Everyone wants to talk about AI agents, autonomous marketing, and orchestration.
But new research from Scott Brinker suggests the biggest barriers to AI adoption are not actually about AI at all. In his latest report, The New Martech Stack for the AI Age, marketers identified poor data quality, integration friction, and fragmented customer identity as the primary obstacles standing in the way of AI implementation today.
Those challenges ranked ahead of budget, talent, or access to AI tools, which says something important about where the market really is. The biggest obstacle to AI is not access to models. It is the inability to create trusted customer context that AI systems can act on in real time.
As AI becomes more deeply embedded into marketing operations and real-time customer experiences, the gap between brands with trusted customer understanding and brands without it is about to widen dramatically.
Much of the conversation around AI in marketing assumes the hard part is the model itself. In reality, the harder challenge is ensuring AI systems understand customers accurately enough to make relevant decisions in the moments that matter.
AI does not fix fragmented customer understanding. It accelerates whatever understanding already exists inside the business, whether that understanding is accurate, incomplete, or inconsistent.
AI Is Exposing the Cracks That Already Existed
For years, marketers have been able to work around disconnected systems and fragmented customer records.
Teams built manual processes. Analysts stitched together reports. Campaigns were optimized channel by channel. Identity gaps were tolerated because humans compensated for them.
AI changes that equation.
AI agents operate on speed, scale, and automation. They make decisions continuously, often without human intervention. Which means they rely entirely on the quality of the context they receive.
As Brinker writes, “agents are ravenous for context.”
If the customer context is incomplete, duplicated, stale, or inconsistent, the output will be too.
That's why the report matters. The market is realizing that AI readiness is fundamentally tied to data readiness.
And more specifically, customer intelligence readiness.
The Three Problems Blocking AI Right Now
The challenges highlighted in the report are not isolated technical issues. Together, they point to a larger operational problem: most enterprises still struggle to create a consistent, real-time understanding of the customer across the business.
That becomes significantly more important in an AI-driven environment where systems are no longer just analyzing customer signals. They are increasingly acting on them.
1. Brands still struggle to create consistent customer context
Most enterprises still operate across dozens of disconnected systems.
Customer data exists across ecommerce platforms, loyalty systems, CRM platforms, service interactions, email engagement, mobile apps, advertising platforms, and in-store systems. Each source describes the customer differently.
One system stores “first_name.”
Another stores “given_name.”
A third stores “fname.”
That inconsistency sounds small until an AI system tries to reason across all of it in real time.
AI models don’t normalize data automatically. They amplify whatever patterns already exist.
This is why foundational capabilities like schema-free ingestion and semantic normalization matter more than ever. The goal is not simply consolidating data in a lakehouse. It’s creating shared meaning across systems so AI can operate on trusted context instead of fragmented records.
2. Integration friction is becoming an AI bottleneck
The report repeatedly returns to one central issue: integration complexity.
For years, marketing technology evolved as a collection of disconnected applications stitched together through APIs and pipelines. That architecture was already difficult for humans to manage.
It becomes exponentially harder in an agentic world.
AI agents don’t operate well across disconnected systems with inconsistent schemas, duplicated data, and delayed synchronization. They require immediate access to shared context.
That’s why the report describes the future architecture of marketing not as a rigid stack, but as a “composable canvas” built on a unified data foundation.
This shift matters.
The future of AI marketing isn’t about adding more tools. It’s about reducing the friction between systems so customer context can move fluidly across the organization in real time.
3. Most brands still don’t truly know who their customer is
This one is the real problem.
Fragmented identity quietly undermines almost every AI initiative in marketing.
If the same customer exists as five separate records across the business, every recommendation engine, suppression model, audience segment, and next-best-action system is operating from an incomplete picture.
The report found that customer profiles from CDPs and CRMs are now the number one internal data source connected to AI agents, cited by 61.2% of organizations.
That means the quality of customer identity directly shapes the quality of AI decisions.
This is why customer intelligence is becoming strategically important again. Identity resolution remains a critical component, but only as part of a broader understanding that combines behavioral, transactional, and contextual signals into a complete view of the customer.
The brands moving fastest in AI are often the brands that have already invested in creating trusted customer understanding first. They are able to recognize customers consistently across channels, understand changing behavior in real time, and respond with more relevant experiences because the systems making decisions are working from the same customer context.
The Real Shift Happening in Martech
One of the most important ideas in the Chiefmartec report is that the martech stack itself is changing shape.
The industry is moving away from isolated systems and toward what Brinker describes as a unified “context-as-a-service” architecture.
That shift reflects a broader change happening across the market. AI is pushing businesses away from static campaigns and disconnected workflows toward operating models built around continuous decisioning, adaptive experiences, and real-time responsiveness.
In that environment, the role of the customer data layer evolves significantly. It is no longer just a system for segmentation or activation. It becomes the governed customer context layer that AI agents rely on to understand who the customer is, what they have done, what matters right now, and what should happen next.
This is ultimately where the market is heading: toward AI systems that can continuously translate customer signals into intelligent action across the business.
As businesses move toward AI-native operating models, a new set of standards is emerging to help agents access and share context across systems.
That is also where emerging standards like MCP (Model Context Protocol) become increasingly important. The report highlights MCP as one of the foundational standards enabling AI systems to access trusted context across tools and environments.
But standards like MCP do not solve fragmented customer understanding on their own. They simply make customer context more accessible to the AI systems that depend on it. AI orchestration cannot compensate for inconsistent identity, and agents cannot generate relevant decisions from incomplete or fragmented customer context. Those systems only work as well as the understanding of the customer underneath them.
The Emerging Divide Between Brands
Right now, most brands fall into one of two categories.
Phase 1: Experimenting with AI
isolated AI use cases
inconsistent customer understanding
difficulty scaling outcomes
Their AI initiatives may produce isolated wins, but they struggle to scale. The customer intelligence layer underneath is still fragmented, which pushes real AI ambitions further out until the core problems get solved.
Phase 2: Operationalizing AI
trusted customer intelligence
real-time decisioning
AI embedded into everyday customer experiences
They’re positioned to operationalize AI more effectively because the underlying context is already trusted.
That distinction will increasingly determine which brands can move from AI experimentation to AI-powered operations.
The Most Important AI Investment May Not Be Another AI Tool
The market still tends to frame AI as an application-layer challenge, with most of the attention focused on agents, orchestration platforms, and automation frameworks. But the brands pulling ahead are increasingly approaching AI differently. They recognize that AI is only as effective as the customer understanding behind it.
Before investing in another AI tool, organizations should ask a simpler question: can our systems create a trusted, real-time understanding of the customer that AI can act on intelligently?
For many brands, that remains the bigger challenge.
As AI becomes embedded into every customer interaction, customer intelligence becomes the operational layer that determines whether businesses can recognize customers, interpret signals in context, and make relevant decisions in the moments that matter.
The emerging divide in the market will not be between brands that use AI and brands that do not. It will be between brands that can provide AI with trusted customer context and those that cannot.
The winners in the AI era will not simply have better AI. They will have a better understanding of their customers.
If you're evaluating your own AI readiness, join us at Databricks Data + AI Summit 2026. We'll be sharing how leading organizations are solving the customer context challenge by unifying fragmented customer data, improving identity resolution, and creating the trusted foundation that AI systems depend on.
Visit us at Booth #501 to see how brands are turning fragmented customer data into trusted customer context that powers AI, continuous decisioning, and real-time customer experiences. Explore the sessions and event details: https://amperity.com/events/databricks-data-ai-summit-2026
This is the challenge Amperity has been helping some of the world's most recognizable brands solve for years. By transforming fragmented customer data into trusted customer intelligence, organizations can create the context AI needs to deliver more relevant experiences, make better decisions, and adapt in real time.
As the industry moves toward context-as-a-service architectures, embedded AI, and emerging standards like MCP, that customer intelligence layer becomes increasingly important. The brands best positioned for the AI era will be the ones that invested in understanding their customers first.