Things to consider about Oracle's Agentic Fusion
Beyond the Oracle Fusion Agentic Applications announcement that coincided with the Oracle AI Tour London event, there is the inevitable question of what it actually does.
By default, it seems the agentic AI functionality is switched off, but available to subscribers for an additional fee. That seems to be the business model the major IT providers have adopted. But Oracle is also trying hard to differentiate its offering.
Businesses are able to continue running their Oracle enterprise applications as is, but they have the option to add on agentic AI functionality. And to entice people to give it a go, it starts off for free.
In a LinkedIn post Sid Malik, co-lead for Oracle’s Centre of Excellence, Europe, Middle East and Africa wrote: “Every agent is free, regardless of who built it. Every Fusion customer gets 20,000 AI Units per month. If you’ve been sitting on the fence waiting for the commercial model to make sense, there’s no excuse left now.”
AI units represent a new consumption-based pricing metric, which Oracle uses to price use of Large Language Models (LLMs) running on the Oracle Cloud Infrastructure (OCI).
Within the order-to-cash business process, invoice processing is a domain-specific tasks Steve Miranda, executive vice president of Oracle applications development, gave as an example of agentic AI use during the London AI Tour.
NHS Shared Business Services (SBS) processes 7.1m invoices using Oracle Fusion Applications. The application handles up to £355bn in NHS transactions per year and there is an opportunity to AI-enable this.
Obviously there is a cost in having AI handle the processing. Every character in a machine-readable invoice will add a tiny amount to the Oracle AI unit count. And as the volume of invoices rises, so too does the cost. This is the basis of consumption-based pricing. Statistically speaking, a certain proportion will be flagged for additional investigation, possibly requiring a human in the end-to-end automated invoice processing task.
Miranda believes less people will be required. He said: “Nobody is in business to run ERP. The more we can save them on the ERP side, the more they invest in what they do.”
And this, to quote NHS SBS managing director, Erika Bannerman, means “the focus stays on providing the highest standard of care for patients.”
No one is likely to be against the NHS focusing more on patients, but in any organisation, there is going to be a big question over what happens to those back-office staff once agentic AI replaces a large part of their work. Success requires training the AI model to encapsulate knowledge from the very people whose jobs are in jeopardy. Within the NHS, they can’t simply retrain as clinicians. What are they supposed to do?