Personalisation that uses AI at scale only works when it’s built on trusted data, sound governance, and experiences that feel human.
Allan Christian, Senior Vice President and General Manager - Engage at Precisely, sets out how firms can use first-party data and agentic AI for customer communications without eroding trust or compliance.
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.
Regulated firms are under pressure to tailor customer messages to context and behaviour, while privacy rules and consumer expectations limit how far third-party data can stretch. The sections below explain where personalisation programmes fail, what CCM and IXM change in practice, and how governance and data integrity need to work before automation scales.
Useful personalisation depends on data teams, legal, and CX agreeing what “helpful” means before models or orchestration scale. The sections below set out how first-party data, IXM, governance, and integrity fit together in regulated customer communications.
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.
"Personalisation works when it’s driven by trusted data and clear purpose, not gimmicks."
Allan Christian, Senior Vice President and General Manager - Engage, Precisely
Consented behaviour and location data can then drive the next message from context, consent, and recent activity.
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.
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, dialogue-driven experiences.
These approaches lay the groundwork for advanced capabilities, including agentic AI that can coordinate steps, retrieve relevant context, and propose content variations, all under human oversight for sensitive decisions.
"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
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.
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.
Regulatory Context For Personalised Comms
Firms using profiling or automated decisioning in customer journeys should map how ICO AI guidance and Consumer Duty expectations apply before scaling new channels or models.
Legal, data, and CX teams need shared criteria for when a message can be sent, what must be explained to the customer, and what evidence to retain if a complaint or audit follows.
"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
Those guardrails only hold if the data feeding personalisation is accurate and consistent across channels.
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.
Sending Important Documents At Scale?
Learn how Mailock Automated helps organisations protect high-volume customer communications without forcing every recipient through a portal.
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 - as well as 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.
"When a personalised message carries sensitive information, the delivery route still needs recipient authentication, controlled access, and a record of what was sent and opened. Personalisation and protection have to be designed together."
Paul Holland, Founder and CEO, Beyond Encryption (Mailock)
Firms that send tailored statements, policy updates, or account messages by email often pair orchestration with secure email so recipients can open and reply without a separate portal account for every interaction.
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.
Scale personalisation only after consent, data quality, and governance rules are clear enough for every channel to use.
That sequencing matters because a single bad send in a regulated sector can undermine trust across the rest of the programme.
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
Allan Christian, Senior Vice President and General Manager - Engage, Precisely
Sam Kendall works on digital marketing at Beyond Encryption, helping build B2B marketing activity around research, first principles, and sustainable growth. He writes about marketing effectiveness, positioning, customer communications, and digital culture, with longer-form work published at ATNL.
Sam Kendall
.svg)
Previous post