Exploring 5 Use Cases of Large Language Models

Chatbots and Virtual Agents

The first and perhaps most relatable use case is about chatbots and virtual agents. These AI-driven interfaces simulate human conversation, providing customer support, answering queries, and facilitating interactions 24/7. By leveraging LLMs, businesses can ensure their customers receive timely and relevant responses, enhancing user satisfaction and engagement.

We already see great success here, for example, Klarna reported an AI assistant handled two-thirds of all customer service chats.

• The AI assistant has had 2.3 million conversations, two-thirds of Klarna’s customer service chats
• It is doing the equivalent work of 700 full-time agents
• It is on par with human agents regarding customer satisfaction score
• It is more accurate in errand resolution, leading to a 25% drop in repeat inquiries
• Customers now resolve their errands in less than 2 mins compared to 11 mins previously

It's projected to boost Klarna's profits by USD 40 million in 2024, marking a remarkable return on investment from their technology investment in the single-digit millions. 

While Klarna's support team may require minimal intervention, even high-touch customer support departments can benefit from such systems, aiding staff in responding to inquiries more quickly.

Co-pilots 

Co-pilots are meant to work alongside humans to boost productivity and creativity. These LLMs assist in writing, coding, and designing tasks by offering suggestions, corrections, and even generating content. This collaboration between human intelligence and AI enriches the creative process and streamlines workflow, making complex tasks more manageable.

As co-pilots are already vastly embedded in the daily workflow of developers, there are evident reports about the success of these systems as well. Github, who released their co-pilot already 2 years ago, did an a/b test, with the following results;

• Higher rate of completing the task (78%, compared to 70% in the group without Copilot).
• A striking difference was that developers who used Copilot completed the task significantly faster (55%) than the developers who didn’t use Copilot.
  

Maybe even more importantly they also reported that;

• Between 60–75% of users reported they feel more fulfilled with their job, feel less frustrated when coding, and can focus on more satisfying work.
• Copilot helped them stay in the flow (73%) and preserve mental effort during repetitive tasks (87%).

Scott Galloway, an NYU professor, wrote an extensive piece on why AI is already responsible for the recent round of layoffs and record profits posted by tech companies across the board. 

Chatting with Data (RAGs)

RAG or Retrieval-Augmented Generation helps users to interact with and utilize data. This use case enables LLMs to pull information from various sources in real time to answer questions, make recommendations, or provide insights. It's like having a chat with data, where the LLM understands context and delivers precise information, making data more accessible and actionable for decision-making.

These use-case is are most often applicable to internal data within companies and therefore clear success stories with measurable effects are less available. However, OpenAI published a case study how they implemented such a system at a financial institution. The effectiveness of that system is very impressive. Initially, they started with an accuracy of 45%, but over time they managed to increase that to over 95% making it a handy tool for employees to retrieve internal information.
‍

Classification and Typical NLP Tasks

LLMs are very good in classification and other Natural Language Processing (NLP) tasks, such as sentiment analysis. These capabilities allow businesses to automate the understanding and categorization of text data, from customer feedback to market research. By using LLMs for these tasks, companies can gain deeper insights into their data, enhancing their strategies and operations.

For example, by leveraging classification capabilities, LLMs can effectively organize and categorize emails, support tickets, and social media posts, making it easier for businesses to prioritize and respond to customer needs. Such applications are critical for large businesses aiming to understand and engage with diverse customer bases efficiently. Moreover, LLMs play a crucial role in enhancing content moderation systems, identifying and filtering inappropriate or irrelevant material automatically. This ensures that online platforms remain safe and relevant for users, while reducing the workload on human moderators.

Historically, creating effective NLP systems was a challenging task, often requiring extensive manual effort to tailor algorithms for specific languages or tasks. With LLMs, not only have previously complex NLP use cases become more viable, but the results often surpass those of traditional systems. LLMs, with their advanced understanding and processing capabilities, can adapt to a wide range of tasks with minimal customization, offering better accuracy and efficiency.

Autonomous Agents

The development of autonomous agents represents the cutting edge of LLM applications. These agents can perform tasks, make decisions, and even learn from their interactions, all with minimal human oversight. From automating routine tasks to managing complex systems, autonomous agents offer a glimpse into the future of AI, where machines can take on more responsibility and initiative.

Even though this is probably the most complex use case, we also see a lot of development here. A clear example is in sales automation. For example, you can give a virtual SDR the assignment to grow your pipeline and give it access to your LinkedIn account. The virtual assistant will start:

• Discover the best-fit customers and establish a lead list
• Research all the leads in-depth
• Personalize outreach to drive conversions
• Engage with prospects on multiple channels to get you meetings

The results for systems like this are very early and mostly self-reported by companies that offer these systems, however, you don’t need to have a lot of fantasy to realize, that given these agents do their work well, the results can be amazing, allowing people in sales to just focus on the end of the funnel. 

And your business?

Large Language Models are transforming industries by offering innovative solutions to age-old problems and opening up new possibilities for interaction, creativity, and automation. We have just begun exploring and expanding the capabilities of LLMs, and their role in shaping the future of business and technology becomes increasingly evident. We hope that by understanding these five key use cases, leaders can start tinkering with how their organizations should position themselves, to drive efficiency, innovation, and growth.

‍