Two video-production bids can show the same headline price and still describe completely different purchases. One may include global usage, vertical cutdowns and two consolidated review rounds; the other may leave all three as add-ons. A buying agent can compare them quickly only if those conditions are stated in a consistent form. It still cannot decide which creative idea deserves the brand.
That distinction matters for overseas brands working with China-linked supply chains. Product availability, sample timing, localization, usage rights and approval dependencies can change the real cost of a film more than the first number on a quote.
The signal is moving upstream—but its scope is narrow
On September 22, 2026, IAB Tech Lab updated AAMP 3.0 to extend agentic advertising into the RFP or brief stage of digital media. Its OpenProposal work is intended to help buying agents parse and evaluate what is being bought and sold, from the initial brief toward commercial terms.
The important boundary is in the source itself. The OpenProposal repository says the specification is a draft, not a ratified standard. Its public-comment window runs from September 22 to October 22, 2026, and the machine-readable schemas were not yet published with the first draft.
It is also a media-planning object, not a standard for selecting film-production companies. Applying the same design logic to production procurement is ONCE's inference. The practical signal is not “automate vendor selection now.” It is that proposals will become easier for agents to inspect when commercial facts are separated from persuasive presentation.
Why conventional production bids are difficult to compare
A polished treatment deck is good at communicating taste. It is often poor at exposing dependencies. “One hero film plus social cutdowns” does not tell a buyer how many versions, which aspect ratios, which languages, which usage term, how many feedback rounds or who pays when samples arrive late.
Production-side guidance points to the same operational problem. Griffith Media's May 2026 proposal guide recommends defining exact deliverables, pre-production, production and post scope, schedule, review process, rights, file handling, assumptions, exclusions and total cost. Its examples show why versioning is work: reframing, captions, timing, graphics, music and quality control do not happen automatically.
A procurement agent can normalize fields that exist. It cannot safely infer terms a supplier never stated. If a comparison workflow rewards the lowest visible number, the most incomplete bid may look best.
The ONCE three-layer bid: Terms, Proof, Treatment
For creative production, the goal should be structured comparison without creative flattening. We recommend keeping three layers distinct.
1. Terms: facts an agent may normalize
This layer should use comparable units. Include the number and duration of master films; aspect ratios; languages and subtitle files; cutdowns; shoot days and locations; delivery formats; milestones; feedback deadlines; included revision rounds; change-order triggers; usage territories and terms; ownership of raw footage and project files; and client-supplied dependencies such as products, claims and approvers.
Each field should be marked as included, optional, assumed or excluded. A price or consequence should accompany every optional or assumed item. The Terms layer helps a buyer identify questions; it does not replace the contract.
2. Proof: evidence the supplier can deliver
Proof should connect claims to relevant material rather than present an undifferentiated portfolio. Which project resembles the product, market or delivery complexity? Who will actually direct, produce and finish the work? Who obtains and records talent, music, stock, AI or location permissions? How will the team respond to a late prototype, regulatory review, weather disruption or a new language version?
An agent may check completeness and highlight contradictions. It should not score capability from the number of case-study links or from keyword overlap alone.
3. Treatment: judgment that belongs to people
The Treatment explains why a narrative, camera language, performance direction and production method serve this brand. It should reveal the trade-offs: what changes first if the schedule tightens, what can move to AI-assisted production, and what must remain live action or human-led.
Creative fit, cultural sensitivity, trust and the final award require human review by marketing, creative and procurement. An unconventional idea should not be rejected merely because it is harder to map into a historical scoring model.
Compare total proposal cost, not the cover price
A useful normalization formula is:
Comparable total proposal cost = base production + mandatory options + versioning and localization + rights and licenses + review and revisions + dependency-delay risk + exit and rework cost.
Suppose Bid A excludes music, global paid-media rights and vertical edits while Bid B includes them. Comparing only base production is not a price comparison; it is a scope mismatch. The formula makes the hidden difference visible without pretending that creative quality can be reduced to one number.
What the buying agent may do—and where it must stop
Start with four low-risk tasks: normalize deliverables into common units; flag missing fields; recalculate comparable total cost; and produce a supplier-clarification list.
Keep three decisions behind human approval: rejecting a nonstandard treatment, using historic lowest prices as a creative ceiling, and sending a counteroffer or award notice. The human gate becomes more important for new-market launches, regulated claims, celebrity or synthetic talent, complex rights and large media commitments.
McKinsey's February 2026 analysis of agentic procurement describes a similar division: agents can analyze supplier bids and prepare negotiation materials, while people retain creative problem solving, relationships and complex judgment. That is a consulting view, not proof that every procurement team has deployed the model.
A practical pilot for brands sourcing production
Do not wait for an industry schema. Keep the human-facing Treatment deck, require every vendor to complete the same Terms appendix, and give the Proof layer a fixed index. Run the agent in shadow mode on one real procurement cycle: let it normalize and flag, but do not let its output change the official score.
Record the omissions, false positives and questions that still require a producer. Once the fields are stable, changes are traceable and creative review remains human-led, the structured appendix can connect to a procurement system.
The purpose is not to let software choose a director. It is to stop people spending senior attention on avoidable scope mismatches.
If you are preparing a cross-border brand film or product-content program, explore how ONCE approaches reusable brand inputs for platform-generated ads and the human boundary for content agents. To align the brief, deliverable matrix and proposal structure for an upcoming project, start a conversation with ONCE.
Production examples and next steps
Use the TVC brief to connect the selling point, performance and product presentation. The Red Devils × Tencent Games commercial is a finished-film reference.
Review TVC and commercial production and Red Devils × Tencent Games: A Passionate Competitive TVC for Young Players. To discuss your project, send the script, storyboard, required formats and approval dates through the English project enquiry form.
