$9M+ · AI Automation

What they're building

CodeWords is not a workflow builder. That distinction is the whole product thesis. Zapier, Make, and n8n require the user to understand trigger logic, map fields between systems, and maintain the workflow when integrations break. CodeWords removes that layer entirely. You describe what you want done. Cody, the AI agent at the centre of the platform, understands the business context, connects the relevant tools from 3,000+ integrations, writes Python logic to handle branching and API calls, deploys the workflow on CodeWords' own infrastructure, and keeps it running 24/7 without the user touching a config file. The output is not a prototype. It is a production automation. The platform has three main surfaces. Meet Cody is the core agent: it asks clarifying questions before building, shows you its plan as a visual flow, and only runs after you have approved what it intends to do. CodeWords UI turns descriptions into branded live apps, not just backend automations. WhatsApp Agents lets Cody reach users over WhatsApp for approvals, updates, and human-in-the-loop steps in any workflow. The agent runs on real Python under the hood, branches, transforms, and API calls, rather than a fixed menu of no-code blocks. Alongside the seed announcement, CodeWords launched three new capabilities: Contextual Memory (Cody learns from past activity and institutional context to proactively suggest and build automations before being asked), WhatsApp support, and Cody Modes (which adapt how the agent plans and executes based on the task type and required reliability level). The company runs 500,000 workflows per month in live production environments. Customers include agencies, go-to-market operators, automation studios, and individual non-technical professionals. A two-person audiovisual studio with no coding background used CodeWords to build a full social content pipeline and onboard paying clients. An automation agency built a fleet of agents to deliver lead generation services to their client base. These are not demos.

Why this matters

The no-code automation market is not a small market. Zapier processes hundreds of millions of tasks per month and has been valued at over $5B. Make (formerly Integromat) serves millions of users across Europe and the US. n8n has been growing rapidly as the open-source alternative for technical teams. The market has validated that automation demand is enormous. What it has not solved is the human expertise still required to use these tools. Zapier is easier than code, but it still requires the user to design the logic, understand the trigger-action paradigm, and debug when something breaks. For non-technical users, that gap is still too wide. CodeWords's bet is that natural language is the correct interface for that market, not a simplified GUI. The user describes the outcome. The agent infers the logic. That shift unlocks the population of users who have automation problems but no automation skills: the marketing manager, the solopreneur, the small agency owner, the sales operations lead who knows what they want done but not how to wire it up. The first Firstminute Capital investor quote from the original launch captures the scale of this framing: "enabling billions of people to create software." That is not a SMB-tools company. That is an infrastructure company for the AI-enabled economy. The angel roster is a secondary thesis. Ilkka Paananen (Supercell), Robert Gentz (Zalando), and Mati Staniszewski (ElevenLabs) are not casual investors. They are founders who have built the largest European consumer and B2C software companies of the last two decades. Their checks are bets that CodeWords becomes the automation layer for the next generation of those companies, built by people who could not have built them before.