New AI developments are coming down the pike so rapidly that numbers can and do go out of date quite quickly. But two basic facts seem to be pretty clear for the moment: more and more businesses across industries are adopting AI capabilities, and many of those same businesses are “bleeding capital” on AI investments that don’t pay off.
Logistics in general—and last mile logistics in particular—isn’t a sector that’s especially susceptible to technology hype. When it comes to the newest and shiniest tech there will be a handful of early adopters, but in the past most logistics operators have taken a wait and see approach.
That’s undoubtedly sensible. The challenge is speeding up that “wait and see” approach so that it moves at the pace of technological change. Given that the scope and capabilities of AI technology seem to change on a timescale of months, rather than years, it’s crucial to evaluate your options early and often. It’s all too easy to dismiss something as too risky and miss the moment when it becomes an easy, low-risk proposition for saving time and money.
Luckily, we’re here to help. We’ve been developing AI capabilities in our product for years, and we’ve seen first hand the ways in which it does and doesn’t contribute to lower delivery costs, happier customers, and more connected logistics operations. Based on that, we’ve compiled the top 5 ways to actually reduce logistics costs with AI.
At CBC Transport, a huge percentage of the team’s time was taken up answering simple, repetitive questions from customers.
"Most of the messages we were getting were simple things like, ‘What time is my delivery?’ or ‘Are you on your way?’” said owner Cesar Bermudez. “They weren’t hard questions, but they were slowing us down and we needed to respond to our valued customers."
The introduction of an AI-powered chat agent into CBC’s two-way customer communications solved that problem in a heartbeat. The agent was configured to answer the most common questions in a clear, conversational way and then escalate to a human for anything that can’t be answered easily.
You can get a quick, 5-step guide to implementing this kind of technology here: The 5-Step Implementation Guide for AI in Delivery Experience. But the high level overview is this: once you have two-way communications between your team and your customers enabled, implementing a chat agent on top of that workflow is the work of a few minutes, and it can act as a front line of defense for customer questions and inquiries.
This is a huge win-win. Customers get faster answers to routine questions, and you save a huge amount of time and money since you don’t have to leverage serious person-hours answering the same questions over and over again.
Not only does the use of an AI powered chat agent like the one we described above make it possible for you to reduce customer phone calls and messages and the costs associated with them, it also empowers you to speed up response rates (CBC decreased them by 30%, for instance) and provide round-the-clock support without increasing the size of your team.
This saves on labor costs while simultaneously improving customer experience. The best part about this is that it’s an extremely low risk way to deploy AI in your logistics chain. Agentic AI always has the risk of errors, but if you configure your chat agent with a clear scope then you ensure that it’s only ever performing actions that are fairly foolproof.
Anything you can do to make life smoother and easier for drivers can directly reduce delivery costs in the form of fuel and driver pay. But it’s difficult to insert technology into driver workflows without having an interruptive or disruptive impact, which is the last thing you want.
This is where AI can actually play a really valuable role. How? By providing location-based briefings via voice note for each and every stop along a route.
The best practice here is to use a mix of order details from your system (e.g. any instructions the customer has provided) and data from around the web (such as what the parking situation is typically like or where the delivery access is for the building) to automatically generate location-based intelligence. This can then be weaved into the driver’s normal workflow via their driver mobile app.
The result here is that drivers are less likely to get bogged down with the most frustrating parts of the job—hunting for parking, navigating delivery sites, gaining access to buildings, etc. Instead, they can focus on delighting customers while getting jobs done more smoothly.
With AI that’s powerful enough to enable just one additional delivery per route, you can bring down cost per delivery and boost customer satisfaction.
In all of the hype about generative AI, large language models, agentic AI, and the like, it can be easy to forget that artificial intelligence as broadly conceived is a pretty old field of study. AI research can be said to have begun as early as the 1950s. The result of the rapid progress in the field over time has been the AI effect, which can be summed up as: “AI is whatever hasn't been done yet.”
But let’s not forget that AI and machine learning have been leveraged to improve route optimization for many years at this point. By leveraging previous delivery data as well as up-to-date information about traffic and weather patterns, AI can enable you to improve route density and more accurately match your delivery capacity to your fleet size and your order volumes.
The result is that you can decrease miles driven per stop by 10% or more—saving you money in terms of fuel and labor costs. This is something where the route optimization solution you implement should already have these capabilities baked in, and your teams shouldn’t have to worry about the technology.
By the same logic we saw in the previous section, AI and machine learning can also help improve your route accuracy and thereby increase first attempt delivery rates.
Ryder, one of the premier 3PLs in the country, was able to use machine learning-powered routing to achieve exactly that. They were able to achieve 98% on-time delivery rates, which meant that the number of late deliveries they had to remediate plummeted.
The impact here can be significant. Redelivery attempts are incredibly expensive—you essentially double your delivery costs if you’re lucky. And there’s always the added risk of damage to items when they’re driven back to the warehouse and potentially unloaded and loaded again. There’s your warehouse footprint to think about as well; if you’re constantly dealing with unplanned returns due to missed deliveries, you’re going to wind up paying for more warehouse space.
Those costs drop out of the picture rapidly when you’re able to get the right goods to the right place at the right time at a rate of 98% or better.
Each of the AI-powered cost reduction methods we’ve been discussing is based on solid, trustworthy technology that logistics leaders are already adopting. As AI in logistics evolves, more use cases like this will emerge (and the power of leveraging AI for existing use cases will multiply), but you’ll always need to find ways of separating the hype from the real impact.
Leveraging AI in logistics in a scattershot way and adopting individual new technologies piecemeal isn’t necessarily the path to long term success. Instead, the best way to make sure that you’re getting value out of emerging technologies without adopting risky or unproven capabilities is to partner with a trusted technology provider with a track record of smart, measured AI innovation.
In this way, you can continue to find new and practical ways to reduce logistics costs using AI over time.