Intel

Intel is a legacy semiconductor giant that helped define the PC and server eras and continues to operate at a massive global scale today.

In 2025, it generates over 50 billion dollars in annual revenue, employs around 100,000 people, and remains one of the world’s largest chip makers, powering PCs, data centers, and AI infrastructure.

In today’s AI at the Top, we’ll explore how Intel uses AI internally to reinvent its legacy and compete with modern enterprises.

Similar to JPMC and McKinsey, Intel developed internal, proprietary LLMs trained on its enterprise data.

The platform has two parts. iForge and iGPT.

iForge is the backend engine that runs and connects models with data sources.

iGPT is the front end that provides no-code AI tools, enabling employees to build AI assistants for specific use cases.

Employees use iGPT to create personal assistants without writing code. Think of use cases like supply chain forecasting or legal contract reviews.

In a report published by Intel, numbers show 30,000 monthly users (internal employees) have created over 7,000 personal assistants.

The platform supports 436 AI use cases across Intel. The company aims to hit 1000 use cases by 2026.

Intel tracks every AI use case through an internal dashboard.

The dashboard serves two purposes.

This prevents teams from building redundant solutions and helps Intel achieve managed, intelligent growth of AI.

The impact so far: 321,000 hours saved in 2024 with $305 million in business value.

Intel’s enterprise architecture team uses an AI agent to discover and summarize existing artifacts, such as policies, principles, standards, business process flows, and solution architecture diagrams.

Complex architecture documents become accessible to both technical and business audiences.

The goal is self-service architecture governance. Instead of waiting weeks for manual reviews, employees can upload a design document and receive instant feedback to improve quality.

Intel rebuilt its IT support bot, AskIT, using iGPT. The conversational AI delivers better responses to help employees solve IT issues.

As of mid-May 2025, over 11,000 employees have had more than 26,000 conversations with AskIT. The result is a 30% reduction in “How Do I…” support tickets. Intel expects to hit 50% reduction by now (end of 2025). We’ll know once the company publishes its annual report.

Intel’s factories use robots equipped with IoT sensors to roam facilities and perform inspections autonomously. The robots are:

The robots detect acoustic leaks, read gauges, and check thermal status with nearly 100% accuracy. This enables preemptive dispatch of repair and maintenance based on real-time data.

For predictive maintenance, Intel deployed sensors on pumps and blowers to monitor vibration. When vibration exceeds specific thresholds, the system alerts maintenance teams before equipment fails.

TL;DR: Saves thousands of dollars. Sometimes millions if you prevent blower failures.

Intel deployed 25,000 AI PCs equipped with Intel Core Ultra processors. These aren’t traditional laptops. They have a CPU, GPU, and neural processing unit (NPU) working together.

AI runs locally instead of the cloud, reducing latency and network load. This also mitigates data-in-motion security risks.

Intel anticipates cost reductions by eliminating mid-cycle refreshes and reducing IT support requirements. Thanks to AI’s self-diagnosis and self-repair capabilities.

Intel operates 54 data center modules at 15 sites with a total capacity of 133 megawatts, housing more than 454,000 servers.

The company’s data center strategy has generated cumulative savings of more than $9.6 billion from 2010 to 2023.

But hardware is just a part of the story. Intel is obsessed with data quality because poor data means poor AI decisions.

In the second phase of their ERP transformation, Intel achieved a 98% data quality pass rate, exceeding the 80% goal. Fewer than 10% of total defects identified during system testing were due to data quality issues.

Intel completed a massive data analytics platform modernization, migrating a large business unit’s data to a cloud-based data lakehouse solution. The project moved over 4,000 objects, supports more than 3,000 users, and integrates data from over 100 sources.

All this in just four quarters.

The results: 65% decrease in view processing times, 18% reduction of object footprint, 50% drop in incidents, and 58% decrease in job failures.

Clean, accessible data enables Intel’s 436 AI use cases.

Without the data foundation, the AI tools would produce unreliable results. We have seen this repeatedly in companies we have covered previously.

Intel aims to achieve operational excellence with its tech investments. In the company’s words:

Intel’s 2025 priorities included developing a roadmap for future AI capabilities, defining additional use cases, and driving employee awareness through its “AI Everywhere” program.

The program focuses on culture, awareness, and training. Intel uses surveys to determine employees’ baseline AI awareness, then offers training sessions and meetups.

There isn’t much publicly available information about Intel, so we’ll have to wait till the next earnings call or AI reports to learn about tangible ROI on its investments.

With Lip-Bu Tan as the new CEO since March 2025, Intel is set to make its processes faster and reliable with AI.

2026 will be an interesting year to observe enterprises.

YOU DECIDE

YOUR TURN