Unicorn Ideas: AI Prompt Analysis | Superblocks CEO
Uncover the secrets of AI innovation with Superblocks CEO Brad Menezes! He reveals how mastering system prompts is the key to unlocking the next generation of AI applications.Learn how role, contextual, and tool-use prompting strategies are revolutionizing advancement. Superblocks, backed by $60 million in funding, is empowering non-programmers with “vibe coding” tools, enabling them to build enterprise apps. Discover their approach to “prompt enrichment,” the art of building infrastructure around large language model calls. This approach, currently used by Instacart and Paypal, helps businesses build AI agents. News directory 3 has the scoop! what strategies will emerge to help organizations find even better ways of building apps? discover what’s next for AI-driven development.
Superblocks CEO Reveals AI Startup Secrets: System Prompts and Vibe Coding
Updated June 7, 2025
Brad Menezes, CEO of Superblocks, believes the key to the next wave of AI innovation lies in understanding the system prompts used by today’s leading AI startups. These prompts, often thousands of words long, guide foundational models from companies like OpenAI and Anthropic in generating application-level AI products. Menezes views them as a master class in prompt engineering.
While not always publicly available, these system prompts offer valuable insights. Superblocks, in announcing its new enterprise coding AI agent Clark, shared a file of 19 system prompts from popular AI coding products. This move gained meaningful attention, wiht Menezes’s tweet going viral and viewed by millions.
Superblocks recently raised $23 million in Series A funding, bringing its total to $60 million. The company focuses on “vibe coding” tools designed for non-developers in enterprises.
Menezes said that the system prompt itself is only about 20% of the ”secret sauce.” The remaining 80% involves “prompt enrichment,” the infrastructure built around calls to the large language model (LLM). This includes instructions attached to a user’s prompt and actions taken when returning the response,such as accuracy checks.
Roles, Context and Tools
Menezes identified three key areas for studying system prompts: role prompting, contextual prompting, and tool use. He emphasized the need for specificity in these prompts,akin to instructing a human co-worker.
Role prompting helps LLMs maintain consistency by providing purpose and personality. Contextual prompting gives models the necessary background, reducing costs and ensuring clarity. Tool use enables agentic tasks,allowing models to go beyond simple text generation.
By studying system prompts, Menezes observed different emphases among vibe coders. some prioritize fast iteration, while others focus on creating full-stack applications. This led to Superblocks’ focus on enabling non-programmers to build apps by handling security and access to enterprise data sources.
Superblocks’ customers include Instacart and Paypaya Global. The company also uses its own product internally, with buisness teams building AI agents for tasks like lead identification and support metric tracking.
“This is basically a way for us to build the tools and not buy the tools,” Menezes said.
What’s next
Superblocks plans to continue refining its AI-powered platform, focusing on expanding its capabilities for enterprise users and further simplifying the app development process for non-programmers through innovative use of system prompts and vibe coding techniques.