What did Google change at I/O 2026 that threatens some AI startups?
Google embedded advanced models and agent capabilities across Search, Chrome, Workspace and Android, turning many generic agents, thin wrappers, and simple app builders into free, built-in features with massive distribution.
Who is most at risk from this shift?
Founders whose products are 'below the line' — generic access to models, single-prompt apps, or thin interface wrappers without proprietary data, workflows, or trust — face direct competition from Google's free features.
What is 'the line' and how should founders use it?
'The line' separates commoditized AI features (below) from non-commoditized value (above): industry-specific workflows, proprietary context, customer trust, and unique distribution. Founders should stop selling below-the-line features and instead build unique offerings above it using free tools as fuel.
What concrete actions can startups take right away?
Ship quick prototypes via prompt→product tools, set up market-monitoring agents to capture demand, automate admin with personal agents (e.g., inbox summaries), and repurpose content using creation tools — then package those systems into industry-specific services or workflows.
How can businesses still monetize AI after these changes?
By selling outcomes and bespoke systems: implementing market monitors, internal dashboards, custom workflows, or managed services that combine proprietary data, domain expertise, and trust — things Google can't easily commoditize.