Oracle AI agents to help automate dozens of processes as businesses seek productivity gains

As companies ramp up artificial intelligence spending and hunt for investment returns, they’re augmenting single-shot, question-and-answer interactions with AI software “agents” that help complete entire tasks.

The self-directed software leverages generative AI models—trained on vast amounts of data to make associations among concepts and learn new tasks—to find job candidates, analyze talent, research sales calls, and onboard suppliers. Oracle on September 11 said it would ship more than 50 such AI agents as part of its Fusion cloud business applications suite in areas spanning finance, HR, supply chain management, quality control, sales, and customer service.

Instead of relying on keyword triggers or pre-coded business rules and workflows, agents powered by large language models can carry out multi-step processes that span software tools, adapt to novel scenarios, understand users’ roles in an organization, and respond to natural-language prompts instead of code. That helps make them more flexible and less error-prone than previous systems that were more rigid and less aware of incoming information. AI agents can also pull data from businesses’ own documents, helping workflows stay relevant.

In one example, Oracle’s document agent could help a sales exec with the following functions: snap a photo on his phone of an overseas vendors’ price quote, extract relevant information and translate it from Japanese, then create a purchase request. Later on, the software can assist to automatically process the vendors’ invoice for review by a payment manager.

“It blows away anything we’ve ever been able to deliver before,” said Oracle Executive Vice President Steve Miranda during a speech at the company’s CloudWorld conference for customers and industry partners in Las Vegas. “It’s a tremendous change in terms of input and output from our applications across the board.”

AI’s pace of improvement—with agents just the latest example—may incent more companies to move their applications and data to software-as-a-service subscriptions and online platforms.

“We’re getting more customers to the cloud, but there are still laggards out there who haven’t moved,” says Oracle Group Vice President for Applications Development and Strategy Miranda Nash. “If there’s any incentive to getting to quarterly updates with new innovations, it’s got to be AI. How can you stay on the sidelines?”

Oracle plans to make Fusion cloud applications’ AI agents available within the next 12 months at no additional cost. The agents couple large language models that Oracle hosts on its public cloud with a technique called retrieval augmented generation (RAG) that extracts information from a company’s own documents and feeds the LLM with a continual history of what a computer user has done so it has more context for its answers. “It’s not easy to get consistent, reliable results out of these LLMs,” says Nash. “That’s not something customers should take lightly.”

In search of AI ROI

AI that’s embedded right into business processes, aware of who’s using it and able to help accurately automate time-consuming jobs could be a boon at a time when businesses and investors are looking for proof that massive investments in AI will pay off. Technology companies and utilities are set to spend more than $1 trillion on AI infrastructure in the coming years, including data centers and chips, according to Goldman Sachs. Downstream productivity gains so far haven’t matched the investments.

Businesses expect their spending on AI to increase by 30% in 2024, compared with a 3.3% average increase in overall IT budgets, according to a Boston Consulting Group survey of 330 North American and European IT buyers published in July. More than 30% of the respondents BCG called “high-maturity” AI implementers expect no more than a 10% return on their investments over the next three years—with even more pessimistic prognoses at companies that have adopted less of the technology.

Agents that help guide workers through tasks, search the web, compile staff feedback, and retain information throughout processes could help companies get more value from their data than the one-shot queries users often send to LLMs. They could also provide a sophisticated update to business software automation tools and productivity assistants that tech companies have long built into applications, PC operating systems, and phones. “Without foundation models, these capabilities would require extensive manual efforts to integrate systems, or tedious manual efforts to collate outputs from different software systems,” consultancy McKinsey & Co. said in a July report.

As one example of productivity gains, generative AI agents could help reduce the multi-week process of a bank creating a credit risk memo by 20% to 60%, according to McKinsey. AI agent software has “traditionally been difficult to implement, requiring laborious, rule-based programming or highly specific training of machine-learning models. GenAI changes that.”

Breadth of data

Technology companies are infusing business software with self-directed agents. Microsoft has embedded agents into its productivity applications, including a co-pilot for its Teams meeting software that can manage agendas, track follow-up actions, and notify workers when colleagues need their input. Google has demonstrated a smartphone assistant that can answer questions about objects and drawings viewed through the camera.

Oracle is differentiating its offering by taking advantage of the breadth of enterprise data stored in its Fusion applications for financial management, HR, marketing, sales, customer service, and supply chain management, according to Nash. Its applications also run on Oracle Cloud Infrastructure (OCI), which means customers can source agent-enhanced business apps, cloud infrastructure, and databases from one vendor. The company is hosting an LLM from partner Cohere for Fusion agents and offering customers online services to help users employ appropriate language and get results consistent with what a system is trying to accomplish.

For example, Fusion agents can help finance departments create and track forecasts, assist writing job postings and offer letters for HR departments, or help summarize account information and upselling opportunities for sales reps. In the future, they could let HR conduct talent reviews and set up employee meetings or procure supplier quotes for necessary equipment repairs. Businesses that want to build agents themselves can use Oracle’s OCI GenAI Agents technology.

View more Oracle Connect articles

Future Product Disclaimer
The preceding is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions. The development, release, timing, and pricing of any features or functionality described for Oracle’s products may change and remains at the sole discretion of Oracle Corporation.

Forward-Looking Statements Disclaimer
Statements in this article relating to Oracle’s future plans, expectations, beliefs, and intentions are “forward-looking statements” and are subject to material risks and uncertainties. Many factors could affect Oracle’s current expectations and actual results, and could cause actual results to differ materially. A discussion of such factors and other risks that affect Oracle’s business is contained in Oracle’s Securities and Exchange Commission (SEC) filings, including Oracle’s most recent reports on Form 10-K and Form 10-Q under the heading “Risk Factors.” These filings are available on the SEC’s website or on Oracle’s website at http://www.oracle.com/investor. All information in this article is current as of September 11, 2024 and Oracle undertakes no duty to update any statement in light of new information or future events