• Since 2015, Daily Harvest's goal has been delivering healthy meals to customers across the US.
• It's using AI across facets of the business, including customer experience and packaging.
• This article is part of "How AI is Changing Everything: Small Business," a series exploring how small companies are using AI for success.

Daily Harvest is on a mission to bring nutritious food to consumers — with products that span heart-healthy harvest bowls to high-protein smoothies.

The meal delivery service, founded in 2015, ships orders throughout the US and sells its frozen food products in big-box and grocery stores like Costco and Kroger, and that all involves complex logistics.

While the brand's reach is large, its team is not — Daily Harvest employs fewer than 200 people.

As it has built an omnichannel business, Daily Harvest has integrated artificial intelligence throughout its operations, Jackson Mlawer, the company's director of product management, said. It's using AI for product recommendations, customer experience improvements, and packaging efficiency.

"Over the last year, we've invested a lot more in hyperfocusing on areas that we know over the course of our history as a business have driven success with AI," Mlawer said.

AI now touches nearly every facet of business operations, but the conversation often centers on how large enterprises or Big Tech companies are using it to scale their operations.

Small businesses like Daily Harvest are also tapping into AI, which shows how this segment can use the technology as a force multiplier.

Adopting a new AI approach

Mlawer said Daily Harvest has long been a tech-forward company. It initially built a custom tech stack, website, and mobile app, as well as internal tools. But beginning last year, the company turned to specialized, third-party AI solutions and has taken a more strategic approach with its AI investments.

Daily Harvest uses AI to drive product searches and recommendations for the 100 or so SKUs, or product codes, in its catalog. Sifting through all this data to recommend products based on each customer's dietary needs and unique preferences would be nearly impossible if it relied on humans alone. AI allows the company to quickly analyze customers' orders and website browsing history to incentivize them to try new products and reorder.

"We don't want customers to be fatigued with our food," Mlawer said. "We want them to try new products that we're launching or try things they've never tried so they can feel like the assortment is evergreen and it meets a bunch of different needs that they have throughout their life cycle." That's why ensuring the company has a robust data engine that can support product recommendations is so important, he added.

AI allows Daily Harvest to enhance other parts of its customer experience. The company has always used chatbots in its customer care channels, but AI allows it to drive more self-service and lets customers easily make changes to their plans, such as skipping their next order.

Previously, its customer care team would receive a help ticket and speak with the customer to resolve their issue. Last year, Daily Harvest implemented an AI chatbot that provides more rapid responses to customer queries. It also developed an AI model that categorizes customers at risk of canceling their subscriptions and directs them to a customer service agent for more personalized service, which improves retention.

"Not only are we seeing cost savings from enabling more of this automation, but we're seeing that our customer success scores — their actual satisfaction scores — have actually increased with the implementation of more AI in our care channels," Mlawer said.

Taking the friction out of fulfillment

AI also helps Daily Harvest enhance its fulfillment experience.

The company ships its boxes directly to consumers, so it's critical they are delivered on time and in good condition. Daily Harvest uses AI to determine how much dry ice to pack based on the number of products in each box and the weather at its final destination.

"How do we properly pack a box? Do we need 10 pounds of dry ice? Five pounds? Should I use a small box or a large box?" Mlawer said. "All these compounding factors are really intended to learn as AI is used, and then deliver, essentially, a frozen, on time, in-full package to customers."

With AI, Daily Harvest is improving packaging efficiency and the customer experience, but the technology isn't a solution for everything.

AI is only as good as the data it ingests, and there can be some growing pains in the beginning. For example, when Daily Harvest launched its new chatbot, the tool would sometimes provide incorrect responses to customers. The company adjusted how it trained and coached its AI model, pivoting to provide manual responses and focusing on a few use cases until the model had enough accurate data to deliver better results.

Within its fulfillment operations, Daily Harvest has also implemented quality-control measures, in which employees spot anomalies and double-check box configurations and packaging.

Mlawer said small businesses exploring AI should be tactical about how they use it. They don't need to apply it in every area. Instead, they can use it to augment specific parts of their operations and address the most pressing problems.

"Be very strategic with the type of AI enhancements that you feel your business will drive the highest value from," Mlawer said.

• Based in Houston, 3 Men Movers is a privately owned moving company.
• The small business uses AI to reduce driver distraction and find the most efficient routes.
• This article is part of "How AI is Changing Everything: Small Business," a series exploring how small businesses are using AI for success.

More than 25 million Americans move each year. And they expect their most precious items to arrive in one piece — whether they're traveling hundreds or thousands of miles.

So moving companies such as Texas' 3 Men Movers are increasingly turning to AI to meet these expectations.

Founded in 1985, long before artificial intelligence was ubiquitous, 3 Men Movers has adapted its business for the digital age. The company started testing and implementing AI-powered solutions in the late 2010s, its CEO, Jacky Fischer, said.

Today, 3 Men Movers uses the technology to detect distracted drivers and optimize route planning. It's just one of the millions of small businesses in the US that leverage AI in their everyday operations, illustrating how much the technology has redefined business.

Creating an AI-powered moving company

Higher insurance costs and steadily growing accident rates led Fischer's team to implement AI. The company also wanted to improve efficiency and reduce costs, critical in the fiercely competitive moving industry.

"To prosper, we had to focus on safety and liability as early as possible," Fischer said in an email interview.

To improve safety, the company installed cameras in the driver's cabin. Live video is transferred to its server through an AI-enabled distracted-driver detection system. The system is trained to recognize when a driver is using a smartphone, eating, drinking, doing some personal grooming, smoking, or even yawning. If it detects distracted behavior, both the driver and the supervisor receive a notification.

In addition, 3 Men Movers uses advanced open-source routing-machine technology, which identifies the best routes between points.

"There is a big misconception that AI was born in 2020 with the launch of ChatGPT," Fischer said. "While it was a major turning point in AI adoption, machine learning and data analytics for OSRM were already used in pioneering industries such as finance, telecommunications, and logistics."

The tech helps drivers at 3 Men Movers avoid high-traffic areas, high-crime zones, places with high crash rates, restricted areas, and even environmental hazards. It also allows the company to route in the most optimal way that will avoid liability and reduce risks, Fischer said.

Balancing AI benefits and risks

So far, implementing AI has paid dividends for 3 Men Movers. Comparing the AI detection reports to manual reports, the company has determined that the distracted driver detection system has an accuracy of 91% and prevents 80% of distractions. In the first three months of use, the system reduced the company's accident rate by 4.5%, according to Fischer.

Despite these results, implementing AI wasn't seamless. One of the biggest challenges, Fischer said, was integrating the distracted-driver detection system. Some of the initial products the company tried returned too many false positives, such as misinterpreting a driver's moving patterns at different speeds.

"It was quite challenging to keep the balance between staying competitive and avoiding more drawbacks than benefits due to adopting bleeding-edge tech," Fischer said. "This is why we always keep an eye on the benchmarks and do microtests before adopting any tech, including AI."

Fischer added that testing is key for any small business adopting AI. She urges other small-business owners to remember the risks of AI. To validate the technology, companies should always ask tech providers for proof points, such as case studies, information on how their results compare with industry benchmarks, and who will be responsible for implementation and false positives or hallucinations, meaning when AI tools deliver misleading or inaccurate results.

Along with testing, Fischer said transparency is critical for AI implementation. Her company is transparent with its team about when, how, and why the AI solution is interacting with them and their data. The company has also created a feedback loop so that every team member can share their suggestions and complaints about each solution.

"AI will only augment and empower," Fischer said, "but it will never replace or lead the people."