Intel AI Assistant Builder enables developers to create, optimize, and deploy AI-powered assistants with cutting-edge technology.
In order to improve user productivity, creativity, gaming, entertainment, security, and more, this new generation of PCs which debuted in December 2023 with the Intel Core Ultra processor family combines the power of a central processing unit (CPU), graphics processing unit (GPU), and neural processing unit (NPU).
What makes a great PC an AI PC is not just powerful hardware that provides the performance required for today’s artificial intelligence (AI) applications. The PC business has a huge opportunity: create new AI applications and user interfaces that capitalize on this cutting-edge technology.
The next step in ensuring AI PCs provide the greatest AI experience is to develop quick and effective methods for developing new AI assistant that aid with daily tasks.
Intel presented Intel AI Assistant Builder (code-named Project SuperBuilder) at CES Las Vegas in January. With this hassle-free method, software developers and computer manufacturers can quickly construct customized AI helpers for any use case.
Building these models no longer requires precise guesswork with AI Assistant Builder. It simplifies the process into three simple steps: select an AI assistant model that has already been created, download and install it, and then start the application.
Intel giving the sector the foundational elements it needs to produce its own AI-focused content. According to Olena Zhu, senior principal engineer at Intel’s Client Computing Group, “that allows a customers to shorten their overall development time and accelerate the rollout of their own smart solutions.”
How does it work?
Here’s how it operates:
- Users can upload or remove documents in a variety of formats (such as PDFs, spreadsheets, presentations, and text files) and ask questions via the app’s chat box.
- The AI assistant reacts according to the user’s question each time. It gains knowledge from the discussion history, relevant documents, and the large language models (LLMs) innate knowledge.
- Additionally, it has sophisticated AI safety safeguards, such as a secondary safety system that directs the model to react impartially and safely, and profanity filters.
- With its constantly expanding feature sets, customers can create their own AI assistant by selecting from a variety of sophisticated capabilities.
AI Assistant Builder is already being used by Intel’s major partners to create specialized models for specific purposes:
- In just three days, Acer integrated its camera vision technology to create a proof of concept (POC) AI sales assistant.
- At Kyobo Bookstore in South Korea, Intel and Samsung Electronics unveiled a creative AI chatbot that operates locally rather than on the cloud. Based on intelligent LLMs, it allows users to look for books and get tailored information.
- Asus developed a technical Q&A assistance for its NUC clients by utilizing Intel’s technologies.
AI Assistant Builder Speeds New Age of AI Helpers
It takes weeks or months to perfect AI assistants from the ground up, and it requires extensive software programming and AI expertise.
Only a few months after its initial proposal, AI Assistant Builder was being actively utilized by independent software providers and PC OEMs worldwide. Additionally, as Intel’s AI PC platforms advance, more clients are coming up to collaborate with the company.
“Large language model (LLM), retrieval-augmented generation (RAG), and other components optimized for performance and accuracy when run on AI PCs based on Intel CPUs, GPUs, and NPUs are the foundation of these AI assistants,” Zhu stated.
Each vertical or domain has its own set of procedures and feature requirements in addition to vertically tailored LLMs. At Intel, its collaborate with clients to develop AI capabilities and workflows that are specific to their needs.
Any Intel-powered AI PC that satisfies certain hardware requirements can run AI Assistant Builder: A PC with an Intel Core Ultra Processor (Series 1), 16 gigabytes of RAM, and integrated Intel graphics is the very minimum need. (In September, Intel unveiled the Intel Core Ultra Processors (Series 2), which are up to 50% less power-intensive than their predecessors and offer a 3x boost in AI performance.)
Another significant advantage of the software, as noted by Zhu, is that it operates fully locally, requiring no cloud connections to execute queries (although initially an Internet connection is required to download the toolkit). Security worries are also allayed by using AI assistants locally on the computer.
“Users gain peace of mind and have total control over the security of their personal data by running AI assistants locally on their AI PCs rather than relying on cloud-based instances,” Zhu said.
In the Hands of Major Customers
When Zhu and her team originally drew up the basic structure for AI Assistant Builder in February 2024, they were foresighted in seeing the need for a rapid and simple AI builder that would enable AI PCs to come to life.
Five months later, in June, the team had a working prototype that was prepared for clients.
Top customers of Intel receive the software, technical resources, and engineering consulting at no cost.
A Three-Day Smart Camera Solution
In just three days, Acer used Intel AI Assistant Builder to create a proof of concept for a camera-assisted online sales assistant.
The assistant can identify items during video conferences and offer product information in the video feed by utilizing Acer’s camera vision technology.
The Intel AI Assistant Builder‘s RAG mechanisms enable the assistant to learn product features. Customers of Acer’s AI PCs with Intel Core Ultra CPUs can take advantage of this technology. (Acer’s new AI PC, the Acer Swift 14 AI, has an all-day battery life of up to 29 hours and is powered by Intel Core Ultra Processors (Series 2).
More developers can now design their own AI helpers with Intel AI Assistant Builder. The possibilities this presents are exciting, particularly when using Acer’s camera vision technology,” stated Mark Yang, Acer Inc.’s associate vice president of compute software technology.
“Acer will keep advancing its long-standing collaboration with Intel and assist clients in enabling these applications by providing hardware and software solutions.”
