By Betsy Burton
The Next Boom: Vertical Generative AI Solutions
In the last 18 months, we’ve witnessed the emergence of several generalized large language models, such as those developed by OpenAI, Microsoft, and Google. These generative AI technologies are designed to address non-industrial or topic-specific interactions. While these models are intriguing for their broad applicability, they may prove too heavyweight for more specialized applications, particularly when considering edge devices.
We anticipate that generative AI will become significantly more compelling as we witness the emergence of industry-specific or application-specific models.
Industry and Application-Specific AI Models
In recent weeks, there have been numerous significant announcements regarding products and services tailored to address specific industries or tasks.
For instance, specific generative AI models are already being utilized in the medical field to streamline administrative tasks, such as taking doctor’s notes during patient interactions.
While these notes still require review by doctors or their assistants, they have substantially increased the accuracy of documentation while reducing the time needed to make these notes available to patients.
Just this week, Nvidia and Johnson & Johnson unveiled plans to develop AI applications for surgery. Similarly, Salesforce and Quilt are both independently building AI models aimed at assisting salespeople in closing deals.
These examples are just a glimpse of the targeted generative AI models being developed for various sectors, including legal firms, advertisers, financial institutions, defense, and manufacturing.
Small Language Models
These announcements mark the initial phase of vertical AI solutions utilizing small language models (SLMs).
SMLs are AI models focused on specific language generation and processing tasks, tailored to particular use cases, applications, or industries. They are characterized by their lightweight neural networks, fewer parameters, and streamlined training data.
Bottom Line
While recent attention has been rightfully focused on generative AI models, they represent just the tip of the iceberg.
Over the next three years, we anticipate a substantial growth in more targeted and focused industry-specific generative AI models supported by SLMs and lightweight neural networks.
Clients evaluating these industry or application-specific models must ensure that the provider fully comprehends their specific industry and task requirements, both current and future. This entails a clear understanding of their future business strategy, needs, and governance, not just the AI technologies involved.
