In a guest piece for InternetRetailing, Radu Săndulescu – Data Analytics & AI Services Director, Zitec, explores how the fast-changing AI landscape is impacting on retailers – and how to position your organisation for success.
Artificial Intelligence (AI) is everywhere in retail, from chatbots to predictive analytics, promising to change how businesses interact with customers and streamline operations. But after working with all sorts of organisations throughout the years, I’ve learned one thing for sure: AI only delivers real value when it’s embedded into the core of retail operations, with clear objectives and a structured approach.
Our intel tells us that organisations of all sizes experience a decrease in manual work and significant cost-optimisation when using AI models in areas like CX and personalisation, automation, and more. What’s more, an Accenture study shows that companies modernising operations with AI-led strategies experience 2.5x faster revenue growth, while AI-driven personalisation and process optimisation yield around 20% higher sales ROI.
The data is there. If years ago, we were discussing AI’s potential, now we’re talking about results. But getting to these results isn’t an occurrence, but rather a result of careful strategic planning.
Start with the problem, not the technology
It’s easy to get caught up in the latest AI tools and technologies, but the most important first step is to figure out what you’re trying to solve. It sounds obvious, but many retailers dive into AI without a clear understanding of the problem they want to address.
AI works best when it’s focused on high-impact areas like improving the customer experience, streamlining inventory management, or making smarter pricing decisions.
Fix your data and tech readiness
For AI to deliver results, it needs solid foundations. This means getting your data and technology in shape first.
The data challenge
Retailers have a wealth of data, but it’s often scattered across different systems and in formats that are hard to work with. For AI to do its job right, this data needs to be organised and accessible. It’s crucial to get rid of silos between ERP, WMS, e-commerce platforms, or POS systems and centralise your data in one place. Only then can AI provide meaningful insights that drive decisions.
Data quality is key. No matter how advanced an AI model is, its performance is directly limited by the quality of the data it receives.
That’s why, for us, any AI project begins with building a clean data foundation. Our process typically involves consolidating multiple sources into a single platform, standardising the data to ensure consistency, and retiring legacy systems to create a single source of truth. For example, for a major retail partner whose e-commerce platform now drives 10% of their national sales, our work involved consolidating their reporting in Looker, unifying data from various systems to provide their teams with a clear and rapid overview.
The tech side of things
Once your data is in order, the next step is making sure your tech systems are ready. Our internal findings indicate that many retailers still rely on legacy systems that struggle with modern AI tools, with 58% running on e-commerce platforms older than 5 years. This technical debt helps explain why 59% of European retail CIOs report challenges in delivering digital projects on time and on target: it’s difficult to build the future on an outdated foundation.
Outdated tech can create roadblocks down the line. If your current systems can’t integrate with AI solutions, it’s time to assess whether they need an upgrade or replacement.
Don’t forget the people
AI is often seen as a technical challenge, but in reality, one of the biggest obstacles is organisational readiness. AI should be seen as a tool to help your team, not replace them. The key is getting people on board and ensuring they understand that AI is here to make their jobs easier.
Test, don’t guess. Start small
It’s easy to get carried away with big AI promises, but the best approach is to start small, with a clear goal and a willingness to learn along the way. Our experience tells us that a Proof of Concept (PoC) can help you test an AI solution in a controlled environment, so you can assess its impact before rolling it out across the entire business and making big investments.
A PoC should not only aim to prove that AI can work but also measure its potential ROI. Take, for example, a leading skincare brand we worked with. They used AI to recommend personalized products based on selfie analysis. The solution was possible by running a PoC using Google Gemini Pro. The goal was to validate whether users would trust AI to suggest skincare tailored to visible conditions, and we did.
Pick the right AI technology
Choosing the right AI models can feel overwhelming, with so many options available. The key is to choose a solution that fits your business needs and partner with the right technology partner that can support you in your AI journey.
AI implementation isn’t a one-time fix, since it requires constant monitoring, adjustment, and iteration.
As regulations like the EU AI Act evolve, retailers must ensure their systems are both effective and transparent. Low-risk tools like chatbots should disclose they’re AI-driven, while high-risk systems need detailed logging and documentation for accountability.
Beyond compliance, addressing bias and fairness is essential. Regular audits, diverse datasets, and transparent decision-making help prevent discrimination and build trust in AI systems.
Stay informed
Our editor carefully curates two newsletters a week filled with up-to-date news, analysis and research. Click here to subscribe to the FREE newsletter sent straight to your inbox. Why not follow us on LinkedIn to receive the latest updates on our research and analysis?
