Transforming Grocery Shopping With AI-Powered Personalization

In a world where personalized experiences are increasingly central to consumer satisfaction, Ben McKean has built Hungryroot into a leader in the personalized grocery service industry. McKean, who founded the company in 2015, shares his journey from a traditional packaged food business to a cutting-edge, AI-driven platform that tailors grocery shopping to customers’ goals and preferences. In this interview, McKean discusses the pivotal moments that shaped Hungryroot’s success, the role of AI in revolutionizing food delivery, and his advice for entrepreneurs looking to disrupt industries.

Dave Knox: I want to start by asking if you can tell us more about Hungryroot and describe the early days when you founded the company.

Ben McKean: Hungryroot helps people eat healthy. Customers share their goals, budget, and lifestyle with us—whether they’re aiming to lose weight, boost energy, save time, or save money. Based on that, we recommend and deliver healthy groceries, easy recipes, and essential supplements tailored for them and their families.

Knox: About 10 years ago, you started Hungryroot, but around 2018, you made a major pivot in the business model. What drove that decision, and how has the company evolved since?

McKean: When I started Hungryroot in 2015, the vision was very different from what it is today. My goal was to create the next generation of consumer packaged goods (CPG) companies—essentially, a modern General Mills or Kraft Heinz. The idea was rooted in how deeply personal and emotional food is. It impacts how we feel every day, yet big packaged food companies are largely disconnected from the end consumer because they sell through grocery stores.

The premise was to build a direct-to-consumer CPG company and innovate at the food level. We launched with six SKUs, essentially reinventing the frozen dinner in a fresh, healthy format. These meals were packaged similarly to frozen dinners but featured fresh ingredients that could be sautéed in 5-7 minutes for a quick, healthy meal. Customers loved the product, but we quickly realized something crucial: to succeed in online food sales, you need to mimic the grocery shopping experience as closely as possible.

There were two reasons for this shift. First, convenience. People don’t want to buy yogurt from one website, ice cream from another, and fresh produce from yet another. They want a one-stop shop. Second, order value. Delivering food directly to consumers is expensive, and larger orders improve the economics for both the company and the customer.

Over our first four years, we expanded from six SKUs to 60 by 2019. While we grew to $25 million in revenue, we weren’t growing at the pace we wanted, and we weren’t profitable. That led us to make three significant pivots in our history. The biggest came in 2019, when we shifted from being a packaged food company to becoming a personalized grocery service. Instead of selling individual products like the best cookie dough or clean-ingredient sauces, we focused on curating and delivering groceries, recipes, and meals personalized to each customer’s lifestyle, budget, and goals.

This pivot allowed us to offer a more comprehensive and convenient service while improving our economics. It’s been the foundation of our growth and success since then. We also made earlier pivots in 2017 and 2018 to lay the groundwork, but the 2019 shift is what truly transformed the company into what it is today.

Knox: When meal kits came on the market as a new version of subscription commerce, there was a flood of competitors and significant venture capital investment in the space. With the pivots you just mentioned, how did you differentiate Hungryroot from other meal kits and food delivery services?

McKean: When I started the company, meal kits had been around for a year or two. A core principle of mine from the beginning was that you can either innovate on the demand side or the supply side in the food industry, but trying to do both is a recipe for failure because it’s simply too complex. Historically, food companies have focused on supply-side innovation—whether it’s introducing new ingredients, new packaging, or new production processes—and then outsourced demand to grocery stores.

Hungryroot, on the other hand, has always focused on demand-side innovation. We aim to create the best possible shopping experience for customers, while outsourcing as much of the supply side as possible. This approach took a few years to refine. Initially, in 2017, we were producing all our food ourselves. At the time, we offered about 30 SKUs, ranging from chopped produce to sauces to baked goods. Managing such a diverse supply chain was overly complex and inefficient. We realized we needed to shut down our in-house supply chain and outsource production to third parties.

This shift allowed us to focus on delivering a highly personalized consumer experience that helps people feel better on a daily basis. One major challenge for meal kit companies was their attempt to deliver a direct-to-consumer experience while building highly complex operations and supply chains. This often led to significant hurdles in scaling their businesses. By streamlining our operations and focusing solely on customer experience, we’ve been able to differentiate ourselves and succeed where others have struggled.

Knox: One of the things that’s always fascinated me about your business is how AI has been built into its core, long before it became the buzzword it is today. How did you come up with the idea of incorporating AI into what you’re doing?

McKean: We started building our AI in 2019. At that time, we were transitioning from a packaged food company with 60 SKUs to a grocery platform. The question was: how could we build a valuable and unique grocery service in a competitive market with players like FreshDirect, Instacart, and Amazon Fresh? With only 120 SKUs initially, compared to Instacart’s 100,000 SKUs, we needed a distinctive premise. That premise was personalization.

Our value proposition became filling your grocery cart for you based on your preferences, goals, and lifestyle. While customers have complete control and can modify every recommendation, the idea is to save time, achieve health goals, discover new foods, and even save money. To achieve this, we relied on AI.

In 2019, we built an operations research algorithm, a subfield of AI, which allowed us to encode human logic and hard constraints into our system. For example, if an item isn’t in stock, the algorithm knows not to recommend it. If a customer indicates they’re vegan, it ensures non-vegan products aren’t suggested, no matter how much the system might think they’d enjoy them.

Initially, the algorithm worked with about 100 parameters. For example, if you love salmon and a recipe includes it, the system might assign a “plus seven” to that recipe. These weights were manually fine-tuned by our team through debate and testing. While it may sound archaic now, that foundational algorithm still underpins our AI system today.

The difference now is that we’ve integrated 10 machine learning models to replace manual tuning, making the system more scalable and sophisticated. This evolution has allowed us to continuously improve our recommendations and deliver an even more personalized experience for our customers

Knox: Do you showcase AI’s role to customers, or let them experience the results without emphasizing the technology?

McKean: Until mid-2024, we didn’t mention AI to consumers due to skepticism. Instead, we branded our technology as “SmartCart.” This branding highlights the benefits without directly referencing AI.

SmartCart encourages customer engagement, such as reviewing and adjusting their carts or rating items. This feedback is invaluable for improving the system. The branding has fostered positive reactions and engagement. For example, in our Facebook group of over 30,000 customers, people now discuss the algorithm’s improvements. This focus on benefits, rather than technical jargon, has resonated well.

Knox: The pandemic presented both opportunities and challenges. How did Hungryroot navigate this period?

McKean: Our COVID story is unique. Since we pivoted in late 2019, we were essentially a six-month-old company when the pandemic began. We didn’t experience the dramatic spikes and declines others did. Instead, we steadily grew, increasing revenue from $25 million in 2019 to over $500 million annually.

Our SKU count expanded significantly during this time, from 60 to nearly 1,000 today. While we had a brief surge in demand, we paused marketing to manage the influx. More impactful was the shift to remote work. We’ve remained a remote-first company, which we see as a competitive advantage.

From a community perspective, we launched a frontline workers initiative, providing free deliveries to support those on the frontlines. This initiative reflected our commitment to supporting both our employees and the broader community.

Knox: What does the future hold for Hungryroot?

McKean: We believe that in the future, people will rely on AI to help make their food choices. Humans are notoriously bad at choosing food. The average American spends five hours a week planning meals, shopping, and deciding what to buy—five hours that are rarely enjoyable. Our target customer is the overwhelmed suburban mom, who already has a long to-do list and loves that Hungryroot helps her save time on something that feels like a burden.

Despite all this planning, 70% of what we buy at the grocery store is the same as the week before. The challenge of meal planning leads us to fall back on what’s familiar. However, 30% of what we buy ends up being thrown out because it spoils before we can use it. And yet, 82% of Americans are metabolically unhealthy, meaning our food choices are often harming our health. While not everyone fits this mold, the majority of Americans struggle to make good food choices. AI can address most of this.

AI saves time by pre-populating your shopping cart based on explicit preferences and feedback, but more importantly, it uses implicit data—how you shop, what you remove or add to your cart. It fills your cart with items you’re likely to love. Our long-term retention is closely tied to how much you trust the recommendations: the more you rely on the AI, the better your experience.

When it works, AI takes a lot off your plate—literally and figuratively. It helps you achieve complex health goals for yourself and your family. Whether you’re trying to lose weight, your daughter is eating vegan, or your son needs more protein, the AI can sift through all the variables and provide you with food that supports your goals while making you feel good.

We strongly believe this is the future of shopping, and at Hungryroot, we are pioneering this AI-driven shopping experience. It’s all about refining the process, being data-driven, and creating an exceptional user experience, all while becoming known for offering that experience.

Knox: What advice do you give to entrepreneurs looking to disrupt industries and build successful businesses?

McKean: It depends on where the entrepreneur is in their journey. Entrepreneurship is a process. I fell in love with it when I was 18 and started my first company, and now, with Hungryroot being my third business, I’m probably more passionate about it than ever. Each stage of the journey is different. In the early stages, it’s all about action and learning. I talk to a lot of aspiring entrepreneurs, and most of them don’t end up actually starting a business. There are so many reasons not to—why it’s a bad idea, why it won’t work—and you’ve got to push through those obstacles. That doesn’t mean ignoring them, but when you receive advice, it’s okay to adjust your direction by just a few degrees. Over time, those small shifts can lead to a major pivot. Early on, especially for first-time entrepreneurs, it’s easy to be swayed by advice and constantly shift directions, which can be discouraging and prevent you from learning and making progress.

For early-stage entrepreneurs, my advice is simple: stay focused. As you get further along, the approach becomes more personal, depending on your goals and market conditions. But you must aim to be lean and profitable as early as possible. It’s easy to become reliant on investor capital, but one of the things that helped Hungryroot was my experience with my second business, which was highly unprofitable. That experience made me paranoid, so I prioritized making the business profitable—or at least ensuring strong unit economics—early on. That approach has paid off, especially as the market has shifted over the past few years. Many companies in our category have exited or shut down, primarily due to their inability to achieve profitability.