Vercer
Built for small to mid-size businesses that rely on disconnected tools and manual processes but lack the technical resources to build AI infrastructure on their own. Vercer understands how a business operates and constructs the solution itself.
Discovery
Observed the same pattern repeating across businesses of every size — data living in five different places, reports assembled by hand weekly, workflows that only run because someone remembered to do them. Researched Zapier, Make.com, n8n, and Latenode — none combine natural language intake, an AI planning layer, and an AI build-and-deploy layer in a single product for non-technical users.
Wireframe
Designed two distinct modes: Plan Mode, a structured AI conversation that asks about the user's problem, tools, triggers, desired outcomes, and edge cases — with every message vectorized and stored per conversation so the AI builds a complete picture before building anything. Build Mode hands context to a second AI agent that constructs the workflow JSON, connects to the user's tools via n8n, deploys it live, and renders the result as a visual node graph using React Flow.
Prototype
Built the solutions architect AI prompt system first — defining how the AI asks questions, what it needs to know before building, and how it signals readiness. Defined the Supabase schema with pgvector setup and seeded the tool knowledge base with native n8n integrations. Added a Connect Tools credential management dialog and a deploy state machine with explicit states — draft, connecting, ready, active, unpublish — so the UI always reflects exactly where the user is.
Ship
Currently in active development with four pilot customers being onboarded and interviewed for validation. Key features include AI-guided Plan Mode with vectorized memory, Build Mode with RAG-based context retrieval, n8n workflow construction and deployment, React Flow visual rendering, and full lifecycle state management from draft to active automation.
