Top 10 Business Use Cases for Generative AI

We present a list of the top 10 business use cases that were recently considered too difficult or impossible and can now be enabled with generative AI.

Responding to RFPs
Creating a sales proposal may require answering hundreds of questions, especially in highly regulated industries. Responding to RFPs is the least liked business activity. Teams of overworked salespeople often pull all-nighters to hit a deadline, and some companies do hundreds of them per year. It is also the kind of work – question answering – that generative AI models excel at.

Strategic Account-based Marketing
With strategic account-based marketing (ABM), marketing teams analyze the strategic imperatives driving their customers’ purchasing decisions and design highly customized solutions uniquely enabled by their products. No marketing team could ever scale strategic ABM beyond just a few accounts. With generative AI, ABM can be applied to every account in the CRM system.

Market Research
95% of new products fail, and most do because companies build products that no one wants. Companies test new products by generating ideas, running experiments, and analyzing the results. A generative AI model can significantly increase the number and quality of experiments by creating numerous variations of landing pages, product images, ad copy, and other artifacts.

Unlocking Revenue Synergies
Personalization is the driving force of commerce. In the enterprise, personalization enables cross-selling and upselling. Unfortunately, the state-of-the-art algorithms break down without historical data, which is often the case in enterprise M&A. Generative AI enables effective personalization based on content rather than history. It can create jobs and unlock billions via revenue synergies across product groups and business units.

Customer Onboarding
In e-commerce, the product returns rate can be as high as 20-30%, creating a huge profit drag. Many customers return products because they cannot figure out how to assemble, install, or use them. Large foundation models are capable of problem-solving and commonsense reasoning. Armed with nothing but a product manual, such a model can make a massive difference by assisting with customer onboarding.

Legal Research
When working on briefs, motions, and opinions, law professionals search through thousands of cases and court decisions stored in public and private repositories. Unfortunately, with keyword-based searches, they get many misses and false positives. Large language models solve this problem by enabling search based on meaning, similarity, and relevance. They are also very good at summarizing documents.

Multi-lingual HR Chatbots
With generative AI, the entire company intranet – text, images, audio recordings, and videos – can be indexed and provided to employees in the form of a question-answering chatbot. Unlike search engines, foundation models can derive answers that cannot be looked up, such as calculating the number of vacation days given a start date. They can also effortlessly communicate in many languages.

Risk Management
Sentiment analysis of news, social media posts, product reviews, earnings call transcripts, and other content is essential to risk management. It is also a source of alpha in many investment management strategies. Due to industry jargon and specifics, it was previously nearly impossible to train a single sentiment analysis model that works across industries. LLMs perform this task out of the box.

Enterprise Application Integration
Business processes depend on dozens of applications. They are typically hardcoded in process management systems. Changing a process is expensive, negatively impacting business velocity and adaptability. With some tuning, foundation models can auto-assemble business processes in response to task definitions, reducing the cost of digital transformation.

Multi-modal Diagnostics
In healthcare, manufacturing, and other industries, diagnostic tasks are performed by task-specific models trained to detect a narrow range of problems based on signals of one modality. For example, a computer vision model can be trained to detect a specific type of lung cancer in X-ray images. With multi-modal foundation models, detection can be based on several types of signals and cover a broader range of problems.

Can GPT-4 Outsmart Wall Street Stock Pickers?

I provided a large language model with last year's market outlook reports and asked it to generate aggressive growth portfolios. Next, I backtested the portfolios against historical market data. Here are the results.

Market Segmentation and Lead Enrichment with LLMs

Market segmentation is an essential part of any go-to-market strategy. Segmenting a large list of prospects can be very challenging. Fortunately, it is precisely the kind of work that large language models excel at, but it does take a few tricks.

AI Takers vs. Shapers: Harnessing Generative AI for Competitive Edge

With AI assistants popping up in every tool and service, it's easy to be an AI taker. However, using generally available tools does not build a lasting competitive advantage. One must learn to shape AI to make a difference.