A company is deploying a generative AI model on Amazon Bedrock and needs to reduce the cost of usage while using prompt examples of up to 10 sample tasks as part of each input. Which approach would be the most effective in minimizing the costs associated with model usage?