Agentforce Marketing needs a CMO-owned optimization contract

Optimization is a policy choice

Salesforce describes AI and agents that can assist with audiences, content, campaigns, timing, journeys, and optimization. That establishes a broad product possibility, not the marketing policy the system should pursue. An agent can improve the metric it is given while damaging margin, brand trust, customer experience, service capacity, or long-term demand. A CMO should therefore treat the optimization target as a governed business decision. The first question is not whether the platform can act across a journey. It is which outcome marketing is prepared to optimize, for which people and channels, under which constraints, and who can stop or reverse the change when the observed result departs from the intended customer and brand outcome.

Write the optimization contract

For each use case, the contract should name the primary outcome, the attribution window, the comparison baseline, and the guardrail measures. It should identify eligible and excluded audiences, approved data and consent states, frequency and suppression rules, brand and accessibility requirements, channel budgets, inventory or service constraints, market restrictions, and content categories that require human approval. It should also state which variables the agent may change and the maximum change permitted in one cycle. That prevents a vague instruction such as improve conversion from silently becoming authority to widen targeting, increase contact, alter an offer, or privilege a short-term response at the expense of customer value.

Preserve a reviewable experiment

Marketing operations should retain the audience definition, data snapshot, prompt or configured objective, generated variants, approvals, distribution record, control group, spend, and outcome data for each material optimization. Measures should separate delivery, response, qualified conversion, margin, unsubscribe or complaint, service load, and longer-term customer behavior. A named channel or journey owner should review exceptions and confirm that a winning result is not explained by audience leakage, offer differences, measurement error, or an unrecorded operational change. If the team cannot reconstruct what changed and who received it, it does not have enough evidence to claim the agent improved the journey or to promote the variant more widely.

Keep the CMO's stop authority explicit

The CMO can delegate routine execution within the contract while retaining authority over the objective, protected audiences, brand boundaries, material spend, sensitive claims, and escalation threshold. Legal, privacy, security, accessibility, sales, service, and finance owners should approve constraints in their domains, but no technical configuration should be treated as a substitute for that business accountability. The contract needs pause triggers for complaints, consent uncertainty, brand or accessibility failures, unexpected spend, inventory or service strain, and measurement instability. At the trigger, the agent should stop expanding the change, preserve evidence, and return the decision to a named person. Scale is earned only when both the target outcome and the guardrails hold.

Turn this source into a reviewable decision

For AI for CMOs, use this briefing as a dated decision record rather than a substitute for the source. Preserve Salesforce Marketing AI, the exact URL, the August 19, 2026 review date, the supported facts above, the editorial interpretation, the limitations, and any buyer-specific evidence. Link that record to the decisions most directly affected: Customer journeys and personalization; Media planning and activation; Measurement and performance explanation; Brand, disclosure, and synthetic-media risk. State whether the source changes the scope, evidence requirement, control, sequence, or only the language used to describe the decision.

Before action, name the accountable owner, affected population and workflow, exact offering or configuration, source data and rights, human decision point, exception and appeal path, complete cost, expected benefit, failure and stop conditions, retained evidence, and next review date. Keep official facts, provider statements, buyer observations, representative tests, measured outcomes, editorial inferences, and unknowns visibly separate. Reopen the record when the source, offer, model, integration, data, policy, population, responsible person, or measured result changes.

Limitations and unknowns

Salesforce is the provider source. Its current Marketing AI page describes AI and agent support for audiences, timing, content, campaigns, journeys, and optimization. It does not independently establish a buyer's entitlement, connected data, consent and suppression state, configured objective, audience rules, brand and accessibility controls, channel capacity, experiment quality, attribution, margin, customer effect, compliance, or business outcome. Current contracts, configuration, audience and consent records, creative and distribution evidence, experiment data, and qualified marketing, analytics, brand, accessibility, privacy, security, procurement, finance, service, sales, and legal review control.