Runway launches Agent 2.0 for briefs, campaign assets, and performance analysis

• Runway says Agent 2.0 turns a prompt into a full marketing workflow, from briefs to campaign assets, inside Runway itself, according to runwayml's Agent 2.0 announcement.
• In c_valenzuelab's launch post, Runway CEO Cristóbal Valenzuela framed the jump as moving from single-model clip generation to finished videos you can get by describing the result you want.
• The launch pitch pairs generation with iteration: runwayml says Agent 2.0 can analyze performance data, then adapt creative across platforms, formats, and markets.
• Runway had already shipped adjacent ad tooling, and runwayml's ad localization post shows one-click language localization from a single source image.

You can [get started through Runway's linked product page]Get started with Agent, watch the launch demo, and compare the new agent pitch with Runway's earlier ad localization demo, which already pushed the platform toward multi-market campaign work.

Runway is pitching Agent 2.0 as a higher-level creative system, not just another video generator. The product claim in runwayml's post is specific: a user starts with a simple prompt and gets marketing briefs plus campaign assets without leaving the tool.

Valenzuela's phrasing in his launch post adds the intended category shift. He compared the jump to coding models that moved from autocomplete to writing working software, then said that same moment has arrived for video.

The workflow Runway is describing breaks into three steps:

1. Start from a prompt.
2. Generate a marketing brief and creative assets.
3. Iterate from performance data and scale the outputs across channels.

The unusual part of this launch is that Runway is not stopping at asset generation. In runwayml's announcement, the company says Agent 2.0 can analyze performance data to improve creative, which pushes it closer to campaign operations software than a pure media model.

That matters for creative teams because the surface area is broader than making one hero asset. The feature set Runway names in the announcement spans briefs, asset production, analysis, and scaling across platforms, formats, and markets.

Runway has not said much more in the tweet evidence about how that analysis works, but the product entry point is live through the Agent signup page.

Before Agent 2.0, Runway had already started bundling campaign adaptation tools into the product. In runwayml's earlier post, the company said users could input one ad image and generate versions for every market in any language with one click.

That older feature makes the new Agent 2.0 pitch easier to read. Runway is combining two layers that were often separate: generating the campaign itself, then localizing and extending it for distribution.

The ad localization rollout also came with its own live product link in runwayml's try-it-now post, which suggests this marketing workflow push is not a one-off demo but an ongoing product direction.