November 19, 2021
Benj Cohen, founder and CEO of Proton, recently joined Rob May, host of the “Investing in AI” podcast, to talk about how artificial intelligence (AI) affects the distribution industry. In this episode, Benj shared that Proton was founded to help the distribution industry with its biggest challenge: data.
Distributors sell many products to a large volume of customers, which creates unique challenges. For example, a sales rep can’t be an expert on each of the thousands of SKUs typically sold by a distributor. Additionally, distributors’ customers purchase from multiple channels. The activity and data from each channel are typically siloed and managed from disparate systems. Benj combined his knowledge of AI and familiarity with the distribution industry to create an AI solution that helps distributors address these problems.
Distributors are unsung heroes in the American economy and have a tremendous impact nationally and globally. One-third of the U.S. GDP passes through a distributor’s warehouse at some point in the supply chain. In total, distributors accumulate over $6 trillion in volume per year, even more globally. And although distributors have always had an important role in keeping our lives running, they’ve received more recognition over the past few years due to the pandemic.
Distributors get supplies to where they need to be. For example, when the pandemic began, that meant getting masks, gloves, and other PPE equipment to hospitals and doctors.
Benj and Rob dove into how AI can improve the process of distribution and help make this critical component of society run more smoothly and efficiently. Here are just a few highlights from the conversation.
Rob: What’s the history of Proton? How did the company begin?
Benj: I grew up in my family’s distribution business. We sell dental products to dentists across the U.S. Growing up, I traveled with my dad when he would visit customers. I watched sales reps and learned how the warehouse worked. I even worked in the business for a bit. That’s where I became familiar with the problems and challenges that distributors face.
In college, I got interested in AI. It occurred to me that one of the biggest challenges distributors face from a sales perspective is that they sell a lot of products to a lot of different customers. For example, my family business sells 400,000 different products. At Proton, our average customer sells 687,000 different SKUs to tens of thousands of customers. Basically, it’s really hard for reps to figure out what customers need and for customers to find what they’re looking for.
It occurred to me that all of this is really just a data problem. It’s a large number of products and customers and transactions – high volume. This is the kind of optimizing that AI is good at.
Rob: How do distributors benefit from this large amount of data? What’s so important or useful about it?
Benj: When Proton started, our first customer was my family’s business, and we started with maybe a hundred million transactions right out of the gate. There were 400,000 products and 50,000 customers. Having this much data, which all distributors have a massive amount of, is a huge advantage for building an AI system. Pulling from this history, we’re able to enable sales reps with product recommendations for repeat customers as well as opportunities for up-selling and cross-selling right away.
Rob: What was one of the biggest issues you have faced when launching your company?
Benj: In the beginning, we were too focused on the technology and not focused enough on the delivery – how do you get a salesperson to actually act on one of the system’s suggestions. When we showed my family the earliest version of the system, we basically got laughed out of the room because we hadn’t figured out how to plug the software into their workflow. The data was there, and the models worked. We even increased their revenue per customer by five times, but we overlooked the most important part, which is where the model meets the person. Once we got that right, then the product really started to work well.
Rob: How do you deal with this user adoption issue? When you plug your AI into their workflows, as you mentioned, how do you get the sales reps to act on your system’s suggestions?
Benj: First, a little bit of context on the setup: We can plug into other interfaces, but we try not to. We’re really focused on owning the desktop experience. The reason for that is reps today in distribution are working across many fragmented tools, so there’s value in pulling all of that information together and putting an AI layer on top. We simplify things with AI. We take their workflow and pull their Microsoft BI insights as well as our AI insights and put them into a single place.
For example, one of the use cases that we work with is field sales reps. Before we help them, they might have a CRM on their phone and an ERP system they’re pulling up on their laptop. They might also have spreadsheets full of different customers’ information. They probably have an order entry system on their laptop. So, they’re looking at all this disparate data before they go visit a customer.
We plug into that workflow by pulling all the important information about their accounts into one application that exists on their phone. Then we layer on top of that AI to help the rep understand where they need to go with that customer. Rather than digging through spreadsheets, the ERP system, and CRM to figure out what the customer last purchased, AI simplifies that by saying, “Hey, based on all we know about this customer, here are the next couple of opportunities with this account.” These could be wallet share opportunities or a reorder opportunity. For desktop sales reps, one of the key things we figured out was that they loved being able to easily copy the recommendations that our software made and paste it into the emails they were sending to customers.
So, these product recommendations, and reps seeing how the technology can simplify their lives while also helping them sell more, aids the user adoption and user experience. People might be hesitant at first about changes and new technology, but we’re not hiding the AI. We’re using AI to simplify their existing workflows.
Rob: How do you approach the culture-building at your company, specifically for folks who don’t understand distribution or see the value that distributors provide?
Benj: Distributors are not a business that many people in high tech understand. The way we frame it as a company is in terms of the opportunity to transform this massive part of the market. Yet, there aren’t many people focused on this industry from a technology perspective, so we have this amazing opportunity to make a dent in this part of the world and this part of the economy, which I find really exciting. At Proton, we’re focused on getting everyone in the company – not just the sales reps – to spend time one-on-one with distributors. Employees visit with distributors. They go on-site and sit next to sales reps. They travel. Every two weeks, they spend at least 30 minutes talking to an end-user of our product. It’s important to give people that context and empathy for this industry that often lacks exposure. And the response is generally surprise. “Wow, I did not understand how complicated this business was.” What I’ve found is that people have an “aha” moment when they sit with reps and start to understand how distribution works. They gain a lot of respect for the industry, which I think is awesome because it’s the industry I grew up in, and I have so much respect for these companies. It’s hard work, and it’s not as simple as it may look.
Rob: Where is distribution headed with the help of technology? How will AI impact this industry?
Benj: There is so much opportunity in distribution to leverage AI simply because of the volume of data about their customers and their products. They also have a lot of data about how products move through the warehouse and what the most efficient way is to get certain products to certain customers.
We’re focused right now on enabling the salesperson. How do we help the sales rep better understand what their customer needs so that they can better serve their customer and, ultimately, sell more at a higher margin to the customer? The solution is building an omnichannel experience, where we help the reps get smarter. We also use AI to adjust their ecommerce site, so the customer has a more personalized experience while browsing and/or shopping.
The next step is connecting that system with inventory management. Today, inventory is managed with a forecasting model or process of guessing future needs based on prior volumes. It’s extremely complicated. But this problem could be solved using the data that we’re already looking at. The future of AI in distribution is merging sales and purchasing together to not only enable the sales force but to also forecast inventory. When you marry these two systems, you gain efficiency. Once you can automate pieces of the sales model, we’ll be able to connect that with what’s happening in inventory.
Rob: Is there anything in this industry that investors should keep an eye on? Or anything that you’re hoping for the future in terms of technology to make Proton run more smoothly?
Benj: One big area, and one that I know already has a lot of investment going toward it, is integrations, specifically making it easier for us to integrate with legacy systems. A lot of our customers are using ERP systems that are 10-20 years old. Integrating with them can be challenging at times. We’re certainly getting better and much faster than we were when we began, but I’m hoping that the process can become more efficient with new technology in the future.
Rob: It sounds like all of this technology will really reinvigorate the distribution industry.
Benj: Yes, that's why I'm so excited about what we’re doing. Most people don’t understand what distributors do or see the value in their contribution to the supply chain. But, the secret is this is a really hard business, and these companies are run expertly, yet no one has thought to build technology for them. So now there’s a huge opportunity for distributors to jump ahead simply because they haven’t had the opportunity to previously. People have been building sales software for companies that focus on sales forever, but no one has built software to help a distributor sell more. That’s a huge roadblock for this market, which will be transformational for the distributors willing to jump on board.
Listen to the entire podcast here.