By Alistair Bastian, CTO, Data and Technology Solutions at WPP
The data-driven marketing industry has spent years treating identity as the holy grail. But that era is closing fast, squeezed from one side by tightening privacy regulation and from the other by the growing inadequacy of static IDs in a world that demands richer, real-time insight. Now the conversation is moving away from identity alone, and towards something more powerful: collaborative intelligence.
Putting intelligence at the centre of multi-party collaboration
When we talk about collaborative intelligence, we’re really talking about moving beyond isolated data silos toward a model where organisations can work together on insight without ever moving or exposing raw data. There’s no sharing; it’s about allowing multiple parties to securely ask questions of their combined signals while each retains full control of their data.
Historically, brands and partners would each build their own dataset, analyse it in isolation, then try to join the dots. Collaborative intelligence is about enabling richer insights by letting everyone query against aligned signals within a governed framework.
With privacy expectations rising and regulatory scrutiny tightening, organisations can no longer rely on copying or centralising data to get results. Built on privacy-by-default technology, collaborative intelligence provides a path forward that’s both scalable and directly aligned with how modern data strategies need to work. It provides a pathway for teams to unlock collective insights that lead to better decisions and build trust with their customers and partners at the same time.
The evolution of identity
The industry shift from traditional identifiers to broader signals reflects a growing recognition that identity alone provides only a limited view of human behaviour. Legacy IDs, such as cookies and hashed emails, were built for a different era. New approaches move the focus beyond simply who someone is to why they behave the way they do, moving from identity to intelligence by using contextual, behavioural, location, and engagement signals to create a more complete picture.
Identity doesn’t disappear, but it becomes one input within a richer, AI-powered framework. By analysing diverse, real-time signals together, organisations can uncover patterns and predictions that static ID systems don’t surface. This shift enables more meaningful activation and measurement, rather than relying purely on match rates as a proxy for effectiveness.
For organisations, the transition starts with investing in signal diversity and AI-driven modelling, supported by strong governance. Privacy regulations and platform changes will continue to constrain legacy identity models, while the number of high-quality signals continues to grow. The organisations that adapt best will be those that focus on insight and outcomes, building flexible systems that can scale up quickly.
Everyone needs PETs
Richer signals and smarter modelling are only possible, though, if organisations can access and analyse data responsibly. That’s where Privacy Enhancing Technologies (PETs) are moving from a nice-to-have to a non-negotiable.
PETs matter because they provide a practical way to balance two forces that are pulling in opposite directions: the enormous appetite for data required to build useful AI systems, and the equally strong demand from consumers and regulators for privacy and control over personal information. PETs such as federated learning, private set intersections, differential privacy, and synthetic data allow companies to analyse or model rich data without ever exposing the underlying personal information.
By protecting privacy at the technical level, organisations can work with partners, combine signals from multiple sources, and train better models, all while each party retains control and visibility into how their data is used. For example, brands and retail partners can collaboratively measure sales lift without exchanging sensitive customer records, and even competing organisations can jointly train predictive models without revealing proprietary information. Used in the right way, PETs mean businesses don’t have to choose between privacy and performance.
AI readiness requires strong foundations
Which brings us to the bigger picture. The real prize here is genuine AI readiness, not just better privacy compliance. This requires organisations to think carefully about the foundations they’re building on.
Having the latest AI tools isn’t the differentiator. What matters is whether your data infrastructure and governance are built to actually deliver results from them. For many companies, the real challenge is fragmented systems and inconsistent data quality, which undermine AI performance. Private data networks help address this by allowing organisations to analyse and collaborate on data where it resides, without moving or exposing raw datasets. This creates cleaner, connected signals that are far more reliable for AI-driven insight.
Becoming AI-ready requires aligning data strategy to business outcomes, establishing governance that balances collaboration with control, and access to scalable infrastructure. Organisations that take a more disciplined approach by strengthening architecture, enforcing governance, and enabling secure collaboration are the ones positioned to unlock AI’s value responsibly and at scale.
Collaborative intelligence is where the real competitive advantage lies
The organisations that will thrive in the next phase of data-driven marketing are not necessarily those with the biggest datasets, but those with the most collaborative and privacy-focused approach to using and connecting them. Intelligence built on secure collaboration, richer signals, and sound data infrastructure is where the real competitive advantage lies. The tools to get there already exist. The question is whether the industry is willing to move fast enough to use them.
