Mikko Lehtimäki
Perustaja, Chief Data Scientist
icon

Agents, assistants and GPTs - news from OpenAI devday 2023

Language model agents and GPTs

OpenAI has just announced a new feature: GPT agents on their platform & API. These are customizable versions of their language model, ChatGPT, that you can easily tailor to your specific needs. The potential applications of these agents are vast and could be a game-changer for your business.

The new OpenAI GPTs and Assistants API come with several key features that enhance their functionality and customization options. Here are the highlights:

1. Customization: With GPTs, you can now create custom versions of ChatGPT that combine instructions, extra knowledge, and any combination of skills. This allows you to tailor the AI model to fit specific use cases and requirements.

2. New Assistants API: OpenAI has introduced the Assistants API, which makes it easier for developers to build their own assistive AI apps. This API enables developers to create AI applications with goals and the ability to call models and tools and run code. It also allows making full use of agents outside the ChatGPT interface.

3. Multimodal capabilities: The new GPTs also offer multimodal capabilities, including vision, image creation (DALL·E 3), and text-to-speech (TTS). This means that the AI models can process and generate content based on text, audio and visual inputs and interact with you through sound.

4. GPT-4 Turbo: OpenAI has launched a preview of GPT-4 Turbo, the next generation of the GPT-4 model, which is still the best LLM. No other model achieves the same text understanding and tool usage capabilities. GPT-4 Turbo is more capable and has knowledge of world events up to April 2023 and will keep updating. It features a 128k context window, allowing it to handle the equivalent of more than 300 pages of text in a single prompt. Additionally, GPT-4 Turbo offers improved performance and is available at a more cost-effective price compared to GPT-4.

5. Price cuts, repeatability, JSON compatibility and speed improvements across the board.

These new features provide users with greater control, customization options, and enhanced capabilities when using OpenAI's GPT models.

Unlocking Agent potential

What should your organization do? Start exploring the possibilities. Think about the repetitive tasks that consume your employees' time. Maybe it's handling customer service emails or compiling weekly sales reports. These are the kind of tasks that GPT agents excel at. By automating these tasks, you free up your team to focus on more strategic, value-adding activities.

Enterprises can unlock a wide range of potential use cases with agents. Here are a few examples:

1. Crafting Marketing Materials: Enterprises can use agents to create marketing materials that embody their brand voice and messaging. Agents can generate compelling copy for advertisements, social media posts, website content, and more.

2. Customer Support: Agents can assist support staff by answering customer questions and providing solutions to common problems. They can handle basic inquiries, provide product recommendations, and offer troubleshooting guidance.

3. Onboarding and Training: Enterprises can utilize agents to help onboard new employees and provide training materials. Agents can answer questions, provide explanations, and offer guidance on company policies, procedures, and best practices.

4. Data Analysis and Insights: Agents can assist in analyzing large datasets and extracting valuable insights. They can help identify trends, patterns, and correlations, enabling enterprises to make data-driven decisions more efficiently.

5. Personalized Recommendations: Agents can offer personalized recommendations to customers based on their preferences, past behavior, and purchase history. This can enhance the customer experience and drive sales by suggesting relevant products or services.

6. Workflow Automation: Agents can automate repetitive tasks and streamline workflows within the enterprise. They can handle administrative tasks, schedule appointments, manage calendars, and perform other routine activities, freeing up time for employees to focus on higher-value work.

These are just a few examples of the potential use cases that enterprises can unlock with agents. The flexibility and customization of agents allow them to adapt to various business needs and provide valuable assistance across different departments and industries.

Next steps

We've developed our own Conversational AI solution called YOKOT.AI, and with these announcements it just got a lot better. We are excited about the opportunities the new Assistants API and better multi-modality will enable. Softlandia understands the importance of privacy when deploying LLMs at scale. We have expertise in addressing the key aspects of privacy and LLMs, ensuring that organizations can utilize these tools while maintaining data security and compliance. Don't hesitate to reach out to their team to learn more about building on top of LLMs and exploring collaboration possibilities.