Many organizations launched dozens of AI pilot projects in the last two and half years, with mixed success and disappointing ROI. But all that experimentation has borne some fruit, because several standout use cases for generative AI have emerged.

Many AI experts say the current use cases are just the tip of the iceberg, but here are the ones that have emerged so far.

Advanced chatbots and customer service

While simple chatbots using word and phrase recognition have been around for decades,  with gen AI capabilities can make conversations sound more natural while dealing with many customer requests.

European ridesharing and delivery service Bolt, for example, has  to deal with most customer complaints, creating a huge cost savings.

In another example, Deutsche Telekom has  to improve its Frag Magenta AI assistant, and the company anticipates the chat assistant will be able to handle 38 million customer interactions each year.

Verizon has also bet big on gen AI for customer service. The telecom provider’s call centers receive hundreds of millions of calls a year, and another two billion customer interactions take place over digital channels.

These numbers give Verizon a huge opportunity for gen AI to have a business impact, and the company is already seeing benefits, says , Verizon’s chief customer experience officer.

Verizon began working on developing gen AI use cases a couple of years ago, starting with Google Bard, and moving up to Gemini 1 and then Gemini 1.5, using retrieval augmented generation (RAG) to provide the model used with relevant information.

The gen AI is used to quickly summarize complex documents, listen to conversations and automatically pull up relevant information, answer questions, and provide other functionality. Take for example the average 18-minute conversation a customer might have with a Verizon representative.

“We can shave off about a minute on every call,” says Higgins. “When we think about the number of calls coming in, that has a material impact.”

The amount of time a customer has to wait before they can talk to someone has come down as well, he says.

Today, all of Verizon’s customer service representatives, more than 40,000 employees, use the gen AI tool. The fine-tuning and contextualization that Verizon adds to the base model means that accuracy is currently over 90%, Higgins says.

Digital assistants and productivity

Several large IT companies, including Microsoft and Google, have been touting , or copilots, even though some CIOs were initially suspect of the potential ROI.

These assistants can search the dark corners of the organization for information, create documents and slide presentations, and summarize email chains and videoconferences. Copilot AIs can also generate supply-chain documents, such as requests for quotes from suppliers.

Some videoconferencing applications now generate transcriptions and summaries, as do standalone tools such as . Apps such as Grammarly correct mistakes in grammar, spelling, and punctuation.

Productivity is one of the main areas of gen AI deployment at PGIM, a $1.4 trillion asset management firm, says , CITO there. AI-driven productivity tools, available through Office 365, enable employees to perform tasks such as summarize emails, or help with PowerPoint and Excel documents.

Not all tools are available to all employees, but those who are able to use gen AI take advantage of it, he says. “Once you get Copilot for Office 365, you go through training, and that’s driven up our utilization to around 93%.”

Another organization using Microsoft Copilot for productivity is Oral Roberts University in Tulsa, Okla.

“It’s accelerating the learning process, improving research, and helping students with assessments,” says , the university’s VP for innovation and technology.

For example, Copilot can create a draft outline in seconds when a faculty member wants to teach a new section of a course.

“I’ve not met many professors who don’t feel overworked and exhausted,” he says. “To be able to have the opportunity to do research and grade faster, and to assess students faster, it’s all perfect for them.”

The university is also using AI to process applicant transcripts, the time it takes to respond to applicants has fallen from weeks to hours, and enrollment has grown by nearly 11%, he says.

Coding assistants and software development

One of the use cases for gen AI that pops up the most frequently is the . Gen AI can write basic software code, allowing human programmers to focus on more complicated tasks.

These code copilots can also help programmers keep their focus on code when they run into a problem, instead of turning to a search engine or other resources to find answers, says , CTO at data orchestration startup Astronomer.

“They can instead write a code comment and let an LLM complete their code for them,” he says, referencing large language models. “This keeps developers in what we refer to as the ‘flow state’ and ‘in the zone’ instead of breaking focus to search for examples.”

Gen AI is particularly helpful for web development, adds , founder and managing partner at GenEdge Consulting, an AI consulting firm. Gen AI, by creating website code, can significantly reduce the time and cost needed to update websites.

“By leveraging tools like ChatGPT, even users without deep technical expertise can develop and implement code directly on their websites,” she says. “This democratizes the development process, allowing web specialists to actualize their vision with AI assistance.”

Many enterprises that have been implementing gen AI across the software development lifecycle are currently working through the technology’s limits and team impacts, as well as their own lessons learned.

PGIM also uses gen AI for code generation, specifically using Github Copilot. About 1,000 developers in the company are using it, says CITO Baker.

“We see about 60% of our developers using it on a day-to-day basis,” he says. “And about 70% of the code that’s recommended by Copilot we actually adopt. So a pretty high adoption rate for AI code generation.”

Marketing and sales support

Several AI experts and users point to marketing support as one of gen AI’s sweet spots. Gen AI can create personalized marketing materials, analyze customer data, and aid with content creation, says , co-founder and CEO of Brosix, provider of a secure instant messenger tool.

“In my experience, content creation and social media management are much more efficient with the help of gen AI,” he says. “Less time spent on menial scheduling, optimization, and editing means experts get to focus on high-value tasks, which equals cost savings down the line.”

Gen AI can conduct market analysis based on product reviews, and it can predict customer problems even before they recognize the issues, others say.

“For product companies, understanding customer feedback is crucial,” says , director of AI and data science at banking and insurance provider USAA. “They need to know what customers like or dislike, emerging trends, regional preferences, and how customers will value new products.”

Gen AI can extract customer insights from product reviews instead of companies needing to commission surveys, he says. Before gen AI, data scientists built custom natural language processing (NLP) models for sentiment analysis and intent extraction, but gen AI has added to those earlier efforts.

“Gen AI allows us to craft multiple prompts on the same data set, and with a push of a button, organizations can extract sentiment, topics of discussion, and intended usage,” Thota adds.

PGIM also uses AI to help its sales staff interact with documents in a more user-friendly way.

“Our salespeople go out to meet with financial advisors,” Baker says. “We have a ton of documents we can talk about. With AI, they’re able to get the right information at the right time for business conversations, which was somewhat hard and laborious to get in the past.”

And that’s just the start of it. They also use AI to help with website content, speech writing, and client communications. “Marketing communications is a great area for AI,” he says.

Drug discovery

Gen AI is being used in  by modeling complex molecules and predicting their interactions “at speeds that would make traditional methods look like they’re stuck in dial-up internet days,” says , CMO of CUDO Compute, an AI infrastructure platform. Gen AI can significantly cut down the time it takes to bring new drugs to market, he says.

Gen AI can help pharmaceutical companies , repurpose existing drugs, and create personalized therapies based on a patient’s genetic makeup, according to MSRcosmos, a global IT services provider.

In early 2024,  its AI-driven Clara computing platform targeting the healthcare industry and its BioNeMo, a gen AI platform for drug discovery.

Some biotech and pharma companies, , are promoting gen AI as the next big thing in drug discovery.

Cybersecurity and fraud detection

Several cybersecurity firms are using gen AI to enhance tools that look for suspicious or unusual behavior on a customer’s network and computing infrastructure. AI systems can also be used for  that predict fraudulent activities with great accuracy by analyzing transaction patterns and user behaviors, says , CEO of Conversica, a provider of conversational automation solutions.

For example, Palo Alto Networks offers the  security operations platform, which combines the company’s expertise in ML models and its data store along with Google’s BigQuery enterprise data warehouse and its Gemini AI model. The goal is to alert security analysts to threats in real time, while the cybersecurity platform continually learns about new threats.

Business process augmentation

Generative AI is finding a sweet spot in enterprise business process augmentation. Here, companies are exploring the use of gen AI to provide efficiencies for business-critical workflows, often unique to their verticals.

For example, some firms in the finance and insurance industries are using gen AI to assist underwriters evaluating prospective clients. Credibly, a lending platform for small businesses, uses gen AI, paired with machine learning, to evaluate loan risk and to speed up the lending process, says , co-CEO and founder of the company.

“Gen AI at Credibly is being used to give our underwriters superpowers,” he says. “As a fintech lending company, our success depends on fast and accurate risk assessment of business owners seeking financing.”

Nearly all insurance carriers had adopted gen AI or were interested in it as of late 2023, according to a . About 42% of insurers had already invested in gen AI, and about two-thirds expected a revenue boost of more than 10% through use of gen AI.

In the legal arena, legal information services giant LexisNexis is embracing generative AI to keep in front of what EVP and CTO sees as a disruptive threat in the company’s industry.

“We were all-hands-on-deck,” Reihl told CIO.com. “We did a major pivot because this was a game changer in terms of its interactive abilities, as well as the comprehensiveness of its answers and its data generation capabilities. It was just staggering in terms of its capabilities.”

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

Predictive analytics

While gen AI models have traditionally excelled at retrieving and summarizing information, organizations are now using the technology for .

For example, some companies use gen AI to predict shipping schedules, says , general partner at AI and robotics venture fund Interwoven Ventures.

Predictive analytics powered by traditional AI is nothing new, but gen AI excels in the task because of its ability to deal with unstructured data without requiring predefined algorithms, says Agmoni, former head of AI and robotic deployment at shipping company Maersk.

Shipping schedules can be unpredictable, with several factors affecting the time to get to the final destination, he says. A simple algorithm that looks at historical data isn’t enough to provide an accurate delivery date.

Shippers need multiple systems to share past and current data, including information such as the performance of multiple routes, weather, labor performance, and the financial market situation. “Being able to solve problems like this example can generate billions of dollars for participants, making the eagerness to find solutions very high,” Agmoni says.

Extracting unstructured data from multiple sources

The most sophisticated LLMs can help open organizations’ AI strategies to previously untapped unstructured data from text, videos, and voice messages. For example, some organizations are using gen AI to extract data from video surveillance systems, says , executive vice president at FourKites, vendor of a supply chain visibility platform.

“One of the most exciting breakthroughs we’re seeing in gen AI is the ability to extract unstructured data that’s available across various applications but was previously too cumbersome or time-consuming, which has the potential to revolutionize the market,” he says.

Many shipping ports, for example, have video cameras that refresh regularly. “If we start capturing those frames, we can know exactly what trucks went in and came out by just looking at the license plates or container numbers,” he says. “Most ports have these cameras available 24/7, so the democratization of data and more natural ways of gathering data will be the lowest-hanging fruit that can unlock valuable shipping and tracking insights.”

Nagaswamy sees a big adoption trend toward LLMs with the capability to deal with multimodal inputs and outputs, even though accuracy may not reach 99% right away.

“We’ll see significant strides in the democratization of this capability in the coming year as people get comfortable with the notion of simply talking to an LLM or sending it an image or video,” Nagaswamy says.