Three years ago, Luciano Sieber, chief supply chain officer of Colgate-Palmolive, planned a “Logistics Blitz” as the pandemic caused chaos in companies of all sizes.

The result gave Sieber a better understanding of how Colgate-Palmolivi moved its products around the world. But it encountered another problem: too much data.

About a year ago, Siber said he found a solution to Uber Freight. The long-term logistics and analytics departments of ride-hailing services have been developing new ways to use artificial intelligence to argue with massive amounts of data. Colgate-Palmolive became one of the first companies to use one of its latest products, AA logistics-centric LLM Uber Freight Call Insights AI.

Now, as part of existing supply chain software, Uber Freigh is more formally introducing a suite of AI capabilities to shippers around the world. These include the expansion of Insights AI, Uber Freight quietly launched in 2023, and the more than 30 AI agents built to “perform key logistics tasks throughout the freight lifecycle.”

Uber Freigh is not a person trying to tame unruly supply chains with modern artificial intelligence tools. Flexport announced its own suite of AI tools in February, and there are countless startups trying to help companies quarrel at data, reduce inventory and better predict supply and demand.

But Uber Freight bets that its AI solutions can immediately impact the bottom line of its blue chip customers and the bottom line of nearly 10,000 other shippers they work with. This is mainly because the knowledge base and relationship it built within eight years of its creation was to match long-distance truck drivers with shippers.

"Supply chains are essentially a data-rich issue. It's complex and nuanced, and AI can play a fundamental role in shaping it and accelerating it," said Lior Ron Ron, founder of Uber Freight.

“We’ve been building at this moment”

Uber Freigh’s easiest brokerage business model when it was launched in 2017. However, over the years, Uber's subsidiary has steadily grown into a more service provider to companies that ship goods around the world.

Many modern companies are trying to find ways to incorporate AI (usually a result of a hybrid); it is no surprise that Uber Freight puts the front and center of the technology in the center. After all, Ron's undergraduate work and his main paper are centered on artificial intelligence - "called 'neural networks' in the dark ages," he joked.

Ron continued to use machine learning technology when he ran Google Maps from 2007 to 2016. He said he saw "the potential to digitize the physical universe."

“Nine years ago, this supply chain was fundamentally a data-first, technology-first challenge that could be accelerated through data connectivity, over time, AI,” he said. “I think we’ve been working on it since I started Uber Freight.”

Ron said Uber Freight has used machine learning in its work since the beginning. But about two years ago, the team began experimenting with more advanced generative AI capabilities.

"That's not an easy road," Ron said. Uber Freigh's initial attempt to build a "co-pilot of logistics" was full of hallucinations and returned accurate answers only 60 to 70% of the time.

According to Ron, technology has now been "fight-tested" and is "driven actual business results" with an accuracy rate of 98%. The Insights AI model has received internal and external data training related to $20 billion worth of freight, which helps every year, the company said. According to Uber Freight, it also leverages multiple undisclosed AI models to “providing the best combination of price, accuracy and performance.”

Ron said this AI drives to create new ways for customers to collaborate with data related to their supply chains. They can ask Insights AI to pull up quickly, for example, the origin of a particular cargo that performs the worst. Or they can ask for "all shipments for CVS 2023". Ron stressed that queries can also be much more complex than this, and that the model remains the same all the time.

Like other popular LLM interfaces, Insights AI is also shown to customers; like other inference models, it will also show its work and be clear about where all the data comes from.

All of this allows clients to “get close to 100% accuracy of web insights right away, rather than creating what you want to know, sending it to some analysts and waiting for two weeks of PowerPoint presentation for discussion.”

'What do you want to know?'

Uber Freight has partnered with many Fortune 500 companies, but it found that a particularly willing partner in Colgate-Palmolive could try out Insights AI and its other new tools. According to Sieber, the group has provided an AI model for all its employees. This also allowed these workers to conduct mandatory training on internally developed AI ethics.

"I think it's great because it turns conversations from fear to 'how that makes me more efficient and (do) how to be a better professional and deliver more by accessing and using these new technologies," Sieber said.

For example, Siber said his company has used Insights AI to easily identify carriers that accept fewer carriers than contractual obligations to move. From there, they can figure out why these levels are lower and then come up with solutions to restore the carrier to compliance, or give it up in support of another.

Sieber said it was previously a challenge to solve in real time, as companies like Colgate-Palmolive have thousands of operators. These can all be used with different systems and workflows, and information on all these results has never really been managed through the center.

Both Sieber and Ron said the next step with AI has been looking for ways to create more proactive solutions. This is another place where Uber Freight can play to its data advantages, Ron said. "We know the facilities, we know the driveway, we know the price," he said. "What do you want to know?"

These more proactive integrations come in the form of alerts that tell customers like Colgate-Palmolive that they pay for certain routes or that there are faster options available for specific shipments.

Any such advice could save only a few hundred, or even a few thousand dollars. But summed up across the entire network, this can make a big difference.

That's why when asked, Siber quickly replied that Colgate-Palmolieve's chief financial officer was an executive who was most satisfied with the launch of Uber Freight. “He likes to see the logistics costs drop,” Siber said with a smile.