From Chat to Decision: Building an AI Assistant with SAS Intelligent Decisioning and Azure

[#1]

The integration of conversational AI with decision-making systems is revolutionizing how businesses interact with data and users. This post explores the smooth integration of SAS Intelligent Decisioning with Azure OpenAI, showcasing how these technologies can transform AI assistants into powerful tools for real-time, data-driven decision-making.

Demonstrating the Concept: AI-Powered Risk Assessment

Imagine interacting with an AI assistant that not only answers your queries but also performs complex tasks, such as calculating your financial risk rating. Through a conversational interface powered by Azure OpenAI, the assistant collects inputs like age, income, employment type, debt-to-income ratio, and credit score. These inputs are then passed to a trusted internal workflow. The decision logic originates in SAS Intelligent Decisioning, which evaluates the data and returns an accurate risk rating (e.g., low, medium, or high).

This approach combines the ease of conversational AI with the precision of decision logic, providing an intuitive and reliable user experience. Whether it’s for banking, insurance, or other industries, this solution makes complex decision-making accessible and efficient.

Explaining the Concept: How It Works

The integration relies on a streamlined workflow involving key components:

1. Azure OpenAI Assistant

The assistant in Azure AI Foundry acts as the user interface, collecting inputs and triggering backend workflows. Using the Azure OpenAI Assistants API, the assistant can call external functions, such as an Azure Logic App, to process user requests.

Select any image to see a larger version.
Mobile users: To view the images, select the "Full" version at the bottom of the page.

2. Azure Logic App Workflow

The Logic App serves as the bridge between the AI assistant and the decision logic. It:

• Formats user inputs into a JSON payload.
• Sends the payload to an API endpoint hosting the decision logic.
• Processes the decision logic’s response and sends the result back to the assistant.

3. SAS Intelligent Decisioning

The decision logic resides in SAS Intelligent Decisioning, which evaluates the inputs and produces a result. For example, in a risk rating scenario, it calculates whether the risk is "low," "medium," or "high" based on predefined rules.

4. SAS Container Runtime (SCR) Containers: I decided to deploy the SAS decision logic to an Azure container and invoke it via the Logic App.

Deployment Options

The integration supports several deployment options:

SAS Viya

• SAS Container Runtime (SCR) Containers.
• SAS Decisioning Connector: Use the native connector in Logic Apps to call decisions published to SAS Viya Microanalytic Service. (The connector is currently in preview, and for many reasons, I did not manage to connect directly to a SAS Viya instance deployed in Azure. However, it's a solid option.)

Azure

• Instead of using the Azure AI Foundry interface, the approach can be packaged and deployed as a Python application.
• Further, the Python application can be containerized and deployed in Azure.

End-to-End Workflow: A Risk Rating Calculation Example

Here’s a step-by-step breakdown of the process:

1. User Interaction: The AI assistant offers to calculate a user’s risk rating. The assistant collects inputs such as age, income, and debt-to-income ratio.
2. Data Transformation: The assistant converts these inputs into a JSON payload.
3. API Invocation: The Azure Logic App sends the JSON payload to the SAS Intelligent Decisioning API.
4. Decision Scoring: The decision logic processes the inputs and returns a risk rating.
5. Response Delivery: The Logic App formats the result, and the assistant communicates it back to the user.

Conclusions: Transforming AI Assistants into Decision Engines

The integration of SAS Intelligent Decisioning with Azure OpenAI Assistants represents a significant leap in AI-driven workflows and an opportunity for SAS. By combining conversational AI with robust decision-making systems, businesses can deliver real-time, accurate insights while maintaining a uniform user experience.

Why This Matters:

• Enhanced Efficiency: Automates decision-making processes.
• Improved Accuracy: Leverages trusted decision logic for reliable outcomes.
• User-Friendly: Simplifies complex workflows through a conversational interface.

By bridging conversational AI with decision logic, this integration delivers smarter, faster, and more intuitive workflows. Whether you’re automating risk assessments or other critical decisions, the combination of SAS and Azure technologies is a game-changer for AI-powered solutions.

Additional Resources

I was inspired by the following resources to power an Azure AI Assistant with a SAS decision:

• Call Azure Logic Apps as Functions Using Azure OpenAI Assistants.
• Azure OpenAI Assistants API with Logic Apps.
• SAS Decisioning (Preview).

Thank you for your time reading this post. If you liked the post, give it a thumbs up! Please comment and tell us what you think about having conversations with your data. If you wish to get more information, please write me an email.

Find more articles from SAS Global Enablement and Learning here.