Skip to content

Latest commit

 

History

History
110 lines (75 loc) · 5.18 KB

why-proficient.mdx

File metadata and controls

110 lines (75 loc) · 5.18 KB
title
Why Proficient AI?

With a little bit of understanding of how Proficient AI works, a natural question that may arise is why you'd want to rely on it instead of implementing everything yourself. While you certainly can implement your custom solution, there are several good reasons why you should consider using Proficient instead.

TL;DR

  • an end-to-end solution, with which it takes 3 minutes not weeks to get a user-facing agent up and running in your app
  • minimizes or eliminates the need for custom backend infrastructure so you can focus on implementing the business logic
  • powerful tools built into the dashboard and Admin API including analytics, monitoring, rate-limiting, content moderation, etc.
  • technology-agnostic solution that supports multiple LLM providers allowing you to easily switch between models with 1 click
  • ready-to-use, highly customizable and beautiful UI components rendering complex interaction trees with support for advanced features like streaming

and much, much more.

{/* TODO: Improve */}

{/* ## Key Benefits

No fixed cost for integration

As little as 3 minutes.

Implementing LLM-based AI assistants or chatbots currently involves a pretty significant fixed cost. For an end-to-end solution, you need to:

  • write a collection of backend systems to communicate with LLM providers and moderate the generated content
  • write database code to accommodate user-assistant interactions
  • write UI components and connect them to your backend
  • optimize config values (prompt engineering, temperature adjustments, etc.)

This process can be lengthy, error-prone, and rigid, often requiring new deployments for incremental updates. Developers face similar challenges when implementing solutions with essentially the same core functionality across different domains.

Product velocity

No infrastructure maintenance minimizes or eliminates the need for custom backend infrastructure so you can focus on implementing the business logic. Proficient is doing a lot of the heavy-lifting. Your agents will always be powered by the latest technology without you doing any of the work.

Access to powerful admin tools

Dashboard and API. Gain valuable insights and data through our admin dashboard, allowing you to monitor, analyze, and optimize your agent's performance.

Over-the-Air Config Updates: Make changes to system messages, content moderation, and much more without needing a new deployment.

Integration with multiple LLM providers

We'll give you access to a plethora of models.

Multi-platform support

Low-to-no-effort integration with other platforms using our SDKs

Powerful abstractions

Proficient provides the crisp abstractions that programmers love.

LLMs are implementation details, agents are interfaces. Programmers should rely on interfaces.

When not to use Proficient

Proficient is not optimized for all types of LLM applications. You still can use Proficient but it wouldn't bring huge value.

  • when the application has no interaction mechanisms between agents and users, it's simply text in -> text out.

What does it solve?

With the explosion of ChatGPT, more companies have been adopting AI chatbots. But implementing an AI-powered chat system currently involves a pretty significant fixed cost. For an end-to-end solution, you need to:

  • write a collection of backend systems to communicate with AI providers and moderate the generated content
  • write database code to accommodate user-assistant interactions
  • write UI components and connect them to your backend
  • optimize config values (prompt engineering, temperature adjustments, etc.)

This process can be lengthy, error-prone, and rigid, often requiring new deployments for incremental updates. Developers face similar challenges when implementing solutions with essentially the same core functionality across different domains.

Proficient standardizes the structure of human-AI interactions by offering an API for abstract AI agents, along with a suite of well-tested tools. Our solution streamlines the integration process, reducing the fixed cost and time commitment for developers to just 3 minutes.

We believe that as LLMs become commoditized, companies that use these technologies will soon have plenty of options to choose from. Instead of integrating with each LLM provider separately, they should rely on abstract, technology-agnostic, and easily configurable AI entities. Granted, some businesses may require more granular access to the underlying models but there is an entire class of apps that need to operate on a higher abstraction layer using AI agents. Proficient will lay the foundation for these entities, and post-launch, we plan to continually expand their capabilities.

Roadmap

Talk about what we'll build. Agents capabilities etc.

  • browsing and integration with other services (like ChatGPT plugins)
  • more frontend SDKs (iOS, React Native, etc.)
  • more providers including open-source Hugging Face models
  • admin to allow fine-tuning/further RLHF
  • agent configuration presets and end-user templates to allow reusability
  • custom algorithms to efficiently fill LLM context windows */}