DearFlow approach to automating work

For an AI solution to be truly useful to a business, it needs to be both reliable and affordable.

Multi-Agent Workflow Frameworks

Multi-agent workflow frameworks like Autogen and CrewAI inspired the creation of DearFlow. While we appreciate the autonomous agent approach, we chose to address work automation differently.

We've found that a fully autonomous agent approach is currently too unreliable and costly, especially for business use cases or specific tasks that need context and integrations. Our approach to automation is different but could potentially work alongside these autonomous frameworks in the future.

Reliability

Reliability is what separates a fun demo from a valuable product, especially for business applications.

Businesses usually seek to automate critical tasks essential for day-to-day operations. If a workflow wasn't crucial, they wouldn't invest time and money in exploring AI solutions. Therefore, there is very little room for error.

Cost

AI solutions must not only match the cost of a human employee but be significantly cheaper.

If costs are similar, hiring a human is often more practical. Humans are versatile—they can fill gaps in various departments, contribute to company culture, and assist with physical tasks. AI solutions need to be affordable enough to justify any initial setup friction.

Less AI Can Be Better

This might seem counterintuitive, but our experience in the AI field drives this perspective. We only use AI when it is absolutely necessary in a workflow.

A typical DearFlow automation is about 20% AI and 80% infrastructure. By infrastructure, we mean the loading, manipulation, and writing of data around these AI nodes. Using standard software for the infrastructure ensures that AI nodes receive the right context to perform their tasks, and we can verify and correct their outputs.

Autonomous Agents & DearFlow

We see an exciting future where autonomous agents use DearFlow automations as tools to complete tasks. DearFlow provides a framework for defining tasks with clear beginnings and ends. Autonomous agents could determine when to use these automations, how often, and with what inputs. This way, DearFlow automations become tools in an agent's toolkit, enabling reliable action and solving complex business problems.

Our Solution

This is the state of our automation building UI as of June 22rd 2024

Our automation building approach lets users explicitly define the exact order of operations for a task, leaving no room for error. Think of it like following a GPS with turn-by-turn directions—you'll always reach your destination.