Retailers looking to adopt artificial intelligence may be asking the wrong first question. Instead of deciding which AI platform to buy, the more urgent task is preparing the data that will power it. That was the consistent message from executives speaking at the RETRAK Retail Summit 2026 in Nairobi, where the discussion focused on how data readiness before AI can help retailers make better commercial decisions across pricing, inventory, customer engagement and payments.
The panel, moderated by digital strategist Moses Kemibaro, brought together leaders from Network International, Canvas Cosmetics, LOOP DFS at NCBA Group and Compulynx. Although each speaker approached the topic from a different part of the retail ecosystem, their conclusions pointed in the same direction: businesses already collect vast amounts of information, but much of it never reaches the people making day-to-day decisions.
AI Starts With Better Business Data, Not Better Models
Artificial intelligence has become the centrepiece of many technology roadmaps, yet several panelists argued that retailers risk disappointing results if they introduce AI before fixing the underlying quality of their business data.
Eric Muriuki, Group Director of Digital Business and CEO of LOOP DFS at NCBA Group, described this as a question of data readiness rather than AI readiness.
Every retailer, he noted, already generates operational information through sales, inventory, suppliers, finance and customer interactions. The challenge is whether those records are properly organised, labelled and connected so AI systems can understand them.
Without that foundation, businesses simply replace one problem with another. Dashboards become chatbots, but the answers remain incomplete because the underlying information is fragmented.
That thinking moves the AI conversation away from software selection and towards operational discipline.
Retailers Already Have the Information They Need
One of the strongest themes throughout the discussion was that retailers are not suffering from a shortage of data.
Point-of-sale systems, online stores, loyalty programmes, mobile applications, social commerce, delivery platforms and payment systems all produce continuous streams of commercial information.
The problem is that those systems often operate independently.
A retailer may know what a customer bought in-store but have no visibility into what that same customer viewed online before making the purchase. Marketing teams may measure campaign performance, while finance teams monitor margins and operations teams track stock levels, with each department working from different datasets.
That fragmentation limits the value of AI because no single system captures the complete picture.
The discussion suggested that creating a unified operational view may produce greater commercial benefits than investing in another standalone analytics tool.
Payments Are Becoming a Source of Business Intelligence
For many retailers, payments remain a necessary business expense.
Judy Waruiru, Regional Managing Director at Network International, argued that this perspective overlooks one of retail’s richest sources of customer intelligence.
Every digital payment reveals purchasing behaviour, preferred channels, transaction timing and customer movement across physical and online stores.
When combined with inventory, loyalty and sales information, payment data can help retailers identify customers at risk of leaving, monitor fraud, understand channel preferences and measure the effectiveness of promotions.
It also provides a broader market perspective.
Because payment providers process transactions across multiple merchants and sectors, they can identify wider consumer trends that individual retailers cannot observe from their own businesses alone.
Customer Data Is Still Too Fragmented
The conversation repeatedly returned to one operational challenge.
Retailers interact with customers through websites, stores, social media, mobile applications and digital payment channels, yet those touchpoints rarely feed into a single customer profile.
Without that unified view, businesses struggle to understand the full customer journey.
That affects more than marketing.
Product recommendations become less accurate, loyalty programmes lose effectiveness, pricing decisions rely on partial information and customer churn often becomes visible only after shoppers have already moved elsewhere.
For Sonal Haria, Co-Founder and CEO of Canvas Cosmetics, customer insight also influences product development.
She explained that early market research helped the company identify strong demand for complexion products, leading to investment in research and development before launching more than 30 products in that category.
As the business expanded, internal customer behaviour became more valuable than relying solely on industry benchmarks, allowing product decisions to reflect the preferences of its own community.
AI Works Best When It Solves One Business Problem
While AI dominated the discussion, none of the speakers recommended launching broad enterprise programmes from the outset.
Instead, they advocated for focused implementation.
Retailers should begin with a specific operational challenge such as stock replenishment, pricing, fraud detection, sales outreach or customer retention.
Once measurable improvements are achieved, businesses can expand AI into other parts of the organisation.
Siddesh Narkar, Head of Product at Compulynx, also encouraged retailers to pay closer attention to inventory profitability rather than focusing only on sales growth. Metrics such as Gross Margin Return on Inventory Investment can provide a clearer picture of whether inventory is generating sustainable returns instead of simply driving revenue.
That approach reflects a broader theme from the panel: AI delivers value when it improves commercial outcomes rather than adding another layer of technology.
The Next Retail Customer Could Be an AI Agent
The discussion also looked beyond today’s retail environment.
Waruiru suggested that future shopping journeys may involve AI agents acting on behalf of consumers.
Rather than manually comparing products across multiple websites, customers could delegate routine purchases to intelligent assistants capable of searching, selecting and completing transactions automatically.
If that model becomes commonplace, retailers will need digital storefronts, structured product information and seamless payment options that software agents can access as easily as human shoppers.
In that environment, convenience may become just as important as price.
What Retailers Should Do During the Next 12 Months
Despite covering topics from predictive analytics to agentic AI, the panel’s closing recommendations remained grounded in practical execution.
The advice was remarkably consistent.
Digitise business operations. Organise existing data. Connect customer, inventory, payment and financial systems. Identify one commercially important problem where AI can deliver measurable improvements. Build from those results instead of attempting a company-wide transformation.
That roadmap also reflects a broader lesson for Kenya’s retail sector.
Artificial intelligence is becoming more accessible every month, but competitive advantage will depend less on who adopts AI first than on who prepares their business to use it effectively.
Retailers already possess much of the information they need. The next step is making that information work together before asking AI to make sense of it.
