For artificial intelligence to make a difference, companies need tools that are tailored to specific industries or areas of responsibility. AI chatbots based on LLMs are good at communicating and giving advice, but they often lack specialist knowledge.

AI Olympians - Specialized tools for specific disciplines

To illustrate this, let's take a look at the upcoming Olympic Games. AI base models are like the core characteristics of a good Olympian: they represent fitness, dedication and a relentless pursuit of excellence. However, there are 32 sports with over 400 different events, each requiring different skills and experience - much like the different industries and job roles in business. And while AI provides the core technology for various products and services, each of these individual products must be equipped with the appropriate skills to provide added value.

Each athlete is highly specialized for their sport, a sprinter optimizes themselves to be powerful and fast over short distances, which at the same time means they are not suited for long distance running. The best-known AI chatbots are all-rounders. They are designed to have a general knowledge of the world across a wide range of topics. Although a chatbot can provide superficial information on a wide range of topics, it is not necessarily suitable for more specific tasks.

A new kind of AI ecosystem

Consider, for example, an AI-supported universal search tool: it must be able to find and retrieve the right information quickly. Like a sprinter, it is optimized to save crucial seconds on each run. However, there are other tasks that require an AI that is designed for sustained performance over a longer period of time, much like a long-distance runner. For example, predictive AI models for business forecasting need to learn the activity patterns of individual businesses by analyzing historical data and building this knowledge over time. By specializing in the company's business processes, they can make predictions about the company's future development based on past results. Predictive AI models must also constantly adapt their forecasts to ever-changing operations and external business factors. However, recent research from MIT's Computer Science and Artificial Intelligence Laboratory has shown that multiple large-scale language models working together provide more accurate results, perhaps creating a new kind of AI ecosystem.

Two paths to the future of AI

If we look at the development of the AI ecosystem, there are two different paths the industry can take:

1. The first is a race to create the best AI model for general purposes. This AI system would excel at a variety of tasks, in the same way that a decathlete is able to perform in a variety of events, from sprinting to long jump to pole vaulting. The benefit would be a seamless user experience that streamlines an employee's workflow. However, like the decathlete who can't match the specialist's performance in any single discipline, a general AI model might struggle to achieve the same level of performance as more focused tools.
   
2. The alternative path sees the future AI ecosystem as a network of specialized AI products, similar to a team of specialized athletes. In this model, each AI focuses on a specific area, much like individual athletes focus on specific sports. This approach mirrors how an Olympic team combines the talents of sprinters, swimmers and gymnasts to maximize the collective medal potential for their country. Specialization ensures that each AI performs optimally in its domain, often exceeding the capabilities of an all-purpose system. However, the success of this networked approach requires sophisticated coordination and interoperability to create a seamless experience for users.

Whether strategically focusing on a specialty area to optimize the probability of winning, to a broader approach that aims to win as many gold medals as possible in as many disciplines as possible, the type of AI ecosystem each company will implement is conditioned by its own goals.

The author Andy Wilson is Senior Director of New Product Solutions at Dropbox.

© Dropbox

For some companies, growth through gaining market share in a fluid market requires speed and agility, while customer retention in a stagnant market demands a more strategic, long-term plan for others.