The news: After brands focused heavily on AI acceleration across sales, marketing, and customer success in 2025, data is emerging as a more consequential priority this year. In a conversation with EMARKETER, Craig Dempster, CEO of growth services provider Trilliad, outlined why data integration, not AI tools alone, will shape customer experience and growth in 2026.

Data proves powerful: Quality data will be the leading factor giving revenue-focused companies an edge, according to Dempster. Despite AI’s rapid evolution in 2025, Dempster argued that differentiation will not come from AI tools themselves, but from the data that fuels them. As he put it, competitive advantages will “live at the data layer more so than the AI layer” in 2026.

What data works: Crucially, not all data will deliver the same value—EMARKETER and TransUnion findings show that siloed, incomplete data is the primary reason marketers question the accuracy of their measurement.

• Technological advancements have created “a vast amount of data” that companies previously “never had access to,” making it critical for brands to identify which data sources directly inform revenue decisions and prioritize quality and accessibility over collecting more signals.
• Brands must “unify data across the revenue generation spectrum,” from marketing through sales, to ensure AI systems operate on a complete view of the customer rather than fragmented snapshots.
• Breaking down internal siloes is an imperative, ascross-functional data sharing will help enable the “frictionless customer journey" Dempster identifies as a needed element to drive conversions, retention, and growth.

Implications for brands: Predictable revenue growth in 2026 will be driven less by AI tools and more by how companies integrate data across organizations and use that unified data to power AI more effectively.

For brands, this shifts the focus from adopting new AI capabilities to strengthening the data foundations AI tools are built on. Companies that continue to deploy AI on fragmented, siloed data are likely to face limited returns.