AI Over-Personalisation: KPMG's Balisteri on the Empathy Deficit

KPMG's Dan Balisteri Warns Enterprises Are Mismanaging AI Personalisation — and Losing Customer Trust

AI is giving organisations unprecedented capability to tailor customer experiences, but KPMG's Dan Balisteri argues that most enterprises are deploying that capability without adequate guardrails — creating an empathy deficit that is actively eroding the trust they set out to build.

The Fine Line Between Helpful and Intrusive

Balisteri, Director of Service Transformation at KPMG, draws a clear distinction between personalisation that serves the customer and personalisation that serves the algorithm. When an interaction feels genuinely relevant — anticipating a need, reducing friction, acknowledging context — it strengthens the relationship. When it tips into over-personalisation, customers feel surveilled rather than understood. The result is a paradox: the more data an enterprise deploys, the less human the experience can feel.

The warning is particularly pointed for contact-centre and CX leaders who have rushed AI deployments in response to competitive pressure. Balisteri's position is that speed of deployment has consistently outpaced the organisational maturity needed to use customer data responsibly and empathetically.

Why AI Deployments Keep Falling Short

According to Balisteri, the root causes of underperforming AI programmes cluster around three recurring failures:

• Misaligned objectives. Teams optimise for operational efficiency metrics — handle time, containment rate — rather than for outcomes that customers actually value, such as resolution quality and emotional ease.
• Absent empathy design. AI models are trained on transactional data but rarely on the emotional and contextual signals that human agents read instinctively. The empathy gap is structural, not incidental.
• Governance lag. Policies governing how customer data is collected, stored and applied to personalisation engines frequently trail the pace of deployment, leaving organisations exposed to both regulatory risk and reputational damage.

Balisteri is explicit that these are not technology failures — they are leadership and design failures. The tools, in his view, are largely capable. The organisational frameworks surrounding them are not.

Rebuilding Trust Through Intentional Design

The corrective, as Balisteri frames it, is not to retreat from AI-driven personalisation but to approach it with far greater intentionality. He advocates for what might be described as empathy-first architecture: designing AI interactions around the customer's emotional state and genuine needs before layering in data-driven personalisation. Transparency also features prominently in his thinking — customers who understand why they are receiving a particular recommendation or experience are measurably more comfortable with it than those who encounter it without context.

The organisations getting this right are not necessarily the ones with the most sophisticated models. They are the ones that have asked the harder question: what does this customer actually need from us right now?

That framing aligns with established behavioural-economics principles around autonomy and perceived control. When customers feel that personalisation is being done for them rather than to them, acceptance and satisfaction scores rise significantly — a distinction that is easy to articulate and genuinely difficult to operationalise at scale.

Why It Matters for CX

Balisteri's analysis surfaces a tension that sits at the heart of modern customer-experience strategy: data abundance does not automatically produce empathy, and efficiency gains do not automatically produce loyalty. For CX leaders in the MENA region — where customer relationships are often high-context and trust is a primary driver of retention — the empathy deficit he describes carries amplified risk. Deploying AI without embedding empathy design and robust data governance is not a shortcut to better CX; it is a reliable path to customer alienation. The competitive advantage will accrue to organisations that treat responsible personalisation as a design discipline rather than a compliance checkbox.