NAC 2026: AI Is Not Growing Brand Loyalty, But Dropping It, New Study Finds

A new multi-country study presented at the 2026 National Advertising Conference has found that artificial intelligence is not building brand loyalty in African retail markets but eroding it. The research, led by the distinguished Professor Uchenna Uzo of Lagos Business School and the Africa Retail Academy, suggests that the tools brands are rushing to adopt are poorly suited to the trust-driven, relationship-based shopping behaviours that actually determine whether African consumers return to a brand.
The findings, presented at a breakout session on the final day of the conference in Abuja, draw on a sequential mixed-method study involving 1,109 survey responses across 16 countries, 10 focus group discussions and three research documents. The study’s central conclusion is that in markets where purchase decisions are shaped by human trust, negotiation and community recommendation, AI interventions designed for structured Western retail environments are actively counterproductive.
Uzo framed the research against Nigeria’s brutal consumer reality where inflation, though cooling from its 2024 peak, remains elevated at 14.63 percent as of September 2026, petrol prices have risen by more than 450 percent since the 2023 subsidy removal, and purchasing power has collapsed for ordinary households. The question the study set out to answer was, how do retail businesses keep customers loyal when wallets are shrinking?
The answer, the data suggests, is not through algorithms. Trust in sellers and product quality emerged as the strongest drivers of purchase decisions, ahead of price, convenience, delivery speed and product recommendations. That finding alone should give pause to brands pouring budgets into AI-driven personalisation and recommendation engines. The machines are optimising for variables that African consumers rank below the human relationships they can see, negotiate with and hold accountable.
The loyalty crisis is already here
The study builds on Uzo’s earlier work, which has tracked a dramatic acceleration in brand-switching behaviour in Nigeria. Research presented at the Top Trends summit in early 2026 found that five in ten Nigerian consumers switched brands in 2025, with projections suggesting eight in ten would be willing to switch by 2026. Experience-related factors, not price, were cited by 70 percent of returning customers as the reason they came back.
More strikingly, loyalty to salespeople stood at 78 percent, compared with 72 percent loyalty to brands themselves. When frontline staff leave, customer allegiance frequently follows. This is the terrain on which AI is being deployed, and it is terrain on which AI has no natural advantage. A recommendation engine cannot replicate the relationship a shopper has with the trader who knows her name, remembers her preferences and extends credit when money is tight.
The AI mismatch where designed for a market that does not exist here
The research examined consumer awareness, trust, ease of use and expectations of AI in African retail contexts. It found that 70.9 percent of respondents had interacted with AI, mainly for product recommendations, inquiries, price comparisons and delivery checks. Trust was moderate to high among users. But a substantial portion reported confusion or felt they did not receive enough support, pointing to a gap between how AI systems are designed and how African consumers actually shop.
The core problem, Uzo argued, is structural. Most AI systems are built for structured retail where environments, clean data, standardised product catalogues, predictable pricing, and transactions that can be optimised without negotiation. African retail, by contrast, is informal, negotiation-driven and relationship-oriented. The systems do not understand haggling. They cannot read the trust signals that determine whether a shopper believes a trader’s claim about a product’s quality. They cannot factor in the community reputation that makes a recommendation credible in one neighbourhood and suspect in another.
This is not a technical limitation that better models will solve. It is a category error. The AI is answering a question the consumer did not ask. The consumer is asking whether the person in front of them can be trusted. The AI is offering a product recommendation.
Study recommendations
The study recommends transparent, user-friendly, human-centred AI solutions that support negotiation, promoye trust and adapt to African shopping behaviours. It means using AI to support the human touchpoints that drive loyalty, not to replace them. It means recognising that the trader who extends credit, the salesperson who remembers a regular customer’s brand preference, and the community leader whose endorsement carries weight are not inefficiencies to be automated away.
The research also raises a question whether the industry’s enthusiasm for AI is being driven by what clients want or by what agencies find exciting. Brands under pressure to cut costs are attracted to automated customer service and algorithmic recommendations. Agencies under pressure to show innovation are eager to supply them. Neither impulse necessarily serves the consumer whose loyalty is being sought. The data suggests the consumer is looking for something the machines cannot provide.
