17 Oct 2026 · 5 min read
What does bilateral deal governance mean for brand safety teams?
What does bilateral deal governance mean for brand safety teams?
Bilateral deal governance means that brand safety constraints are committed to in the deal terms before any inventory runs, held in a record that both buyer and publisher own, and auditable by the brand safety team at any point during or after the campaign.
Pre-committed brand safety: ** Brand safety constraints specified in the deal terms before the campaign goes live, rather than enforced at impression level after inventory has been evaluated; the publisher commits to these constraints as a condition of the deal.
Bilateral deal record: ** A committed record of deal terms held by a neutral marketplace and owned equally by buyer and publisher; neither party can alter it after commitment, and either party can retrieve and verify it independently.
Post-campaign reconciliation: ** The process of verifying after a campaign runs that delivery matched what was agreed; bilateral deal records make this comparison straightforward because the committed terms are an immutable reference point.
Brand safety escalation: ** The mechanism within an agent's buying mandate that pauses deal commitment when proposed inventory falls outside the brand safety constraints, requiring human review before the deal proceeds.
How brand safety works in traditional programmatic
Traditional programmatic brand safety operates primarily at impression level. A DSP or verification partner evaluates each impression as it becomes available : checking the page URL, content category, and audience signal against the campaign's block list and safety settings : and either bids or passes. The decision is made in milliseconds, before the ad serves.
This approach has two limitations. First, it is reactive: the impression has already been offered before the safety check runs. Second, it creates an accountability gap: if unsafe inventory runs, the record of what was agreed is typically held by the DSP, not the buyer, and reconstructing what happened requires pulling data from multiple sources that may not tell a consistent story.
How bilateral deal governance changes brand safety
Bilateral deal governance adds an upstream layer. When a buying agent negotiates a deal through an AAMP-compliant marketplace, the brand safety constraints are specified in the deal terms as a condition of the agreement. The publisher's agent accepts or rejects the deal with those constraints in place. Once both parties commit, the constraints are part of the bilateral deal record.
This changes the accountability model in two ways. Before the campaign: the publisher has formally committed to delivering within the brand safety constraints, not just to allow impression-level filtering. After the campaign: if unsafe inventory runs, the buyer can retrieve the bilateral deal record and demonstrate that the constraint was committed to, not merely attempted at impression level.
The brand safety team now has a documented commitment rather than a filtering configuration. This matters for internal sign-off, for regulatory requirements, and for conversations with clients about what protections were in place.
What the audit trail looks like
In a bilateral deal governance model, the audit trail for a campaign has three layers. The deal record: the committed terms including brand safety constraints, timestamped at commitment and immutable. The delivery report: what inventory ran under the deal and when. The verification data: impression-level checks from a third-party verification provider confirming that delivery matched the committed context.
When all three are consistent, the brand safety case is clear. When there is a discrepancy between the deal record and delivery, the brand safety team has a precise reference point for the dispute: the committed record of what was agreed, held by a neutral marketplace, queryable by either party.
How brand safety teams should engage with this model
Brand safety teams should be involved in mandate definition before the agent goes live. The brand safety constraints in the buying mandate are the parameters that the agent will honour autonomously; they need to reflect the brand safety team's actual requirements, not a generic template.
The escalation triggers in the mandate should include brand safety edge cases: publisher categories where the team wants human review even if other mandate conditions are met, pricing thresholds that suggest non-premium inventory, and deal structures that imply reduced transparency. These are the conditions where bilateral deal governance is most valuable, because they are the cases most likely to produce the conversations the brand safety team needs to have.
Post-campaign, brand safety teams should build DealSheet retrieval into their standard campaign review process. Retrieving the committed record and comparing it to delivery data is a structured audit that identifies gaps rather than relying on vendor-supplied reports.
Frequently asked questions
Does bilateral deal governance replace impression-level verification?
No. Bilateral deal governance adds a commitment layer upstream of impression-level verification; it does not replace it. Both are needed: the deal record establishes what was agreed; verification confirms what was delivered.
Can brand safety teams access DealSheets directly?
In Alkimi's implementation, the DealSheet is queryable by any credentialled party to the deal. Brand safety teams with appropriate access can retrieve deal records directly rather than requesting data from the buying team.
What happens if a publisher breaches brand safety terms in a bilateral deal?
The committed deal record establishes the standard against which the breach is measured. With a bilateral record, the buyer can demonstrate precisely what was committed and what was delivered, making the case for remediation materially stronger than in a dispute based on DSP logs alone.
Further reading
AAMP : IAB Tech Lab Agentic Advertising Management Protocols
Concourse : Agentic Advertising Platform
A2A Protocol : Agent-to-Agent Communication Specification