The advent of generative AI tools has been hailed as a revolutionary breakthrough in the world of business, promising to save companies significant amounts of time and money. However, one of the key challenges that these AI models face is the ability to admit when they do not know what to do. According to Microsoft vice president, Vik Singh, the lack of humility in AI models is a critical issue that needs to be addressed. Singh emphasizes the importance of a model being able to raise its hand and ask for help when it is unsure of the correct answer. Despite the rapid deployment of generative AI applications like ChatGPT by tech giants such as Microsoft and Google, these models still “hallucinate” or invent answers, posing a risk to corporate customers who rely on them for accurate information.
Singh highlights the frustration among customers with the meanderings of AI systems like Microsoft’s Copilot, which are expected to provide accurate and reliable responses. The goal is to develop AI chatbots that are capable of admitting when they do not have the right answer and seek assistance when needed. Singh suggests that even if a model has to turn to a human for help in 50 percent of cases, it can still result in significant cost savings for businesses. For instance, at one Microsoft client, the cost of having a customer service representative answer each new request is $8, indicating the potential for substantial savings by utilizing AI models.
Enhancing Productivity and Cost Savings
Microsoft’s Copilot, an AI assistant specializing in sales, accounting, and online services, aims to bring in revenue and justify investments in generative AI. The primary focus is on enhancing productivity and hopefully driving profits for businesses. By automating repetitive tasks and freeing up time for salespeople to focus on customer interactions, companies like Lumen have reported saving around $50 million annually. Singh’s teams are actively working on integrating Copilot into Microsoft’s software to make it more autonomous, allowing it to perform tasks like drafting emails on behalf of sales representatives based on previous interactions.
While some industry leaders predict that generative AI tools will lead to job losses, others like Vik Singh believe that technology has the potential to make humans more creative and open up new job opportunities. Singh references his experience at Yahoo in 2008, where the implementation of AI systems led to faster content renewal, increased click-through rates, and a greater demand for new articles. Rather than displacing jobs, the deployment of AI tools is expected to streamline processes, boost productivity, and drive innovation in the business world.
The promises and pitfalls of generative AI tools in the business world present a unique set of challenges and opportunities. While the potential for cost savings, enhanced productivity, and revenue generation is immense, the need for AI models to exhibit humility, accuracy, and reliability remains paramount. By striking a balance between technological advancement and human creativity, businesses can leverage generative AI tools to achieve sustainable growth and competitive advantage in an increasingly digital landscape.