Turning a Newly Built Text-to-Video Model into a Launched Product
Company: Google
Role: Product Manager
Category: Product Design & Strategy
Difficulty: medium
Interview Round: Other
Your team has just built a new model that generates video from a written prompt. You are the product manager responsible for taking it to market. Walk through how you would turn the model into a product, from deciding who it serves to launching it.
```hint Let the model's limits pick the first market
Pin down what the model does well and badly today (clip length, consistency, speed, cost per clip), then look for a group of users whose real video work fits inside those limits and who would pay for it.
```
### Clarifying Questions
- Will the model ship through the company's existing products and channels, as a new standalone product, as a developer API, or is that choice part of the answer?
- What has the research team established about its capabilities so far, such as maximum clip length, resolution, generation time, audio support and how often outputs are usable?
- Do the training data and its licensing limit which commercial uses the product can offer?
- Is success judged mainly by revenue, by reach, or by establishing a position against other labs' video models?
### What a Strong Answer Covers
- A capability assessment that sets what the product can honestly promise, compared against competing video models
- A choice of first segment and job (for example creators, marketers, film professionals, developers or casual users), justified by fit with the model, depth of pain and willingness to pay
- Product form and distribution, with what ships first and why
- Safety, likeness, provenance and copyright treated as launch requirements
- Pricing that reflects the high compute cost of video, a staged rollout, success metrics with guardrails, and the signals that would change the plan
### Follow-up Questions
- Each video generation costs far more to serve than a text or image request. How does that shape your free tier and pricing?
- A rival lab releases a comparable model a month before your launch. What do you change?
- During the beta, a realistic fake clip of a public figure made with your product goes viral. What do you do in the first 48 hours?
Overview: A product strategy prompt asking how a product manager would take a newly developed text-to-video generation model to market. It tests choosing a first segment, product form and distribution, safety and likeness safeguards, pricing under high compute costs, and a staged launch with clear success metrics.