Harnessing Data for Hyper-Personalised Customer Communications in the AI Era

Personalisation that uses AI at scale only works when it’s built on trusted data, sound governance, and experiences that feel human.

Allan Christian is Senior Vice President and General Manager - Engage at Precisely. He brings two decades of experience in communications and data, spanning HSBC, Pitney Bowes Software & Data, and MapInfo.

At Precisely, Allan has helped shape the delivery of hyper-personalised, compliant communications, powered by data integrity.

He's now turning his focus to intelligent interactivity - applying agentic AI to unify orchestration, decisioning, and design for real-time experiences.

In this episode, Allan explains how to harness first-party data for personalised communications without losing sight of trust or compliance.

Why First-Party Data Is Your Personalisation Flywheel

The move to privacy-friendly customer engagement means first-party data is now the engine of personalisation.

When that data is accurate, enriched, and well-governed, teams can design once and deliver to many, tailoring messages to moments rather than broad demographic segments.

The result should be less channel noise and more timely interventions that feel genuinely helpful.

From Demographics To Decisions: Using Signals That Matter

Modern personalisation programmes go beyond name inserts and age-brackets to combine consented behaviour, context, and risk signals.

Location intelligence helps shape offers and advice, while recent activity drives next-best actions rather than batch-and-blast campaigns.

When this is done well, it can reduce fatigue and increase trust because each interaction has a clear purpose.

Any outreach that includes promotion should adhere to the ICO’s direct marketing guidance.

Intelligent Interactivity: The Next Leap For CCM/IXM

Static templates and one-way notices are being replaced by dynamic, two-way interactions that evolve in real time.

CCM (Customer Communications Management) and IXM (Interaction Experience Management) are broad frameworks designed to move communications beyond static outputs and towards responsive, dialog-driven experiences.

These approaches lay the groundwork for a spectrum of advanced capabilities - including enabling technologies like agentic AI that can coordinate steps, retrieve relevant context, and propose content variations, all under human oversight for sensitive decisions.

If you are deploying AI features, follow the NCSC's AI guidance.

Design Once, Deliver Everywhere

A composable stack - a modular set of tools that work together - lets teams model logic, content, and data centrally, then render consistently across email, SMS, app, web, and print.

This removes duplicated effort and makes compliance controls easier to audit by a wider audience, including less technical team members.

It also improves speed to value because one change can propagate everywhere, reducing manual overhead.

Governance By Design, Not By Afterthought

Personalisation is only sustainable when it is lawful, fair, and understandable.

In regulated sectors, that means meeting duties around clear communications, explainability, and evidence of good outcomes.

It also means conducting Data Protection Impact Assessments (DPIAs) - structured processes that identify and evaluate risks to personal data - so organisations can map potential impacts, assign responsibilities, and put appropriate safeguards in place before launching new initiatives.

Aligning With UK And EU Expectations

UK guidance on AI and data protection sets expectations for fairness, transparency, and purpose limitation in automated decision-making, including profiling - as it should.

The FCA’s Consumer Duty requires communications that meet customer needs, are likely to be understood, and support effective decisions.

For EU audiences, the AI Act introduces risk-based obligations that affect how AI-enabled communications are designed, tested, and monitored.

The importance of aligning with these obligations cannot be understated.

Data Integrity: The Bedrock Of Helpful Automation

Without high-quality data, even the best orchestration will amplify errors rather than outcomes.

Identity resolution (matching records that belong to the same person), deduplication, and provenance tracking stop journeys from fragmenting and keep messages consistent across channels.

This is where Precisely’s heritage in data integrity and location intelligence gives their team a practical advantage in moving from data to action.

Measuring What Matters, Not Just What Moves

Outcome metrics should mirror your duty to customers - clarity, timeliness, and the avoidance of foreseeable harm - not only click-through rates.

Experimentation remains critical, but tests should be governed and reproducible, with audit trails for sensitive cohorts.

Human review is essential wherever model or rule decisions materially affect a person.

What Great Looks Like: A Practical Blueprint

The most effective programmes share a few traits - clarity of purpose, composable design, and shared guardrails between data, legal, and CX.

The playbook below is a starting point you can adapt to your context.

FIRST: A Simple Framework For Safe Personalisation

• F – First-party data: Prioritise consented, high-quality sources with clear retention and access rules.
• I – Integrity: Govern lineage, accuracy, and deduplication to avoid conflicting messages.
• R – Responsible AI: Run DPIAs (Data Protection Impact Assessments), document explainability, and define human-in-the-loop triggers.
• S – Secure orchestration: Apply secure-by-design principles to models, prompts, and connectors.
• T – Trust outcomes: Measure clarity, timeliness, and resolution rates alongside conversion.

Allan’s Takeaways

"Personalisation works when it’s driven by trusted data and clear purpose, not gimmicks."

Allan Christian, Senior Vice President and General Manager - Engage, Precisely

"Agentic AI is exciting, but the real value comes from unifying content, data, and decisions so every channel tells the same story."

Allan Christian, Senior Vice President and General Manager - Engage, Precisely

"Seeing governance as red tape is a mistake - it’s how you scale safely and keep customer trust."

Allan Christian, Senior Vice President and General Manager - Engage, Precisely

FAQs

What’s the Difference Between CCM and IXM?

CCM focuses on producing and delivering compliant communications, while IXM treats communications as interactive experiences that adapt in real time.

How Does Agentic AI Help Without Risking Compliance?

By constraining actions to approved policies, logging every step, and triggering human review for sensitive decisions.

Where Should We Start if Our Data Is Messy?

Start with data integrity work - identity resolution, deduplication, and lineage - before scaling orchestration or generative content.

What Outcomes Should We Measure?

Beyond clicks, measure clarity, timeliness, resolution rates, and evidence that customers made effective, well-informed decisions.

References

Automated Decision-Making and Profiling, ICO, 2023

Guidance on AI and Data Protection, ICO, 2023

PRIN 2A.5 Consumer Duty - Consumer Understanding Outcome, FCA, 2023

Guidelines for Secure AI System Development, NCSC, 2023

Artificial Intelligence Act (Regulation (EU) 2024/1689), EUR-Lex, 2024

Unlocking the Next Frontier of Personalized Marketing, McKinsey, 2025

The State of AI, McKinsey, 2025

Marketing Trends 2025, Deloitte Digital, 2025

Personalization Done Right, Harvard Business Review, 2024

Joint Guidance on Deploying AI Systems Securely, CISA, 2024

Direct Marketing Guidance, ICO, 2022

Precisely Named a Leader in the Aspire Leaderboard for IXM, Precisely, 2024

Reviewed by

Sam Kendall, 01.09.2025