������ý

Toki Sherbakov, head of solutions architecture at OpenAI, shares advice on AI implementation at Samsara Beyond 2025 in San Diego. (Seth Clevenger/Transport Topics)

[Stay on top of transportation news: .]

SAN DIEGO — Rapid advances in artificial intelligence promise to transform operations in many industries, including freight transportation, as more powerful AI models become widely available at lower costs.

Although generative AI can solve increasingly complex problems, trucking companies should begin by applying AI to smaller, relatively simple tasks and expanding from there, said Toki Sherbakov, head of solutions architecture at , developer of the widely used AI-powered chatbot ChatGPT.

“Don’t overcomplicate things early,” he said. “Don’t boil the ocean with AI.”

Sherbakov shared that advice and offered his outlook on how AI will continue to improve in the near future while speaking at fleet technology vendor Samsara’s 2025 user conference, held June 23-26.

“People think AI should be used for the most complex, highest-value problem first. We don’t encourage you to start there,” Sherbakov said. “There’s usually a lot of low hanging fruit that’s high value, lower complexity.”

GPT-5 is here.

Rolling out to everyone starting today. — OpenAI (@OpenAI)

After putting AI tools such at ChatGPT in the hands of employees, companies can build greater AI literacy and eventually move on to more complex use cases.

“I guarantee you, as you start using it, you’ll have new ideas that spark up,” Sherbakov said.

He cited three ways for organizations to embrace AI.

First and foremost is building an AI-enabled workforce.

Putting AI to Work

“What this means is getting AI in the hands of all your workers,” Sherbakov said. “This is building AI literacy from the ground up, getting that groundswell of people to be very familiar with how to use AI in their day to day.”

Next is identifying ways for AI to automate some operations, especially repetitive workflows.

Finally, companies can begin building custom AI tools into their products and services to support their customers.

To serve as examples, Sherbakov highlighted how OpenAI has worked with home improvement retailer Lowe’s and agricultural equipment manufacturer John Deere to implement AI capabilities.

John Deere is using AI and machine vision to improve the efficiency of crop spraying. The company also has added an AI customer service feature that answers users’ questions about equipment and maintenance issues.

Lowe’s, meanwhile, has rolled out its Mylow AI assistant to its in-store workers, who use it as they interact with customers to help them make the right purchasing decisions for their specific projects.

Although AI and machine learning have existed for a long time, the current wave of generative AI took off with the release of ChatGPT in November 2022. OpenAI had planned that rollout as a “silent launch,” but the AI chatbot quickly went viral and became “one of the fastest-growing products of all time,” Sherbakov said. ChatGPT has expanded to more than 700 million weekly users as of August 2025, according to its developer.

Looking ahead, Sherbakov predicted that generative AI will continue improving at a swift pace.

AI models will have better reasoning capabilities to produce more accurate, higher-quality answers to more complex questions with fewer “hallucinations,” he said.

Affordability Factor

These AI models are also becoming much cheaper.

“The cost of intelligence is just dramatically going down,” Sherbakov said. “I expect this trend to continue. It’s kind of like the cost of electricity. It ends up hopefully becoming ubiquitous, where you would barely notice the cost.”

Finally, AI tools that are predominantly text based today will become more multimodal — capable of processing and interacting with different types of data, such as audio, images and video.

At the same time, AI is beginning to move beyond answering questions in real time to acting as independent “AI agents” that can handle certain tasks on their own without the need for a human to actively steer it.

This year in particular, agentic AI has started to become good enough to reliably perform those tasks and solve problems autonomously and independently, Sherbakov said.

Image

From left, Samsara’s McGee, OpenAI’s Sherbakov, and Samsara’s Welbourne and Merritt discuss use cases for generative AI.��(Seth Clevenger/Transport Topics)

Speakers from Samsara also emphasized the transformative nature of AI.

Sean McGee, Samsara’s head of product management, compared the dramatic acceleration of AI capabilities in current times to Moore’s law — the observation that the computing power of semiconductor chips has doubled roughly every two years for most of the past half century, sparking a wave of innovation.

“This is happening way faster than Moore’s Law,” McGee said of AI development. “It’s the fastest thing that we’ve really seen in technology. And it’s happening right now.”

As a result, it could become possible for fleets to apply AI everywhere in their businesses.

“You can really monitor all of your data, all of your operations with AI, just because of how incredibly cheap it’s becoming,” he said.

Samsara, for its part, has been implementing forms of AI in its fleet management products and in the company’s internal operations.

.⁦⁦⁩ has introduced a variety of new AI-powered products and capabilities, including DVIR automation, maintenance tools, integrated routing and navigation, additional onboard camera views, and even a wearable safety device for frontline workers. — Seth Clevenger (@SethClevenger)

Evan Welbourne, vice president of machine learning engineering at Samsara, outlined how the company has incorporated computer vision models into its safety and AI dash cam products and generative AI in its Samsara Assistant tool.

AI-enabled automation also has accelerated internal development work at Samsara. Welbourne said the company’s engineers have been able to generate about 20% more code in the same amount of time when they are using AI.

The company’s sales teams also use generative AI to better understand and support customers, added Zach Merritt, Samsara’s director of data science and AI.

“You can use these tools to really become an expert quickly and kind of treat it as this genius second brain or little buddy you’ve got to help you navigate the world,” Merritt said.

Want more news? Listen to today's daily briefing below��or go here for more info: