October 2022 Summaries

Modern businesses often face the challenge of efficiently extracting and categorizing data from the large volume of invoices they handle as part of their accounting processes. This task is complicated by the varied formats and layouts of invoices, making manual data extraction time-consuming and prone to errors. To address this, the Cohere Platform offers a solution by utilizing pre-trained large language models for natural language processing (NLP) to automate the extraction of specific information from invoices, such as customer names, addresses, and email addresses. By leveraging these models, businesses can bypass the need for extensive training and expertise in machine learning or AI, as the Cohere models can quickly process and extract relevant data with minimal human intervention. This approach not only improves efficiency but also reduces inaccuracies and labor costs associated with manual data extraction, demonstrating the potential of integrating advanced AI tools into everyday business tasks.

Cohere has introduced a revamped pricing strategy for its platform, aiming to offer more flexibility and value to developers and businesses by providing a new free developer tier and simplified pricing plans. This new pricing model includes a freemium tier replacing the previous $75 credit program, allowing developers to experience Cohere's API with trial and production API keys, the latter offering significantly higher rate limits. The pricing changes also involve standardizing endpoint costs across model sizes and reducing embedding prices to encourage broader application development. Additionally, Cohere is refining its product terminology by renaming "finetuned models" to "custom models" and "baseline models" to "default models," alongside implementing user-friendly features such as QuickStart guides, accuracy metrics, and training logs to enhance the developer experience. These updates are part of Cohere's commitment to simplifying its platform and enabling developers to build innovative applications with ease.