TL;DR: Not all agentic advertising platforms offer the same level of deal record transparency. For buyers, the difference between a platform that holds the authoritative record and one that only reports against it is significant: one gives you audit rights, the other gives you a dashboard. This piece sets out the transparency questions buyers should use as an evaluation framework, the technical markers of a genuine deal record, and the red flags that suggest a platform is not yet ready for enterprise spend.
Deal record transparency is becoming the central question in agentic advertising procurement. As agent-to-agent transactions move from pilots to production budgets, buyers and their procurement teams are discovering that "transparent" means very different things depending on which vendor is using the word. Some platforms provide access to a shared, immutable record of what was agreed. Others provide reporting interfaces that show what happened, but do not give buyers independent access to the underlying record. That distinction matters more than almost anything else in the vendor selection process.
This is an evaluation framework, not a vendor ranking. The criteria below are technical and contractual, and they apply to any platform that claims to support agentic media buying.
What is a deal record and why does it matter?
A deal record is the structured document that captures what was agreed between a buy-side agent and a sell-side agent at the point of negotiation. It is distinct from a delivery report, which records what was served, and distinct from an invoice, which records what was billed. The deal record is the reference document for all subsequent reconciliation.
In traditional programmatic, the closest equivalent is the insertion order: a document that describes the agreed terms, signed by both parties. The insertion order is often a PDF, sent by email, and stored in a CRM. It is not machine-readable, not updated in real time, and not accessible to both parties simultaneously from a single authoritative source.
In agent-negotiated deals, the deal record should be machine-readable, timestamped at execution, and accessible to both the buy-side and sell-side without either party being able to alter it after the fact. The technical term used by the IAB Tech Lab's Agent Communication Protocol specification is a "bilateral transaction record": a record that is held in a shared state rather than in either party's own system. The significance of this is that neither party can produce a different version of what was agreed. The record is the record.
For buyers, this matters for three reasons. First, it enables genuine post-campaign reconciliation: you can compare your agent's delivery against the agreed terms without relying on the platform's own reporting. Second, it reduces the scope for fee disputes: the agreed price is in the record. Third, it creates an audit trail that can be produced to a compliance team or an external auditor without depending on the vendor to reconstruct it.
What questions should buyers ask about deal record access?
The following six questions form a practical evaluation framework. They should be asked directly of any platform before a procurement decision is made.
Where is the deal record held? The answer should name a specific system, not a reporting interface. If the answer is "in our platform," the follow-up question is whether the buyer can access it independently or only through the platform's own dashboard.
Who holds the authoritative copy? In a bilateral record, neither party should hold the authoritative copy independently; the record should exist in a shared state that neither can edit. If the vendor holds the master record and provides buyers with a read copy, the buyer does not have audit rights; they have reporting access.
Can the deal record be exported? On request, in a standard machine-readable format (JSON or XML), within a defined timeframe? If the answer involves a manual export process or a bespoke request to the vendor's technical team, the access model is operational, not systematic.
Is the deal record immutable after execution? No field in the deal record should be editable by either party after the deal is executed. Any change to deal terms after execution should generate a new record, not overwrite the existing one, and the change log should be accessible to both parties.
What is the retention period? Deal records should be retained for at least as long as the applicable financial audit cycle. In the UK, the Companies Act 2006 requires financial records to be retained for six years from the end of the financial year in which the transaction occurred. A platform that deletes records after 12 months is not fit for enterprise procurement.
What access does the buyer have if they leave the platform? This is the portability question, and it is the one most vendors do not have a clear answer to. If a buyer moves to a different platform, can they export their complete deal record history? Without portability, historical audit rights are contingent on maintaining the vendor relationship.
What are the technical markers of a genuine deal record?
Beyond the contractual questions, buyers with technical teams should look for specific markers in the platform's architecture.
A genuine deal record will include: a unique deal identifier generated at execution; a timestamp in a standard format (ISO 8601); the agreed price, including any breakdowns between net media cost and fees; the inventory specification, including placement, format, and targeting parameters; the mandate reference, linking the deal back to the buyer-approved mandate under which the agent acted; and the parties to the deal, identified in a way that can be verified independently.
A platform that cannot provide all of these fields in its deal record specification is not offering genuine transparency; it is offering reporting. The difference matters because reporting is curated by the vendor, while a deal record is generated at execution and should not be subject to curation.
The distinction is analogous to the difference between an ad server's delivery log and a verification vendor's independent measurement. Both tell you something about what happened. Only one is independent.
What are the red flags in a vendor's transparency claims?
Several patterns in vendor conversations are worth treating as caution signals.
Conflating delivery reporting with deal record access. A vendor that answers transparency questions by demonstrating their reporting dashboard is not answering the question. Reporting shows delivery; a deal record shows what was agreed. They are different documents serving different purposes.
"Full transparency" claims without specifics. Transparency is a technical property, not a marketing claim. A vendor who says their platform is "fully transparent" without specifying the access model, the retention period, or the export mechanism is using the word as a brand attribute rather than as a description of a system capability.
Bilateral access described as a future roadmap item. If deal record access for both parties is on the roadmap rather than in production, the platform does not yet support bilateral transparency. Buyers running production budgets on such a platform are taking on audit risk in exchange for early access.
No answer on portability. A platform that does not have a clear answer to the portability question has not designed for buyer audit rights. This may not be malicious; it may simply reflect that the platform was designed for buyers who remain on the platform. Either way, it is a constraint worth knowing before commitment.
How to weight these criteria in a procurement decision
The six questions above are not equally weighted. For buyers running enterprise budgets through agentic deals, deal record access and immutability are non-negotiable. Retention period and portability are significant but may be acceptable to manage contractually if the vendor will agree to binding terms. Export format and timeline are operational preferences that can be negotiated.
The red flags are useful as screening criteria rather than disqualifiers. A vendor that conflates delivery reporting with deal record access may be genuinely early in their transparency build; a follow-up question about their roadmap and their contractual commitments will tell you whether they are building toward genuine bilateral access or not.
The honest summary is that deal record transparency is a spectrum. Some platforms have built bilateral record infrastructure from the ground up. Others are adding transparency tooling to an existing delivery-first architecture. The buyer's job is to understand where on that spectrum each vendor sits, and to ensure the contract reflects the actual access model, not the marketed one.