Show summary Hide summary
As AI moves from experimental labs into daily marketing routines, brands and agencies are racing to make sense of what works, what’s risky and what still feels futuristic. A new Modern Retail+ Research study, drawing on a survey of 142 brand and agency professionals and interviews with marketing and tech leaders, maps where AI is already reshaping workflows in 2026 — and where hesitation persists.
How generative and predictive AI are reshaping marketing tasks
Marketers report that AI has migrated from pilot projects into core activities. But adoption varies by type of AI and by task. The survey shows that generative approaches are most visible in content and creative workflows, while predictive systems are more common in measurement and analytics.
- Creative production: Generative AI leads here, with the majority of respondents using it to produce images, copy and video assets.
- Marketing execution: Generative tools are widely used for campaign content and channel-specific creative.
- Communications: Both external and internal comms see regular use of generative AI for drafts and asset creation.
- Measurement and KPIs: Predictive AI is most often deployed to forecast outcomes and analyze performance data.
Blake Lively wins resounding victory: settles lawsuit with Justin Baldoni
MAFS Australia 2026 finale dates revealed: when the season ends and what’s left
Specific survey figures underline the trend: a large share of marketers rely on generative AI for creative tasks, while predictive models are the go-to for KPI analysis. These patterns show how different AI capabilities fit distinct parts of the marketing stack.
Concrete examples of AI boosting speed and scale
Companies of all sizes are sharing early wins. Some implementations focus on producing more content in less time. Others apply analytics to simplify reporting for internal and external stakeholders.
- In-house creative platforms: Large brands have built internal AI studios that generate social and e-commerce assets at scale. One global consumer brand reported moving from a few dozen assets per campaign to hundreds per product using an AI production pipeline.
- Retail media analytics: A combined retail and data division launched customizable AI-driven email digests for suppliers. These weekly reports synthesize short- and long-term KPI trends automatically.
- Agency tech at work: Design and technology firms say they now use AI in production, not just ideation, and are experimenting with digital twins to simulate brand behavior and accelerate strategy development.
Agency and client leaders highlight that AI shortens production cycles and expands the volume of testable creative. At the same time, many stress that the human role remains vital for quality control and nuanced judgment.
Where AI is used less: media buying and finance
Not all marketing functions have embraced AI equally. The study found more modest usage in media planning and financial modeling.
- Media buying and planning: Around a third of respondents use predictive models for this work. Generative tools are used even less often for planning tasks.
- Financial analysis: A similar share use AI for budgeting or forecasting, while fewer turn to generative models for these functions.
Industry conversations suggest AI can still add value in media procurement. For smaller teams, automated testing and platform-driven optimization from big ad platforms can be transformative. But many agencies treat AI as an assist, not a replacement, for complex RFP crafting and strategy.
Agentic AI: promise, experiments and persistent doubts
Agentic AI — systems that autonomously plan and execute multi-step tasks — generates excitement. Yet adoption lags behind predictive and generative tools.
- Current reach: A majority of surveyed marketers report no active use of agentic AI in their workflows.
- Ambitious experiments: Some creative shops have produced short films and advanced concepts using chained AI agents for scripting, visual planning and direction.
54% of respondents said their organizations do not use agentic AI, underscoring the technology’s early-stage status in many marketing teams.
Trust deficits and the risk of hallucinations
One major reason for slow uptake is trust. Agentic systems operate with less human intervention, which raises concerns about errors that can cascade through a process. Marketers worry about AI fabricating details, then executing subsequent steps based on false inputs.
Leaders advise a cautious approach: give agents limited permissions, test them on low-risk tasks and keep humans in the loop for judgment calls. Many see trust as something built over time through predictable, transparent agent behavior.
Governance and data access questions
Managers also flag governance as a top priority. Agentic tools often require integration with multiple data sources and APIs. That creates difficult choices about what data to expose and what actions an agent should be allowed to take on behalf of a team or customer.
- Decisions about data scope determine whether an agent can act independently.
- Teams must weigh privacy, security and brand risk when granting access.
Technical complexity and the integration challenge
Beyond governance, agentic AI is complex to build and maintain. Unlike a single-model application, agents frequently orchestrate other models and services. That requires engineering depth and a thoughtful architecture.
Experts warn that the industry is moving away from “collect everything” approaches. Instead, the goal is to identify the minimal, necessary data that enables a reliable agentic experience.
Practical paths for teams preparing for agent-driven futures
Industry veterans recommend incremental strategies. Start with micromanaged agents that perform discrete tasks. Use those wins to build confidence and clarify the right balance of automation and human oversight.
- Train internal teams now so they can adapt as tools evolve.
- Audit front-end assets and back-end integrations to be “agent-ready.”
- Experiment with hybrid workflows: automated insight generation paired with human validation.
Technologists emphasize that AI has shifted from assistant roles toward operator roles in some contexts. Brands that prepare their tech stacks and governance frameworks will be better positioned to layer in more autonomous capabilities.
