LOLA — an AI project manager for creative teams
LOLA is a working AI agent I designed, built, and deployed through Technovation, a global tech education nonprofit where I lead creative technology.
Here's the problem she solves. Every creative team has a doorway: the place where requests come in. Forms, DMs, hallway asks, the "quick favor." Before any design work starts, someone has to read each request, figure out what it actually is, find the right owner, and set a deadline. That triage job quietly eats hours every week, and the person stuck doing it is usually the same person the work is waiting on.
LOLA takes over the reading and matching. When a request form comes in, she sends it to Claude through the Anthropic API along with a routing rulebook the team writes in plain language. She proposes an owner, a project, and a priority, and she cites the exact rule that produced the decision. Then she waits. A human approves, edits, or rejects every proposal in one click. If a request doesn't match any rule, she stops and flags it instead of guessing.
She shipped in propose-only mode on purpose. Autonomy is something a team can grant later, rule by rule, once trust is earned. The most advanced thing about this agent is the restraint built into it.
The brand runs on the same logic as the architecture. The mark is a letter with a face inside it: the L is the structure, the chat bubble is the intelligence, and the structure contains the personality, exactly how the product is built. The palette is functional, not decorative. Warm gradients mean LOLA proposed something. Cool tones appear only after a person says yes. Cream marks the moments she asks for help. Anyone scanning a task board can read the system's confidence at a glance.
Built on the Anthropic API, Node.js, and Vercel. First integration: Asana. In production today on a marketing team serving 100+ countries.