In our previous blog, we introduced the AI Assistant and its benefits for streamlining data access and improving decision-making. Now, we’ll dive under the hood to explore how to configure Generic Inquiries to unlock the AI Assistant’s full potential.

Configuring your AI Assistant: A Technical Deep Dive

To provide insightful answers, the AI Assistant needs access to well-designed Generic Inquiries (GIs). While some GIs are exposed by default, maximizing the Assistant’s capabilities requires careful configuration. Let’s explore the process of exposing new GIs to the AI Assistant.

Generic Inquiry Exposure: Step-by-Step

Here is an example of how the Assistant handles a query when it needs more data to work with.

Instead of a dead end, you get a roadmap. Pass these insights to your system admins to jumpstart new inquiries and expand the Assistant’s knowledge base. You can add to the data pool at any time, giving you the power to validate your data on the fly.

Configuring a GI for AI use involves several key steps:

1. Identify or create GIs with the relevant information and ensure they are published to the UI. As an example, I’ve created a GI displaying current stock on hand for each item and warehouse, using left joins to include all stock items regardless of inventory history.

2. Describe the data: Providing AI descriptions for the inquiry and its fields is optional, but highly recommended. This extra context helps the AI Assistant understand user questions and provide more accurate answers. Describe the GI as a whole on the AI Description tab or describe individual columns using the Description for AI field on the results grid.

3. Expose to AI: Toggle on this setting under the Interface Options tab to allow the Assistant to use the data from the inquiry.

4. Process AI Data: Process the inquiry on the new AI Assistant Data Processing screen to index the information. Newly exposed inquiries will initially show an AI Description Pending Update status. Whenever changes are made to exposed inquiries, they must be processed on this screen before the Assistant can use them. While a predefined schedule processes these updates weekly, administrators can customize the schedule or run the process manually.

Talking to Your Data

Once you’ve finished configuration, you can start asking your AI Assistant questions as if they’re another one of your colleagues. Use plain language, like: What is this month’s total income? or List the 10 customers with the oldest outstanding balance. The AI Assistant will utilize data from the data pool of exposed inquiries when forming responses.

Let’s ask our original question again, now that we’ve exposed the Current Stock On Hand GI.

The Assistant picks the best tool for the job, delivering answers as text summaries, tables, or charts. When specific records and IDs appear, these aren’t just names on a screen; they’re direct portals. One click will take you straight to the source, allowing you to validate the Assistant’s work with zero friction. In the example above, you can see this in action. The Assistant provided both a table and a chart, complete with live links to the stock item and warehouse records.

Optimizing AI Response

With the right data pool in place, the final step in your AI journey is mastering the art of the ask. AI can be an incredibly useful tool given the right data, but asking the right questions is just as important. For instance, shifting from row-level data to an aggregate view can uncover trends you might otherwise miss. Let’s look at how a quick rephrase changes the game:

Now FOODCOKD12 replaces MGFILAMENT as the 5th ranked item. This is where the Assistant proves its worth as a partner, not just a search bar. By asking for summed data, you uncover additional insights – like a new top-performing item.

Security and Access Control

You control both what the AI knows and who gets to ask the questions. Using the AI Assistant User role, you can roll out these capabilities gradually, ensuring the right people have access at the right time. This targeted approach also helps you manage your license efficiently by tracking and controlling monthly token limits. Use the License Monitoring Console to keep an eye on daily usage and ensure your team stays within budget.

Better Data, Smarter Decisions

Building a powerful AI Assistant isn’t a “one-and-done” task – it’s an evolving partnership between your team’s questions and your system’s data. By transforming roadblocks into roadmaps and continuously refining your Generic Inquiries, you do more than just answer questions. You create a culture of data-driven confidence where colleagues have the insights they need at their fingertips. Ready to supercharge your workflow? Start by auditing your most frequent queries today and see where a new GI can pave the way to your next big insight!