Skip to main content

Batches

Batches in BotDojo provide a powerful way to run a Flow in bulk, allowing you to process large amounts of data efficiently. With batches, you can automate repetitive tasks and scale your AI applications to handle real-world scenarios.

Here are a few examples of how batches can be used:

  1. Evaluations: Combining batches with evaluations provides the confidence needed to ensure the quality of your Flow. By constructing a dataset of test cases tailored to evaluate your Flow under various conditions, and running these tests using different models and prompts, you can quickly enhance your Flow's performance. This process allows for the comparison of results across different configurations, helping you pinpoint the most effective setup for your specific requirements.

  2. Processing Product Descriptions: Suppose you have 5,000 product descriptions and want to research the selling prices of those products on eBay. You can create a Flow that takes a product description as input, uses an AI agent to search the web for relevant information on eBay, and retrieves the selling prices. By running this Flow as a batch, you can process all 5,000 product descriptions in one go.

  3. Analyzing Customer Tickets: Let's say you have a dataset of open customer tickets and want to determine how likely a recent feature release is related to a customer issue. You can build a Flow that analyzes the content of each ticket and compares it with the release notes to assess the relevance. Running this Flow as a batch allows you to process all the customer tickets efficiently and gain insights into the potential impact of the release note. You can then schedule this batch to run when a new release is published.