.do Business-as-Code

AI is getting extremely good at skills like math & code, because the outputs are verifiable. By representing your Business-as-Code, you can leverage the power of AI to automate, optimize, and scale your business processes.

AI

Workflows let you define business processes as clean & simple code. Functions can also execute code, use AI to generate structured data, or assign tasks to be performed by humans workers or autonomous Agents.

Business

Goals are the strategic business objectives you want to achieve, and KPIs allow you to measure progress towards achieving them. Experiments allow you to rapidly iterate and improve your business processes, while Projects help you Plan how to get there. Tasks are the day-to-day work delivered by humans and agents.

Data

Nouns are the categories or types of things in your system, Verbs are the actions that can be performed, Actions are the subject-predicate-object relationships, Resources are the instances of nouns in your system, and Integrations are the connections between your system and external systems.

Experiments

Experiments provides a powerful framework for testing hypotheses, measuring outcomes, and iteratively improving your AI applications. It enables you to run controlled experiments to validate ideas, optimize performance, and make data-driven decisions. Datasets are the inputs to your experiments, Evals are the outputs, and Benchmarks are the metrics used to measure success.

Integrations

Integrations provides a framework for connecting your system with external services and APIs. Triggers are the events that initiate workflows, Searches are the queries that retrieve data, and Actions are the actions that perform tasks.

Models

Models are the AI models used to generate responses to prompts. LLM is the low-level model used to generate responses, while GPT is the high-level model used to generate responses.