AI Won’t Fix Bad Data — Why Identity and Enrichment Now Drive Revenue
In digital media today, everyone agrees on one thing: data is the most valuable asset in the business.
And yet it is also the most fragmented, misunderstood, and underutilized asset across the ecosystem.
In my conversation with Brian Silver, EVP of Global Marketing Solutions at TransUnion, we unpack what is actually happening at the intersection of identity, measurement, and AI — and why most companies are not as prepared as they think they are.
The Fragmentation Problem Isn’t Going Away
Marketers today operate across 15–16 different systems just to follow and measure a customer journey. Publishers collect vast amounts of data but struggle to activate it at scale. Identity solutions proliferate, but no single universal ID exists.
The result?
- Data silos
- Measurement inconsistencies
- Underutilized first-party assets
- AI systems built on incomplete foundations
Fragmentation is not a new problem — but AI is magnifying its consequences.
60% of AI Initiatives Will Fail
One of the most striking data points Brian shared: roughly 60% of AI initiatives fail due to poor data quality or weak signal strength.
Not because of model sophistication.
Not because of compute.
Because of inputs.
“Garbage in, garbage out” is not a cliché in the AI era — it is a structural constraint.
Many companies are experimenting with AI at the surface level (search, UI assistants, workflow efficiency). Fewer are investing in the foundational layers required to drive real commercial impact.
The Three Foundations of AI-Driven Revenue
Across companies successfully applying AI, three themes consistently emerge:
1. Identity Resolution
If you cannot resolve a customer across environments with high fidelity, your AI models will misfire.
Identity resolution is no longer a targeting enhancement — it is infrastructure. Without it, you cannot measure accurately, enrich intelligently, or optimize confidently.
2. Data Enrichment
Most companies lack depth, not volume.
Enrichment — adding behavioral, demographic, transactional, or contextual layers — increases predictive power. AI systems require high-quality signal density to generate meaningful propensity models.
Scale without enrichment does not create advantage.
3. Connectivity and Measurement
Resolution and enrichment mean little if they are not connected across platforms and measurable in closed loops.
The companies winning are building “flywheels”:
Resolve → Enrich → Activate → Measure → Optimize → Repeat.
That continuous feedback loop is where AI moves from experimentation to revenue impact.
The Shifting Power Dynamic
A particularly interesting theme in our discussion was the evolving balance between demand and supply.
Publishers have long argued that their first-party data is undervalued. As third-party signals deteriorate and identity becomes harder, the supply side’s direct consumer relationships are becoming structurally more important.
The economics of data may shift — but only for publishers who invest in resolution, partnerships, and scalable infrastructure.
The Emerging Risk: Synthetic Traffic and AI Agents
Looking forward, one of the most uncertain dynamics is the rise of AI agents and synthetic traffic.
If automated agents begin acting on behalf of users at scale, how do we:
- Verify identity?
- Attribute outcomes?
- Measure ROI?
- Distinguish human demand from machine demand?
This may become one of the defining challenges of the next phase of digital advertising.
The Bottom Line
AI is not a shortcut.
It is a multiplier.
If your identity infrastructure is weak, AI accelerates errors.
If your data is fragmented, AI amplifies confusion.
If your measurement loop is broken, AI scales inefficiency.
But when identity resolution, enrichment, and connectivity are strong, AI becomes a revenue engine.
For publishers and ad tech companies alike, the question is not whether to adopt AI.
The question is whether your data foundation is ready for it.