsibly and securely, while 74% consider transparency around AI-generated content, recommendations or images important.
Complex customer issues
AI can handle straightforward questions well, but complicated complaints often require context, discretion and empathy.
A customer dealing with a defective product, delivery problem or unusual return shouldn't have to repeatedly explain their situation to an automated system before reaching someone who can help.
Complaints, disputes, unusual orders and sensitive issues are best escalated to a member of the customer service team who can understand the circumstances and decide on an appropriate resolution.
Brand storytelling and creative direction
AI can help teams produce content more efficiently, but a brand's identity shouldn't be reduced to automatically generated copy.
Positioning, tone of voice, creative direction and campaign concepts require an understanding of the brand, its customers and the market it operates in. AI can support the process, but human input remains important in deciding what a brand should actually say.
This is especially relevant as customers become more aware of AI-generated content and increasingly expect transparency around its use.
AI can be used to support research, idea generation or first drafts while brand strategy, creative direction and final approval are best left with the relevant team(s).
High-value customer relationships
For high-value products, complex purchases and B2B ecommerce, customers may need more than an automated recommendation.
A knowledgeable member of a sales or account team can understand the customer's requirements, answer questions that fall outside standard product information and recommend an appropriate solution.
AI can still play a supporting role by surfacing customer information or relevant products, allowing the person handling the relationship to spend more time on the conversation itself.
Strategic decision-making
AI can analyse data and identify patterns, but it doesn't understand a retailer's commercial priorities in the same way an experienced team does.
Decisions such as entering a new market, changing a brand's positioning, launching a new product range or investing in a new sales channel require context that extends beyond the available data.
Teams can leverage AI to identify trends, analyse performance and provide evidence to help them make better-informed strategic decisions.
The goal isn't to automate everything
The most useful question isn't "Can we automate this?". It's "Would automating this make the experience or process better?"
For many retailers, the answer will be yes when a task is repetitive, predictable and high-volume. It may be no when the interaction requires empathy, creativity, judgement or a deep understanding of the customer.
A useful approach is to:
Automate repetitive work where speed and consistency matter.
Use AI to surface insights that help teams make better decisions.
Keep humans involved when interactions are complex or commercially important.
Review AI outputs rather than assuming they are automatically accurate.
Be transparent about where AI is being used, particularly when it affects customer experiences or recommendations.
This doesn't mean AI and human interaction need to operate as separate experiences. The strongest implementations often combine the two, allowing AI to handle the groundwork while people step in where their expertise makes the biggest difference.
How Aero approaches AI
At Aero, we see AI as a tool for supporting developers rather than replacing them.
For ecommerce agencies, some of the most time-consuming parts of a project aren't necessarily the development itself. Understanding an existing environment, checking dependencies, working through configuration, troubleshooting issues and documenting technical information can all take valuable developer time.
The Aero AI Playbook takes a practical approach to these tasks. It helps agency partners assess their development environment, identify what's missing, work through setup requirements and support troubleshooting to support faster and smoother development cycles.
The aim is simple: remove repetitive work so developers can spend more time on the work that requires their expertise.
That could mean more time spent solving a retailer's specific requirements, developing bespoke functionality, improving the customer experience or working on the technical challenges that create genuine value for a project.
This reflects a broader principle in how we approach AI at Aero. We don't see value in using AI simply because it's available. The focus should be on finding practical applications that make ecommerce development more efficient while keeping technical judgement and control with the people doing the work.
AI should make ecommerce better, not simply more automated
AI is already changing how people discover products, interact with retailers and complete everyday tasks. Salesforce's 2025 holiday data, for example, found that AI and agents influenced 20% of global retail sales during the period through personalised recommendations and conversational customer service.
But increased adoption doesn't mean every part of ecommerce should become automated.
The retailers and agencies that get the most from AI will be those that understand where it can remove friction, improve relevance and reduce repetitive work while preserving the human expertise that makes a brand trustworthy and distinctive.
Want to explore how Aero can support a more efficient ecommerce operation?
Book a demo to see how Aero's developer-first platform and approach to AI can help your team work more efficiently without compromising control or flexibility.
