Overview

As GenerativeAI has moved from research to consumer use to enterprise use, a series of enterprise use cases have come to the fore. These areas can benefit from GenAI now.

Software development

Generative AI is helping make software development more productive by automating code generation, documentation, and bug detection, allowing developers to focus on higher-level problem solving and design. It also accelerates maintenance by suggesting optimizations and providing instant solutions for repetitive tasks, reducing the time spent on routine coding activities. This particularly true for monotonous efforts like code upgrades. In their Q2/2024 earnings call, Amazon’s CEO Andy Jassy said that they saved 4,500 work years of effort by using GenAI to migrate 30,000 product applications from Java 8 or 11 to Java 17.

Customer support

Replacing or augmenting call center agents with Generative AI “bots” has become all the rage. In some cases, the customer service chatbots make existing reps more productive. In other cases, they replace the call center reps altogether. Given the recent release of OpenAI’s Advanced Voice Capability, this use case should only get stronger in the future. Fintech company Klarna says that it replaced 700 call center employees with an AI chatbot and will reap $40M of additional profit in 2024 from that move.

Sales & Marketing

In most cases, enterprise sales efforts revolve around better productivity and focus for salespeople. Salesforce’s Einstein GPT integrates with CRM systems to suggest next-best actions, such as recommending which customers to prioritize or what offers to present, leading to more targeted and successful sales efforts. Marketing is about automatic personalization. Dick’s Sporting Goods has used AI, specifically through a partnership with Metrical, to significantly improve its outbound marketing emails by using predictive AI to dynamically present personalized content to customers, such as targeted messages, offers, and relevant product recommendations.

New features

GenAI is being added to existing products to make those products more compelling and useful. Adobe has incorporated generative AI into its Creative Cloud suite with tools like Adobe Firefly, which allows users to create images or modify photos using text-based prompts. Outside of the software industry, H&M has utilized generative AI in a collaboration with Google to create a “Data Dress” that analyzes users’ social media activity and lifestyle data to generate unique, personalized dress designs.

Back-office automation

The summarization capabilities of GenAI are being successfully used to automate or semi-automate back-office tasks. Contract review is one area of high interest. One example of an enterprise using generative AI for legal document review is Addleshaw Goddard LLP, a UK-based law firm. They have developed a solution called AGPT, which is specifically used for M&A transaction legal due diligence. The firm conducted research to optimize their generative AI system, achieving a significant increase in the accuracy of contract reviews—from 74% to 95%. This tool helps the firm automate the identification of legal provisions and improves efficiency by reducing the time required for comprehensive contract analysis

Training

Enterprises are finding GenerativeAI can be a significant catalyst to training efforts. Gong uses generative AI to analyze recorded sales and customer service calls, extracting valuable insights such as customer sentiment, key conversation topics, and compliance issues. It provides real-time guidance to sales reps by suggesting talking points and strategies based on ongoing conversations, enabling more effective sales interactions. Additionally, Gong helps refine sales strategies by pinpointing which phrases or tactics lead to successful sales and offers personalized training modules for agents based on their performance metrics. Providence Health & Services is a not-for-profit Catholic healthcare system headquartered in Washington state. The health system includes 51 hospitals, more than 800 non-acute facilities, and numerous assisted living facilities in the western half of the United States. Providence is using Generative AI to help train physicians to have empathetic end-of-life discussions with patients.

Looking forward

The frontier LLMs continue to improve with new capabilities like true multimodality, reasoning, agentic behavior, etc. Smaller non-frontier models are being developed and released for specific tasks. At the same time, compound AI application architectures are being devised to combine many models into very accurate, useful, and cost-effective applications. Smart companies that implement GenAI for today’s use cases will have a “leg up” on their competitors as the state of the GenAI industry improves.