The production-economics system of record advertising never built — and a proposal for the firms best positioned to help build it.
Each one took an operational function that ran through email and PDFs, became the place that function happened, and now collects on every project, trial, or transaction in their vertical.
The form on which commercial productions bid was introduced by the Association of Independent Commercial Producers in 1975 and last meaningfully revised in 2019. It ships as an Excel file or a PDF. A bid package today contains what it has contained for decades: a cover letter, a top sheet, a line-item breakdown, a calendar, a list of assumptions. It travels by email. An agency producer reviews it and makes the call from memory. At the end of the job, a wrap memo is filed that nobody opens again.
This is not nostalgia.
The tools that produce the documents are everywhere. The system that connects them does not exist.
A production company underbids deliberately to win the work, knowing change orders will recover the margin. The brand pays the firm bid plus the breakage and signs off on a deliverable smaller than they thought they were buying. Or the inverse: a job gets overbid with inflated rates and the firm-bid model leaves client money on the table. The agency producer may or may not catch it. This is open knowledge. No system measures it.
A bid omits something the job will need — wardrobe, a second AD, art department for the props on screen — and the gap surfaces in production as overage or scope compression. The producer absorbs the political cost. The system that incentivized the omission absorbs nothing.
A holding company production lead misses portfolio margin by several points and cannot trace the cause. The decisions that drove the loss happened at award time, in conversations no system captured.
The intelligence to prevent any of this exists in the room. It is in the producer's head. When the producer leaves, the intelligence leaves with them.
Holding companies and production houses sit on years of bids, treatments, schedules, wrap memos, and actuals. Most of it is organized in shared drives. Almost none of it is interpreted.
It is thrown away the moment the wrap actualization is filed.
We capture the operational data agencies, production companies, and cost consultancies already generate, structure it into a connected knowledge graph, and surface decisions in real time. The intelligence engine sits on top: a pattern library that links early signals to downstream outcomes, validated by domain experts, deployed inside the partner's environment.
The first surface is bid evaluation. Three production companies submit on a job. API surfaces the risks in each bid against historical comparables, with reasoning visible at every level. The producer's judgment remains the deciding voice. The system makes the decision faster, more defensible, and legible to the people upstream.
The architecture extends from there:
Before bids come in.
While spend is still steerable.
At the vertical-lead level.
For production companies bidding the work.
Rendered for cost consultancy engagements.
Every holding company already has that and it is commoditizing fast. Different layer of the stack.
That lives downstream, in the media-buying path. We sit in the bid-to-wrap economics, where the margin actually moves.
System integrators organize documents. We connect them. To a holding company that just paid for a data lake: you have the storage, now buy the intelligence.
AICP's ABID (2020) addresses transparency in submission. It does not interpret the bid against history. We are the layer above it.
AI has commoditized creative tooling, which means production economics — bid quality, scope discipline, breakage discipline — are the only durable advantages a holding company has left. Vertical leads know it. Cost consultancies know it. The infrastructure to deliver it does not yet exist. We are building it.
We built the underlying production intelligence architecture for film and television, where it is currently in active pilot on a 25-day scripted series. The structural problem on the advertising side is the same — firm bids meet operational reality, and the gap gets absorbed by mechanisms invisible in standard reporting — but the ontology, the buyer set, and the data are different. API is the advertising application of the same architecture, built deliberately and not alone.
We are selecting design partners — holding companies, cost consultancies, or production companies — to co-develop the first deployment. We bring the architecture, the engineering, and the methodology. The partner brings the data, the domain experts, and the questions worth answering. The output is an intelligence asset the partner owns, with an advantage that compounds with every job processed.