A company needs to extract ingredient names from recipe records stored as text files in an Amazon S3 bucket. A web application will query a DynamoDB table using the ingredient names to determine a nutrition score. The company can tolerate non-food records and errors and has no employees with ML expertise. Which solution is the MOST cost-effective?

The correct solution leverages S3 Event Notifications to trigger a Lambda function when new recipe files are uploaded. This Lambda function then uses Amazon Comprehend, a natural language processing (NLP) service, to extract ingredient names from the text. Comprehend is ideal here because it's a fully managed, pre-trained service requiring no ML expertise, making it cost-effective and suitable for the company's constraints. The extracted data is then stored in DynamoDB for efficient querying. Incorrect options: Amazon Forecast is a time-series forecasting service, not suitable for extracting entities like ingredient names from text. Amazon Polly converts text to speech, which is irrelevant for extracting ingredient names and would require manual effort from employees, making it highly inefficient and not cost-effective. Amazon SageMaker is a machine learning service for building, training, and deploying custom models. While powerful, it requires ML expertise and significant development effort, making it more expensive and complex than necessary for this use case, especially given the company's lack of ML expertise.