Article
March 31, 2026 · 6 min read timeThe traditional online shopping experience is built on people searching, browsing, and comparing products according to their own criteria. Things will be different in the era of AI agents. How should businesses approach data management with the role of AI agents in commerce becoming increasingly central?
The customer journey of e-commerce is in a state of transformation. People are no longer making decisions alone. Personal AI agents are now available to both consumers and businesses, and their role in commerce is set to grow rapidly.
In this second part of our two-part article series, we explore how to optimise online stores and the growing importance of data management. The first part delved into the value of brand and emotional connection in building and maintaining customer loyalty. You can read the first part here.
The era of AI agents requires a new kind of optimisation
Soon, search engine-optimised storefronts won’t be enough. In the age of AI agents, online retailers must learn to communicate with AIs that make choices and recommendations purely based on data. This doesn’t mean the end of traditional e-commerce, but it does mean a radical shift in search behaviour.
As AI agents become more common and start acting as tailored tools capable of autonomous actions on behalf of their users, online retailers will need to provide them with semantically consistent data. Semantic data tells AI what the information means, not just what it is. When AI agents find the information they need in the right format, the initial barrier to purchase – the AI’s pre-selection process – is overcome.
When an AI agent browses an online product catalogue, it can simultaneously check details such as the country of manufacture, return rate, user reviews, supply chain details, and seller reliability. The criteria are many and may vary from agent to agent, but everything must be available in a semantic format. In short, a company’s data pathways must be more tightly integrated than ever.
”Let’s say a product page simply exclaims: ‘Top-class electric bike for city riding, users love it!’ – an agent gets next to nothing from that. However, if the same information is available in a structured format – motor power: 250 W, battery range: 80 km, weight: 22 kg, rating: 4.1/5, return rate: 3% – the agent can make product comparisons and recommendations instantly,” explains Nitor’s Business Strategist Eveliina Lakka.
The reliability of an online store is crucial in the AI agent’s purchasing process. Service providers designing AI agents want to ensure their agents recommend products from the most trustworthy sellers, so users can rely on the information they receive. For online stores, this means providing verifiable data – it’s all about building mutual trust.
If the company can’t provide the information AI agents are looking for in a semantic format, the agent may simply look at the price, place the company’s recommendations at the bottom of the list, and move on.
Control your data – control the agents
A classic challenge becomes more critical than ever in the operating environment of AI agents: data. If the company’s data is siloed, it can’t be used efficiently or combined with other data in a way that supports business operations. Consistent data management ensures that AI agents will guide consumers to the company’s products without stumbling over inconsistencies.
”Most online retailers already have many of the necessary data elements in place, but now it’s essential to ensure that this entirely new sales channel – the realm of increasingly autonomous AI agents – receives the right data feed. This starts by making sure the data is in order thoroughly and ready to serve a new purpose,” says Esa Hallanoro, Creative Director at customer and brand experience design agency Hi Shine.
For many companies, the biggest hurdle may be the rigidity of their current sales platforms, which makes it difficult to operate more flexibly.
”A modular, composable architecture is a great tool for streamlining commerce. However, before adopting a model like composable architecture, a company must ensure its data platform is reliable and internal data management processes are clearly defined. This also makes it easier to ensure that every sales channel (online store, brick-and-mortar shop, pop-up kiosk, customer app) uses exactly the same data, and business decisions are made based on the big picture, not just fragments,” adds Max Packalen, Principal Architect at Nitor.
When data and its flow are standardised across the organisation, resources can be allocated more freely to other areas, such as developing staff expertise and wellbeing, and ensuring the quality of customer service and products.
How to build an online store that is ready for AI agents?
An AI-ready online store is built on five foundations. These should not be viewed as sequential projects but as ongoing areas of development, all built upon the same foundation: machine-readable, semantically structured data.
Review data management acutely to identify issues and gaps. If your data lives in silos or drips from one system to another without consistency, you can’t make better decisions than before. In such cases, it is also impossible to communicate with AI agents as needed or give optimal responses to their queries. The blowback for poor data management will be even worse in the data-driven purchasing of the future.
Internally assess whether your e-commerce platform allows for AI agent optimisation and how smoothly this can be implemented. If your current platform is restrictive, consider an avenue such as a composable architecture to help take on new sales channels. A modular approach to e-commerce enables a controlled and structurally clear way to manage your company’s data – and to provide the necessary information to AI agents in a structured, understandable format.
Build a comprehensive 360-degree view of your customer data. This allows you to thoroughly examine how the customer experience will change with agents and what actions are required from the retailer side: how customers find their way to your products in the future, and how customer relationships can be strengthened post-purchase.
Define metrics and criteria to determine whether your data optimisation efforts are working as intended and serving AI agents properly. Monitoring agent traffic also requires measurement and tracking tools.
Grow your operational understanding by delving into the world of AI, either by using other companies’ agents or by training your own. This takes time and resources, but without a practical understanding of how agents work, it’s impossible to tailor processes to fit agent behaviour into your business strategy. As Esa Hallanoro puts it: Put on your student cap and start experimenting.
Modularity is the competitive advantage for the future of e-commerce
Composable architecture and modular online store structures are becoming increasingly valuable in the age of AI agents. A modular approach to building online stores clarifies the management of product and customer data and enables the optimisation of product catalogues for AI agents in a data-secure and consistent way. It also makes it more agile to adopt new technologies and interfaces and test different functionalities.
”I’d sum up preparation into three levels: ensuring data quality, building platform capabilities, and agent optimisation. You can work on all three at once, but if your data isn’t in order, you won’t get the full benefit from the other two,” says Satu Koivulehto, Director of Strategy Services at Nitor.
The recipe for success in the age of AI agents is built on a strong brand identity, optimising post-purchase activities, and semantic data. While AI is fundamentally changing our world, those who continue to offer clear value to both people and digital assistants without losing the human core of their brand, will find continued success in the era of agents.
In the first part of this article series, we explored the value of brand and emotional connection in building and maintaining customer loyalty. You can read the first part here.
Written by
Esa Hallanoro is a creative leader with over 25 years of experience transforming brands into immersive experiences. He bridges brand storytelling and product utility, ensuring every touchpoint resonates with users, from CX strategy to UI detail. Having partnered with iconic Nordic and global brands, Esa brings a dual perspective: the emotional impact of consumer marketing and the precision of B2B ecosystems.
Eveliina Lakka is a Business Strategist at Nitor. She has ten years of hands-on experience in building digital marketing capabilities, leveraging the newest martech tools, and bringing marketing and business closer together. In her free time, you can find Eveliina lacing up her running shoes, diving into a good book, or planning her next hiking adventure.
Max Packalen has extensive experience at the intersection of technology and business, having worked in diverse roles within the IT industry for 30 years. He is an adaptable architect who understands the challenges of both the consumer and B2B sectors. In recent years, his focus has been on modern marketing automation and tools, particularly from a composable approach. Max acts as a bridge between technology and business and approaches new technologies with curiosity.
Satu Koivulehto carries extensive expertise in business design, digital services, and customer-focused strategy. Known for her approachable style and sharp insights, Satu helps organisations turn change into opportunities and build resilient digital commerce capabilities.
