Cohere’s new chat API boosts conversational AI

Toronto-based startup Cohere, founded by former Google employees, has been making waves in the world of generative AI. Their focus on crafting AI-powered technologies tailored to businesses has drawn considerable attention. Now, Cohere is taking a big leap into the realm of AI Chatbots, and their latest offering has the potential to reshape how businesses engage with their customers and streamline internal operations.

Cohere’s journey in the tech landscape has been quite a remarkable one. Their team, composed of individuals with roots at Google, boasts deep expertise in artificial intelligence. They’ve put that expertise to work to develop innovative solutions for enterprises.

The latest addition to their toolbox is the Cohere Chat API, and it’s poised to revolutionize the way businesses interact with their customers and improve internal processes. This application programming interface allows third-party developers to create chat applications using Cohere’s proprietary large language model, known as Command.

In a recent blog post, Cohere emphasized that their Chat API simplifies the process of building reliable conversational AI products. Whether businesses want to create knowledge assistants or enhance their customer support systems, this API offers a versatile and powerful solution. It nicely complements Cohere’s existing suite of APIs for content generation and text summarization.

One of the most exciting aspects of Cohere’s new offering is the Coral Showcase. It’s a platform that lets users get a firsthand experience with the Chatbot. You can test the Chatbot yourself, but there’s a small caveat: access requires signing in with your Google or Cohere credentials. The Coral Chatbot, which made its debut back in July, can now be seamlessly integrated into both internal and external applications using the newly released API.

The RAG feature

One of the standout features of Cohere’s Chat API is its use of Retrieval-Augmented Generation (RAG). This feature empowers developers to control the Chatbot’s information sources, enabling them to focus on enterprise data or expand their search to include the entire web. Cohere has honed Command specifically for RAG tasks, ensuring top-notch performance.

The incorporation of Retrieval-Augmented Generation (RAG) in Cohere’s Chat API is a significant leap forward for generative AI. RAG systems enhance the relevance and accuracy of AI responses by incorporating information from external data sources, a critical capability in today’s information-rich environment.

Beyond the RAG-enabled Chat API, Cohere’s platform offers three nifty components for third-party developers:

Document Mode: With this, developers can specify which documents their Cohere-powered Chatbot should reference when answering user queries. It’s a potent tool for customizing the Chatbot’s knowledge base.

Query-Generation Mode: This instructs the Chatbot to generate search queries based on the information provided by the user. It enhances the Chatbot’s ability to seek out relevant information.

Connector Mode: This allows developers to connect their Chatbot to the web or other information sources, broadening its capabilities and access to data.

There is no stopping

But Cohere isn’t stopping there. They have big plans to expand their modular ecosystem further, offering developers even more capabilities and options. This strategic move is in direct response to OpenAI’s recent reintroduction of web browsing capabilities for ChatGPT and their efforts to attract enterprise users through the ChatGPT for Enterprise subscription service tier.

Cohere’s success and growth are further underscored by their recent funding milestone. In June, the company secured an impressive $270 million in funding, solidifying its position in the AI landscape. Cohere’s journey from stealth mode in 2021, where it was launched by former Google Brain researchers, to its current standing has been remarkable. Their partnership with Google Cloud, utilising its AI and machine learning infrastructure, played a pivotal role in training their large language models.

Cohere’s impact extends beyond their platform. They’ve also established strategic partnerships with conversational AI provider LivePerson, providing large language models for LivePerson’s Conversational Cloud platform. This collaboration underscores Cohere’s commitment to advancing AI technology and making it accessible to businesses across various industries.