Generative AI (GenAI) is generating a lot of excitement globally. It’s not only transforming industries but also sparking creativity and innovation in ways previously thought impossible. In this blog post, we explore GenAI, how it works, and some of the enterprise use cases where it can be applied.

What is generative AI?

Generative AI is a form of artificial Intelligence that can generate content, including images, video, text, code, or even creative and complex designs based on the user’s request or prompt. Traditionally, AI has been used for tasks like classifying or grouping data based on one or more types, forecasting, finding anomalies and patterns within data, and making decision based on existing data. However, generative AI can perform more sophisticated tasks like summarizing a conversation, generating video content, and generating code to build new applications.

How does generative AI work?

GenAI has been built using a range of advanced machine learning (ML) models within the realm of deep learning. At its core, it relies on neural networks: Computational models inspired by the human brain.

GenAI models are trained on large corpus or data during which it learns about patterns within the data and relationships that exist within the data. You can adjust this training process to make the model more realistic using the parameters of the model. After the model is trained, it can produce content in the form of text, images, or video.

GenAI models come in many forms, depending on the data they work with and other factors, such as generative adversarial networks (GANs), variational autoencoders (VAEs), and transformers. To learn more, refer to the links at the end of this post. A few important concepts are important for better understanding of how GenAI models work. For more on these concepts, see Concepts for Generative AI.

In this blog post, we focus on large language model (LLM) use cases. The most widely used and the most advanced language models available today are based on transformer models.

A language model is a generative AI model that can process, manipulate, and generate natural language. A language model also acts as a probabilistic model of text, such as a model that predicts the probability of a sequence of characters or words in a natural language.

Generative AI use cases in enterprise applications

AI assistants are solutions that utilize AI and GenAI technology to add significant value and revenue to the business by providing insights and simplifying complex activities that usually require higher skills, time and effort.

The following table shows use cases and benefits for the financial domain:

AI product

Use case

Business benefit

Financials assistant

Proactively research out of balance accounts, correct subledgers, gather information

Faster financials closing of books

Receivables assistant

Accounts receivable: Identify uninvoiced sales orders, create invoices and send to customers, follow up-based payment schedule

Better cash flow, reduce collection costs

Payables assistant

Accounts payable: Identify overdue supplier invoices, automate approval of invoices and payments to suppliers per payment terms

Better tracking of days payables outstanding (DPO)

Financials reporting assistant

Identify negative net income trends and automate root cause analysis: Create summary and detailed reports with alerts to stakeholders

Automate and reduce time and cost of creating financial reports with detailed explanations and notes

Financials forecasting assistant

Financial forecast building and analysis: Identifying anomalies and underlying causes

Automate Financial forecast building and reporting

Pricing assistant

Formulate pricing strategy by pulling market pricing, internal costs, competitor analysis and arriving at the best pricing along with qualifiers and modifiers for volume, customer category, and other environmental forces

Dynamic pricing that helps ensure maximum revenue and margin for sustainability and investment, customer retention

 

The following table shows use cases and benefits for order management:

AI product

Use case

Business benefit

Customer Service Representative (CSR) assistant

CSRs get assistance from intelligent digital agents to look up customer history, analyze, lookup parts, push literature and documents, research issues, configure, price, and create quotes for customers

CSR productivity and happier customers

Inside sales assistant

Recommend and personalize offers, upsell and cross-sell based on customer categories, preferences, lifestyle, and past purchase history

Increased revenue through white glove selling

Customer assistant

Voice enabled interactions to reduce dependence on humans, classify intent, detect sentiment and route as appropriate for intervention

Reduce costs in customer service

Field service assistant

Enhance field service engineer efficiency through manuals, virtual and augmented reality (VR and AR), Q&A, digital assistance, automatically create tickets, Return Merchandise Authorization (RMA)

Improve Service technician productivity through assistance and on-the-job training

Sales query assistant

Assistance to retrieve sales and product information using natural language

Eliminate the need to build reports and perform manual analytics

Order processing assistant

Assist to create, cancel and modify orders, Available To Promise (ATP ) lookup and scheduling, responding to supply chain changes

Better customer satisfaction and proactive order status communication

Returns processing assistant

Help CSRs to decide whether to authorize return from customer or send parts from parts guide

Increased customer satisfaction

 

The following table shows use cases and benefits for revenue management:

AI product

Use case

Business benefit

Supply chain monitoring assistant

Monitor supply chain, looking for potential disruptions, alert, conduct simulations, and recommend actions

Zero-latency supply chain

Backlog assistant

Monitor order backlog, alert when supply is available, experiment with options and recommend best course of action, such as expedite.

Increased customer satisfaction, higher revenue and better net income

Sales assistant

Assist Sales team to fulfill customer products, upsell, and cross-sell based on customer category, history, and other parameters including margins

Higher revenue and profits

Forecasting assistant

Orchestrate consensus forecast creation and perform forecast consumption, production and shipment relief to help ensure that true demand is identified for financial and operation planners

Better forecast accuracy

Supply base Assistant

Look at material requirements in the Bill of Materials (BOM) and scour the market looking at credit worthiness, reviews, and other parameters to help identify suppliers who have capacity and lower price to help buyers drive down cost and refresh supply base

Higher margins, satisfied customers, reduce supply risk

 

OCI Generative AI

Oracle Cloud Infrastructure (OCI) offers the following AI tools and services:

  • Embedded GenAI in business applications: Oracle embeds classic and generative AI into its applications, allowing customers to access AI content and results in their application environment itself without having to deploy any GenAI model.
  • OCI Generative AI service: OCI offers a managed generative AI service where customers can deploy Cohere and Meta models and fine-tune the models as needed. OCI also provides APIs to access the service allowing customers to integrate them into wider range of use cases.
  • OCI Generative AI Agents service: OCI offers a managed service where customers can deploy retrieval-augmented generation (RAG)-based LLMs against their enterprise data to ground the responses within the knowledge bases specific to the enterprise application.
  • OCI Data Science: Build, train, deploy, and manage custom LLMs with open source libraries, such as Hugging Face’s Transformers or PyTorch, and generative models from Meta or Mistral AI.
  • AI Vector Search in Oracle Database 23ai: AI Vector Search with Oracle Database 23ai provides semantic search capabilities using AI vectors. Searches on business and semantic data are more precise because a single converged database mangaes both types of data .
  • MySQL Heatwave Vector Store: MySQL HeatWave is a fully managed database service that provides an online transaction processing (OLTP) database, a real-time in-memory data warehouse, in-database automated machine learning, lakehouse, and private preview GenAI capabilities in a single cloud database service, without the complexity, latency, and cost of extract, transform, load (ETL) duplication. It provides the best performance and price performance in the industry for analytics processing in both data warehouse and lakehouse environments.
  • Autonomous Database Select AI: Select AI is a feature within Oracle Autonomous Database that enables applications and analytics to use LLMs to understand users’ natural language questions and generate Oracle SQL to query data.
  • OCI AI Infrastructure:  Run the most demanding AI workloads faster, including generative AI, computer vision, and predictive analytics, anywhere in our distributed cloud. Use OCI Supercluster to scale up to 32,768 GPUs today and 65,536 GPUs in the future.

Running generative AI on OCI provides you with the following benefits:

  • Choice of models: Access proprietary or open source generative LLMs as suited for your needs with high performance and low cost.
  • Embedded AI across the full stack: Use generative AI as managed services for your custom applications, with the data already in our data platform or integrated within Oracle Cloud Applications.
  • Enterprise-grade security and privacy: Prioritize data management, security, and governance with Oracle in the cloud and on-premises—essential for enterprises.
  • Predictable performance and pricing: Gain trusted performance and transparent pricing with our OCI Supercluster technology for AI.

Conclusion

To learn more about Oracle Cloud Infrastructure generative AI, see the following resources: