So far, in the enterprise, advanced chatbots, digital assistants, and coding assistants seem to be some of the areas of strength for GenAI applications.

ChatGPT's release in November 2022 sparked a GenAI gold rush as companies rushed to adopt the technology and demonstrate innovation.

Many of the AI application cases entrenched in today's enterprise use older, more mature forms of AI, such as machine learning, or don't take advantage of AI's "generative" capabilities to generate text, images, and other data. Traditional chatbots, product recommendation engines, and some other useful tools may rely solely on earlier forms of AI.

Kjell Carlsson, head of AI strategy at Domino Data Lab, an enterprise AI platform provider, said some industries, such as the biotech industry, are looking for ways to use GenAI, but so far many businesses experimenting with the technology have found a limited number of use cases. For many businesses, he said, GenAI's return on investment is difficult to achieve.

"The reality for many users is that they don't have enough information to develop an AI strategy that takes advantage of GenAI use cases, and they can't get enough value fast enough," he adds. They're pushing hard on several use cases, but they're also building a mix of traditional machine learning and 'predictive' AI use cases."

Many AI experts say the current use cases of GenAI are just the tip of the iceberg. As GenAI becomes more powerful and users become more creative in their experiments, more use cases will emerge.

However, several use cases of GenAI are already emerging. Here are a few of the most popular and promising.

Advanced chatbots

While simple chatbots using word and phrase recognition have been around for decades, new chatbots with GenAI capabilities can make conversations sound more natural while handling numerous customer requests.

IT analyst Forrester has named language GenAI and AI agents as two of its top 10 emerging technologies for 2024. For example, Bolt, a European ride-sharing and food delivery service, has deployed an intelligent chatbot to handle most customer complaints, resulting in significant cost savings.

Many companies experimenting with GenAI are concerned about hallucinations, but Carlsson notes that a few missteps aren't the end of the world when it comes to low-level customer complaints. "If we accidentally go in and deliver a meal when we should be denying someone a meal credit, the risk is very low." 'he said.

In another example, Deutsche Telekom has used GenAI to improve its Frag MagentaAI assistant, which the company expects will be able to handle 38 million customer interactions per year.

Digital assistants

Several large IT companies, including Microsoft and Google, have been touting GenAI digital assistants (or co-pilots), though CIOs may not be entirely convinced of the return on investment. These assistants can search for information in all corners of the enterprise, create documents and slide presentations, and summarize email chains and video conferences. GenAI digital Assistant can also generate supply chain documents, such as supplier quote requests.

Some video conferencing applications now generate transcripts and summaries, as do stand-alone tools such as Otter.ai. Apps like Grammarly can correct errors in grammar, spelling, and punctuation.

Nick Rioux, co-founder and CTO of Labviva, a provider of AI-assisted procurement solutions, said digital assistants can also be specialized for specific needs. For example, if a company regularly buys sensitive chemical or biological compounds, GenAI can add special handling instructions to the purchase order.

"The most promising use cases for enterprise GenAI are those that simplify human tasks through enhancements such as content generation, suggestions, and manual task automation," he says.

Coding assistants Coding assistants

One of the most frequent use cases in GenAI is the coding assistant. GenAI can write basic software code, allowing human programmers to focus on more complex tasks.

Julian LaNeve, CTO of data orchestration startup Astronomer, said these code co-pilots can also help programmers keep their focus on the code when they run into problems, rather than turning to search engines or other sources to find answers.

"They can just write code comments and let a large language model (LLM) do the code for them," he said, referring to a large language model. "This allows developers to stay in what we call a 'flow state' and a 'focus state' without the distraction of looking up examples."

Ai-generated technology is especially helpful for web development, added Natalie Lambert, founder and managing partner of AI Consulting firm GenEdge Consulting. By creating website code, GenAI can significantly reduce the time and cost required to update a website.

"By leveraging tools like ChatGPT, even users without deep technical expertise can develop and implement code directly on their websites," she said. "This democratizes the development process and allows cyber experts to realize their vision with the help of AI."

Many organizations that have implemented GenAI throughout the software development lifecycle are currently dealing with the technology's limitations and impact on their teams, while also taking stock of their own lessons learned.

Marketing support

Several AI experts and users pointed to marketing support as one of GenAI's strengths. Stefan Chekanov, co-founder and CEO of Brosix, a company that provides secure instant messaging tools, said GenAI can create personalized marketing materials, analyze customer data, and assist in content creation.

"In my experience, content creation and social media management have become much more efficient with GenAI's help," he said. "Less time spent on trivial scheduling, optimization, and editing means experts can focus on high-value tasks, saving follow-up costs."

Others say GenAI can perform market analysis based on product reviews and predict what problems customers might encounter before they even realize they have.

"It's critical for product companies to understand customer feedback," said Aswini Thota, director of AI and data science at USAA, a U.S. banking and insurance provider. "They need to know what customers like or dislike, what emerging trends are, what regional preferences are, and how customers will perceive the new product."

GenAI can extract customer insights from product reviews without the company having to commission surveys, he said. Before GenAI came along, data scientists built custom natural language processing (NLP) models for sentiment analysis and intent extraction, but GenAI goes one step further than those early efforts.

"GenAI allows us to build multiple prompts on the same dataset, and with the push of a button, businesses can extract sentiment, discussion topics, and intended uses." Thota added.

Drug discovery

Lars Nyman, chief marketing officer at CUDO Compute, an AI infrastructure platform, said GenAI is being used for drug discovery by modeling complex molecules and predicting their interactions "at a speed that makes it seem like traditional methods are stuck in the dial-up era." He says GenAI could significantly reduce the time to market for new drugs.

MSRcosmos, a global IT services provider, says GenAI can help pharmaceutical companies predict drug interactions, repurpose existing drugs, and create personalized therapies based on a patient's genetic makeup.

In early 2024, NVIDIA announced the launch of its AI-powered Clara computing platform for the healthcare industry, as well as BioNeMo, a GenAI platform for drug discovery.

Some biotech and pharmaceutical companies, including Johnson & Johnson, are promoting GenAI as the next big breakthrough in drug discovery.

Cybersecurity and fraud detection

Several cybersecurity companies are using GenAI to enhance tools to look for suspicious or unusual behavior in their customers' networks and computing infrastructure. Jim Kaskade, CEO of Conversica, a provider of conversational automation solutions, said AI systems can also be used for advanced fraud detection to predict fraudulent activity with a high degree of accuracy by analyzing transaction patterns and user behavior.

Palo Alto Networks, for example, offers its Cortex XSIAM security operations platform, which combines the company's expertise in machine learning (ML) models and data storage with Google's BigQuery enterprise data warehouse and Gemini AI models. The goal is to alert security analysts to threats in real time, while cybersecurity platforms constantly learn about new threats.

Business process enhancement

GenAI has found an area of strength in enterprise business process enhancement. In this area, enterprises are exploring the use of GenAI to increase efficiency for business-critical workflows that are often unique to their industry.

For example, some companies in the finance and insurance industries are using GenAI to assist underwriters in evaluating potential customers. Ryan Rosett, co-founder and co-CEO of Credibly, a lending platform for small businesses, said the platform uses GenAI along with machine learning to assess loan risk and speed up the lending process.

"At Credibly, we use GenAI to give our underwriters superpowers," he said. "As a fintech company, our success depends on fast and accurate risk assessment for business owners seeking funding."

According to a survey by EY, nearly all insurance companies have adopted or intend to adopt GenAI by the end of 2023. About 42% of insurance companies have already invested in GenAI, and about two-thirds expect revenue growth of more than 10% by using GenAI.

In the legal space, legal information services giant LexisNexis is embracing GenAI to tackle what Jeff Reihl, the company's executive vice president and chief technology officer, sees as a disruptive threat to the industry.

"We are all mobilized," Reihl told reporters. "We made a major transformation because this is a game-changer in terms of the ability to interact, the comprehensiveness of the answers, and the ability to generate data, which is astounding."

LexisNexis has since released its own GenAI solution, Lexis+ AI, which provides linked legal citations to ensure lawyers have access to accurate, up-to-date legal precedents.

Predictive analysis

While GenAI models have traditionally excelled at retrieving and summarizing information, businesses are now leveraging the technology for predictive analytics.

For example, Erez Agmoni, general partner at Interwoven Ventures, an AI and robotics venture capital fund, said some companies use GenAI to predict shipping schedules.

Traditional AI-powered predictive analytics is not new, but GenAI excelled at this task because of its ability to work with unstructured data without the need for predefined algorithms, Agmoni said. He previously served as head of AI and robotics deployment at shipping company Maersk.

Shipping schedules can be difficult to predict, as multiple factors affect the time it takes to reach the final destination, he said. A simple algorithm that simply looks at historical data is not enough to provide an accurate delivery date.

Shippers need multiple systems to share past and current data, including information on the performance of multiple routes, weather, labor performance, and financial market conditions. "Being able to solve a problem like the one in this example can generate billions of dollars for the participants, so there is a great desire to find a solution." "Agmoni said.

Extract unstructured data from multiple sources

State-of-the-art Large language models (LLMS) can help businesses extend their AI strategies to previously unstructured data in text, video, and voice messages. For example, Sriram Nagaswamy, executive vice president of supply chain visibility platform provider FourKites, said some businesses are using GenAI to extract data from video surveillance systems.

"One of the most exciting breakthroughs we see in GenAI is its ability to extract unstructured data from a variety of applications that were previously too cumbersome or time consuming to utilize, which has the potential to revolutionize the market." 'he said.

For example, many shipping ports have cameras that refresh periodically.

"If we start capturing those frames, we can just look at the license plate number or the container number and know exactly which trucks are coming in and out," he said. "Most ports have these cameras 24/7, so the democratization of data and a more natural approach to data collection will be the low-hanging fruit, unlocking valuable shipping and tracking insights."

Nagaswamy sees a clear adoption trend toward large language models that can handle multimodal inputs and outputs, although accuracy may not immediately reach 99%.

"We will see significant progress in democratizing this capability in the coming year as people become accustomed to simply speaking to or sending images or videos to large language models." "Nagaswamy said.