Introduction

Generative AI (GenAI) models are enabled for use in other products and workflows through APIs and services. AI costs can increase significantly due to inefficient API usage, lengthy and complex model training runs, unnecessary fine-tuning, and retaining large volumes of unused data.

Integrating DoiT with AI providers enables you to consolidate your AI cost and usage data, providing granular visibility into your AI spend and attributing costs directly to projects and teams.

DoiT enables you to integrate with AI providers including:

• Amazon Bedrock
• Google Cloud Vertex AI
• Anthropic
• OpenAI

Note

If you have onboarded your AWS or Google Cloud billing accounts and you use Amazon Bedrock or Vertex AI, we utilize your CUR and billing data to break down the relevant SKUs and show detailed costs and usage information for your generative AI workloads in the GenAI lens.

Benefits

• Unified cost structure: DoiT enables you to integrate AI cost and usage data alongside data from other cloud providers, giving you a comprehensive view of your cloud expenditures and facilitating better financial decision-making.

• Enhanced visibility: Once you've imported your AI data into the DoiT platform, you can start using Cloud Analytics and other DoiT features such as Budgets, Allocations, Forecasting and Anomaly detection to analyze and monitor your AI cost and usage.

Key features

• GenAI Intelligence. For each AI provider that you connect to DoiT, the preset GenAI Intelligence dashboard helps you identify and track your AI cost and usage in a single lens, using common tagging and clear dimensions. This allows you to track AI cost and usage across your projects and organizations. Use the Lens dashboard to decide how and when to optimize costs, monitor performance, and allocate cost the correct teams.

• Model usage. DoiT enables you to see your generative AI model costs. This is useful as it helps you understand any unexpected expenses, inefficient resource allocation, and potentially unsustainable AI deployments.

• API calls. Regularly reviewing your API usage and cost is important when integrating AI into your applications. AI costs can increase significantly due to inefficient API usage.

• Fine tuning. Fine tuning existing models requires significant computational resources to adjust the model's weights based on your training data, causing costs to quickly escalate. Monitoring fine-tuning costs ensures that you are getting the desired performance improvements in a cost-effective manner.