In a Digiday sponsored column, Oz Etzioni, CEO and co-founder of Clinch, argues that adding more AI agents alone will not solve persistent advertising problems. Per the column, AI can process engagement signals at scale and surface previously invisible patterns, but those gains are limited when data and workflows remain fragmented across platforms. Etzioni writes that agent-to-agent communication is not equivalent to system-level coordination; agents need access to shared data and consistent logic to compound learning across the campaign lifecycle. The piece warns that without a unified infrastructure, creative, activation and performance data stay siloed and insights fail to carry across workflows.
What happened
In a Digiday sponsored column, Oz Etzioni, CEO and co-founder of Clinch, argues that simply adding more AI agents will not fix advertising's long-standing fragmentation. Per the column, AI can process engagement signals at scale and surface patterns that were previously invisible, but its impact is constrained when data and workflows are split across platforms.
Technical details
Etzioni writes that agent-to-agent communication differs from true system-level coordination. The column says agents require access to shared data and consistent decision logic to align across creative, activation and measurement stages; when those inputs are spread across disconnected systems, agents end up optimizing in isolation and insights remain trapped in silos.
Industry context
Editorial analysis: Companies and vendors promoting autonomous agents frequently emphasize agent interoperability, but industry reporting and practitioner experience show that interoperability alone does not create a single source of truth. Comparable transitions in martech and adtech have historically depended on data integration, canonical identity graphs, and consistent measurement schemas before automation could scale learning effectively.
Context and significance
Editorial analysis: For practitioners, the column underscores a common pattern - automation amplifies the capabilities of underlying infrastructure. In advertising, the payoff from models and agents is limited if pipelines do not deliver timely, reconciled signals about creative variants, audience exposure and conversion events. This matters for teams investing in MLOps, feature stores, and real-time telemetry for campaign analytics.
What to watch
Analysts can monitor three observable signals: vendor claims about shared-data layers or unified activation platforms, improvements in end-to-end measurement latency, and evidence of learning compounding across campaign stages (for example, creative-level signals informing media optimization within the same measurement window). Observers should evaluate vendor architectures against those operational indicators rather than agent counts alone.
Scoring Rationale #
The column is a practical reminder for adtech and ML practitioners that automation depends on integrated data and measurement. It is useful guidance but is commentary rather than a technical breakthrough, so its direct impact is moderate.
Practice with real Ad Tech data
90 SQL & Python problems · 15 industry datasets
[Active Search Campaigns by BudgetEasy](/problems/sql/active-search-campaigns-by-budget)
[High CPC Clicks & Poor Landing PagesMedium](/problems/sql/high-cpc-clicks-poor-landing-page)
[Campaign ROAS by Attribution ModelHard](/problems/sql/campaign-roas-by-attribution-model)
250 free problems · No credit card