The Personalisation Paradox: More AI, Less Difference

For more than a decade, personalisation has been framed as the solution to mass digital experiences. Smarter recommendations, sharper targeting, and more relevant messaging, which meant to ensure that technology finally treated people as individuals rather than averages.

In 2026, as AI systems spread across advertising, retail, enterprise software, and consumer apps, a contradiction has become increasingly visible. Personalisation is everywhere, yet it often feels indistinguishable from one brand to the next.

Open a shopping app, scroll through a promotional email, or interact with a digital service, and the experience tends to follow a familiar pattern. The tone is similar, the prompts feel predictable, and all the content seems to read the same.

The logic behind the recommendations appears interchangeable. What was designed to create differentiation is, in practice, producing uniformity.

This outcome is not the result of artificial intelligence failing to deliver. It reflects how organisations across industries have adopted the same tools, worked within similar constraints, and optimised toward identical definitions of success.

Efficiency Has Become the Dominant Design Principle

From a technical perspective, personalisation has never been more advanced. Modern AI systems can process behavioural data at a scale that was not feasible even a few years ago. McKinsey estimates that companies deploying advanced personalisation strategies see revenue lifts of 10 to 15 percent, while global spending on AI-driven marketing and personalisation is expected to exceed $35 billion by 2027.

Despite that investment, the outputs increasingly converge.

Sue Azari, Industry Lead for eCommerce at AppsFlyer, said AI has fundamentally altered how personalisation is delivered. “Before this, doing personalisation properly required a lot of human effort,” she said. “Teams had to manually build segments, curate recommendations, and adjust campaigns. That approach was expensive and difficult to scale.”

AI replaced that manual effort with speed and automation. Large retailers such as Walmart have publicly described how algorithmic recommendations took over work that would otherwise have required large teams to manage.

The consequence of this efficiency-driven shift is subtle rather than dramatic. When speed, cost reduction, and optimisation become the primary objectives, brands tend to make similar decisions using similar infrastructure. “When personalisation is implemented in a shallow way, everything starts to feel monocultural,” Azari said.

When Optimisation Replaces Differentiation

The flattening effect is especially visible in performance marketing and advertising. An advertising technology specialist working with brands across both emerging and mature markets described the current environment as one of extreme optimisation.

“AI has shifted competitive advantage away from intuition and toward data completeness,” the executive said. “As systems approach peak efficiency, outperforming the market becomes incremental rather than transformational.”

Creative ideas are tested rapidly, measured against short-term performance metrics, and discarded if they do not deliver immediate results. Generative AI, which was expected to expand creative possibilities, often reinforces the same patterns by amplifying what has already worked elsewhere.

Over time, this feedback loop narrows variation. Personalisation exists at scale, yet the space for meaningful difference shrinks.

Organisational Caution Shapes What Users Experience

Jessica Constantinidis Innovation Officer at ServiceNow argues that the sameness many users feel has little to do with technological limits. It is more closely tied to how organisations manage risk and change.

“AI is already capable of far more than most organisations are comfortable deploying,” she said. “What slows progress is uncertainty around governance, data usage, and accountability.”

According to Constantinidis, many companies are still focused on understanding boundaries rather than redesigning experiences. As a result, AI is often layered onto existing workflows instead of being used to rethink how customers are engaged. “When organisations move carefully, they default to familiar patterns,” she said. “That consistency reduces risk, but it also reduces differentiation.”

This cautious approach ensures compliance and stability, while simultaneously narrowing the range of experiences users encounter. Even brands built around identity and intimacy are not immune to this pressure.

Deepika Nagasamy, Founder of Dipsy Store, said her decision to start a skincare brand came from frustration with how similar products sounded despite making very different claims. “There were too many products and too much jargon,” she said. “It was difficult to understand what was real and what was positioning.”

Her response was to build a brand rooted in Indian-origin ingredients, clear documentation, and cultural specificity. She also acknowledged that AI is reshaping how consumers discover products. “People now ask AI systems for skincare routines instead of searching or asking friends,” Nagasamy said. “That means brands are increasingly being interpreted by machines as much as by humans.”

A similar tension exists in fine jewellery. Amreen Iqbal, Founder of Piece of You Jewellery, said technology has enabled personalisation at scale while introducing new challenges. “Consumers today want meaning, customisation, and ethical sourcing,” she said. “AI helps us understand preferences and personalise communication, but preserving emotional depth as we scale requires constant attention.”

In both cases, technology makes personalisation easier to deliver, while making distinctiveness harder to sustain.

Why Personalisation Often Feels Generic

One reason AI-driven personalisation disappoints users is that it prioritises recommendations over responses.

“Personalisation is not only about what you show a customer,” Constantinidis said. “It is also about how you respond to their situation.” When automated systems replace human judgment without encoding nuance, experiences become predictable rather than personal.

Agentic AI, which can take actions instead of simply generating outputs, has the potential to address this gap. Its effectiveness depends on how carefully scenarios, severity, and context are modelled. Without that groundwork, automation tends to enforce policy rather than reflect judgment.

Despite rapid progress, AI remains constrained by infrastructure realities. Compute availability, memory capacity, energy consumption, and latency continue to shape what systems can realistically deliver. Data centre energy usage is rising globally, and memory shortages are expected to persist through much of the decade.

These constraints do not halt personalisation efforts, although they do limit how deeply and dynamically systems can respond in real time.

The deeper issue is not whether AI can personalise more effectively. It is whether companies are clear about what they want personalisation to represent.

“AI is not just a tool,” Constantinidis said. “It reflects how a company chooses to behave.” At present, most organisations use AI to optimise existing processes rather than to rethink the experience they want to create. Until that changes, personalisation will continue to scale across industries, while genuine difference remains rare.

The technology is already capable. The harder transformation lies within organisations themselves.