Back to the future with EBS and AI natural language queries

For the past three years at Inoapps, I’ve reconnected and been deeply engaged with Oracle E-Business Suite (EBS), reconnecting with senior leaders like Cliff Godwin and Nadia Bendjedou.

My first realization was that many believed EBS had a limited future—a myth I addressed in my first blog. I’ve since been pleased to see Oracle clarify its support policy from ‘at least 10 years’ to a ‘rolling 10 years,’ reinforcing its long-term commitment to the product.

However, EBS update and strategy presentations focused on new 12.2.X functionality or additions to Enterprise Command Centers (ECC), alongside technical details on staying supported. Yet most customers aren’t prioritizing these updates. Their EBS systems are stable, customized, and meet their needs through integrated systems or third-party solutions.

At the recent UK Oracle User Group annual conference, more than half the vendors showcased EBS third-party tools, highlighting the rich ecosystem supporting EBS. When discussing ECC adoption, companies often mention they already rely on other tools, with existing reports or queries fulfilling their operational needs.

That said, ECC offers great functionality at no extra license cost, and customers can benefit. Exploring its potential might help reduce customizations, making future upgrades smoother. Still, for many, staying supported remains the priority over adopting new features.

I do think you should be looking more at the developments. One reason is it may mean being able to retire some customizations, which makes your next move or upgrade much easier.

And this year, the Oracle EBS update messages are changing, and I see excitement in the audience as EBS gets ready for Generative AI.

At Oracle CloudWorld this year, Inoapps hosted a roundtable for EBS customers with Nadia Bendjedou, and the question of AI came up.

Independently, Inoapps was already using the Select AI services provided by Oracle in OCI against many data sources. One of those is our Inoapps Archiving Solution, and we’ve found that the ability of those AI services to create queries against that data is incredibly powerful. We just speak questions like ‘tell me about my suppliers’ or ‘tell me about my spend,’ and the AI generates queries that return the information.

The exciting news is that this is now available for EBS itself. At our roundtable, Nadia explained—using coffee cups and serviettes—the architecture they’re using to approach this. To make the EBS data understandable to the AI services, they are introducing a new schema, XX_NLQ, in the database based on the views that are generated for the ECC solution, which can then be interrogated by the AI services.

This XX_NLQ schema, which is read only, cannot be used as a back door to the data and will be secured with the same permissions as the application itself.

The AI services need to be run from an independent front end that can use the new 23ai autonomous database required for the service. The easiest way to do this is with APEX. APEX integrations have long been a favorite with EBS customers. Even slicker now is the ability to embed an APEX screen into EBS itself, like a simple i-frame. This allows a single page to be the front end to the Generative AI you’re going to consume. This ability to embed APEX pages into your EBS gives those already using APEX for other functionality an even slicker user experience.

Nadia explained the theory to this at our roundtable, and at the recent Applications Unlimited Roadshow, she went into more detail and gave examples. Oracle released this at the start of December, and it is available for all releases from 12.2.7 onwards.

Here at Inoapps, we are ready and excited to be working on this with our customers. If you’d like to explore the possibilities, the AI experts at Inoapps are here to talk.

In my next blog, I'll be talking about how Generative AI in EBS differs from AI in Oracle Fusion Cloud.