A launch under embargo, dozens of markets: how to choose the production model

The context

A new model has a reveal date that does not move. Until that day nobody outside the company may see it, and inside the company only those who need to. Meanwhile the launch has to be shot, and then taken to dozens of markets, each with its own language, formats and calendar. For a global automotive manufacturer this was routine: production decided centrally, product often still under embargo, and a far larger body of work starting afterwards, in versions, adaptations and transcreation.

The challenge

There were two questions. For the shoot: traditional, virtual, AI or hybrid? For what follows: how do you optimise cost to multiply versions, adaptations and transcreation across every market without losing quality or control? Meanwhile each supplier used AI in its own way, and a product under embargo made every new tool one more question.

The approach

The first question looked technical and was one of judgement. The model is not chosen on principle: you choose it by looking at what, in that shoot, has to stay real. The second hid another truth: cost did not sit in the shoot, which is one event, but in the multiplication that followed. We did not bring a ready-made model. We brought our framework, four questions always asked in the same order (what you create, how, with which tool, with which guardrails), and shaped it to the manufacturer’s launches.

The work

For the main shoot we chose case by case between traditional and virtual, according to what had to stay real. For versions and adaptations, where the work follows a standard, we designed a central, AI-assisted flow. Choosing tools became a choice about data: licence terms count, and so does the place where the not-yet-public product lives, and we use only authorised platforms. We updated contracts with clauses on the use of AI, on the image of people and on confidentiality, extended to the suppliers’ own suppliers.

Keeping it in control

We started from a real launch as a pilot, with thresholds written in advance and a decision at the end: extend, correct or stop. Then adoption passed through contract renewals. AURA suite kept the project in one place and automated the repetitive steps of diagnosis and operations.

The impact

The manufacturer now has a decision path that is the same for every launch, clear rules on where data lives under embargo, and contracts that cover AI along the whole supply chain. Whoever executes does not measure: results control stays independent from the supplier. Before starting, we wrote an impact forecast, and the comparison with results is read after a year of operation, across several launches.

Do you recognise this?

When teams have to choose a production model, the discussion almost always starts from technology: what AI can do, what virtual costs. It is the wrong starting point. The choice depends on what, in that job, has to stay real, and the biggest cost almost always sits after the shoot, in the multiplication.

It happens wherever a master has to go to many markets, not only in automotive. For each job, can you say what has to stay real and what you can derive? If you have a secret to protect, do you know where the data of the tools your suppliers use lives?

If you are not sure of either answer, it is worth seeing where it happens in your business before talking about how to address it. We do not know your work as well as you do. That is why we start with a short meeting, with no commitment, on a single launch.

Book a short meeting. We look together at one job family in your business and see whether the knot is the same.

Common questions