In a “short” sentence:
ActionBook uses generative AI powered “Agents” to generate automation code, help user to author in a “Web Editor”, then put them in a hosted “Online Container” to run.
Task automation is much larger than RPA + Zapier + Selenium.
– Generative AI (ChatGPT and likes) models made it possible to generate code auto-smartly.
– But current AI models aren’t enough to generate useful code due to
Base on generative AI as “Brain“, Agent added “Memory“, “Tools” to solve knowledge store/fetching, interact with various apps/internet and multi-round planning/reasoning towards a single goal.
Such as write a working automation code.
Traditional RPA or code editor need more, an IDE/install some library/setup environment variable, etc.
It’s cumbersome, we know.
So we build everything into a web browser from new/editing to run/debugging.
Nothing needed other than a browser.
We have quite some experience in automation. One thing repeatedly shoot us in the foot is “Environment”.
Moden OS and apps are all “stateful”, meaning they have “states”. Thus different ways of interacting with the system, will have different results.
Automation depends on the state, so a “clean” environment is essential to a stable automation and can save you more than 2/3 of your time trying to figure out what’s wrong.
So, we provide a “clean” online container every time your automation starts. No headache.