Why we're building Kili

We're excited to announce that Kili is now open to everyone. Kili helps you automate repetitive tasks. We wanted to write a short note explaining our vision and why we're excited about it.

Large language models (LLMs) will help us automate many of the tasks we do today. LLMs provide intelligence on the fly, and we believe we are only scratching the surface of what's possible with this technology.

Importantly, we want this technology to be available to everyone. Most of the work we do today is tedious and repetitive. Teams rely on their engineering counterparts to automate processes, or rely on software that is available off the shelf. With the power of LLMs, we believe it's now possible to build an automation within minutes. This means software can be customised and built for what you need rather than the other way around - you having to mould to what the software does under the hood.

To achieve our vision, we believe the user interface is critical. We've had no-code and low-code apps for many years. They've grown in popularity but are limited either by inflexibility, their complexity or both. Kili is our attempt to build the next version of this using LLMs from the ground up.

We believe chat interfaces are great for their simplicity. They do not work for complex tasks. On the other hand, most people think visually. Visual node-based user interfaces attempt to solve for this but are typically complex with many elements and buttons. Our user interface combines the best of both worlds. It allows users to review their task from a visual point of view, and leverage our AI assistant to build or manage their automation.

Our assistant and the visual canvas share the same information. This means the assistant can take actions for you that update the canvas. The assistant can zoom in and answer a question about a specific step, or zoom out and figure out what you should do next to achieve the goal of your automation.

Why we're building Kili

Extract data from documents