AI Agents Reshape Marketing Operations as Adobe and Salesforce Race to Embed Autonomous Workflows

How AI Agents Are Reshaping Marketing Operations

AI agents in marketing operations are no longer a roadmap item. Adobe and Salesforce have spent the past six months embedding autonomous workflows directly into their marketing clouds. The shift is reshaping how mid-market and enterprise teams structure their operations. Both vendors are betting customers will pay premium subscriptions for software that does the work, not just stores the data.

Adobe vs Salesforce: Two Different Bets on Autonomous Workflows

Adobe disclosed at its annual summit that its Experience Platform now hosts more than 200,000 active AI agent runs per week. That figure was zero a year earlier. It reflects how quickly Adobe’s all-in bet on agent-based marketing has translated into platform activity.

Salesforce launched Agentforce in late 2024. Agent-driven workflows now touch roughly one in five marketing automation customers on the platform. That growth is part of a broader strategic shift as Salesforce quietly sunsets Marketing Cloud in favour of Agentforce.

How the Pricing Models Compare

Competition has spilled into pricing. Adobe sells agents as discrete consumption units and charges per task. Salesforce bundles them into existing Cloud subscriptions and charges for incremental scale. Independent benchmarks remain scarce. Early adopters report cost-per-completed-task between two and five cents for routine workflows — audience segmentation, campaign performance summarization, and abandoned-cart recovery. Agentic AI entering bank back offices is following comparable consumption pricing frameworks.

One VP of marketing operations at a Fortune 500 retailer requested anonymity to discuss internal vendor evaluations. They said teams are redirecting twenty to thirty percent of operational time toward strategy. Agents now handle work that used to consume a coordinator’s day. That pattern is consistent with broader findings — what 35 CROs are getting from AI in revenue operations shows similar time reallocation across sales functions. The same executive cautioned that governance frameworks have lagged. Most teams are running agents without formal review of decisions made on their behalf.

Governance Has Lagged AI Agent Adoption — Here’s What Buyers Must Do

For martech buyers assessing AI agents in marketing operations, the practical question is governance, not capability. Both vendors publish audit logs. Neither has shipped enterprise-grade approval workflows that gate agent decisions before execution. As cybersecurity teams warn, AI agents are fundamentally an identity and access governance problem most enterprises are not yet addressing. Buyers should also consider how CDP consolidation affects the data layer agents depend on before committing to either architecture. Pressure-test rollback procedures. Demand contractual clarity on liability when an autonomous agent makes a customer-impacting error.