Joe Sullivan, host of The Manufacturing Executive Podcast invites Tutor CEO Joshua Gruenstein to be a guest on this weeks episode of the show. Gruenstein and Sullivan discuss how machine learning and robotics can be leveraged to address manufacturing environments with high variability that have previously been difficult to automate. You can listen to the episode on Spotify, Apple Podcasts, or read the audio transcript below.
[0:04]
Joe Sullivan: Welcome to the Manufacturing Executive Podcast, where we explore the strategies and experiences that are driving mid sized manufacturers forward. Here you'll discover new insights from passionate manufacturing leaders who have compelling stories to share about their successes and struggles.
[0:20]
JS: And you'll learn from B2B sales and marketing experts about how to apply actionable business development strategies inside your business. Let's get into the show. Welcome to another episode of the Manufacturing Executive Podcast.
[0:37]
JS: Welcome to another episode of the Manufacturing Executive Podcast. I'm Joe Sullivan, your host and a co-founder of the industrial marketing agency Gorilla 76, where we help B2B manufacturers grow through revenue focused marketing programs. Before we get into today's show, I'd like to take a moment to announce an incredible event our team at Gorilla 76 will be co hosting alongside True Marketing for the second consecutive year this coming February of 2025 in Austin, TX. It's called the Industrial Marketing Summit, and it's just for marketers in the manufacturing sector. Last year, our inaugural event drew over 200 manufacturing marketers who showed up in person to learn, network and bring back smarter, more effective marketing strategies to their organizations.
[1:20]
JS: This coming February, it's going to be bigger and better. You'll see plenty of familiar manufacturing leaders on stage, as well as a wide variety of some of the sharpest minds in the industrial marketing space. Topics for this year's sessions include content marketing strategy for B2B manufacturers, using data for reporting on marketing results, synchronizing marketing efforts between manufacturers, distributors, and sales reps, how generative AI is changing the marketing playbook, using LinkedIn to build brand awareness, planning for a website overhaul that drives business results, employer branding to recruit the next generation, and much more.
[2:01]
JS: Not only will this be an intensive learning event made-up of speaker sessions, panels, and two intensive workshops, but we'll be hosting social events in the evenings with great food and venues for networking with other manufacturing folks who are trying to solve the same kinds of marketing challenges that you are.
[2:17]
JS: The event is being held from February 26th through 28th at the AT&T Center on the University of Texas campus in Austin and seats are limited, so visit industrialmarketingsummit.com to learn more and reserve your ticket. We hope to see you or whoever is responsible for marketing inside of your organization there.
[2:37]
JS: Now let's get into the episode. For many years, robots were used to perform repetitive tasks in low variability environments and required on site automation engineers to program and run those systems. But we're now living in an era where machine learning and AI technology are making high mix, low volume environments ripe for automation.
[2:59]
JS: My guest today is an entrepreneur whose company is doing exactly that. Let me introduce him. Josh Gruenstein is Co founder and CEO of Tutor Intelligence and MIT spin out with the mission to put a robot in every factory. Prior to Tutor, Josh was a graduate researcher and lecturer at MIT where he received his undergraduate and graduate degrees in Computer Engineering and Artificial Intelligence.
[3:25]
JS: Josh, welcome to the show.
Josh Gruenstein: Pleasure to be here.
JS: Well, your bio was very modest. Sometimes people give me paragraphs and paragraphs of stuff to read and sometimes I get a few sentences. But tell me a little bit more about you and you know, what led you to start this company focused on intelligent robotics for manufacturing?
[3:44]
JG: Yeah, well, I think my background is pretty straightforward. I've been doing robots since I was about 8 years old. That's probably the thing to know about me-I am a robot kid at heart. I had a robot birthday party at 8 years old and I still do it for fun to this day.
JS: That's awesome.
JG: And the story of what I'm currently doing, of Tutor Intelligence, this robotics company, is when I was a graduate researcher at MIT, I was seeing all of these advances happening in the field of artificial intelligence.
[4:12]
JG: There are these incredible new models and technologies coming out where software can reason about the world and, and physical objects and images and text. And as a roboticist, the question I had is, “How can we take those advances and bring them into the physical world and make it so we can use this AI to let robots do more jobs in more places and sort of have more productive output in the world?”
[4:39]
JG: So that led me down this path of creating the start up Tutor Intelligence. And what we do is basically we build robots from the ground up with artificial intelligence to help manufacturers, 3 PLS, and co-packers automate basically with smarter robots that can do more jobs and handle more variability. That's basically me.
[4:58]
JS: Awesome. I love the robot birthday party at age 8. It's probably fun to look back on that and now see it as foreshadowing of your future. It's pretty cool.
JG: Not a ton changed
JS: That's great.
[5:14]
JS: So Josh, you know, one thing that you and I were talking about recently was this idea of, you know, there's high mix or high volume, low mix environments and then low volume, high mix manufacturing environments. And I think there's a perception out there about what manufacturing leaders may think is possible, which has, I think, started to change very quickly. And I'm just kind of curious to hear your perspective on that. What do you hear? What do people seem to perceive? What is the reality?
[5:45]
JG: Yeah, well, I'll say, I guess my own perspective as an outsider coming into the world of manufacturing. I expected manufacturing to look a certain way, right? If you watch how it's made or, or go look up videos on YouTube, you might think of the Tesla factory or whatever and you're building the same thing over and over again at very high volumes. And in these environments, you can use robotics and automation because you can justify, you know, a 10 year project for an ROI, you can build processes that are never going to change, so these facilities have been the facilities that can take advantage of automation.
And what's changing, I think, which is really exciting, is everybody else. Which is most of manufacturing. And that's really what surprised me. Most manufacturing doesn't look like that. The strength of most American manufacturers is that they can handle an enormous amount of variability and they can deliver that value to their customers. They can take an order that's different and they can fulfill that order and they can respond to a dynamic supply chain. And what I learned is that robotics and automation, seem not to be built for those types of facilities or wasn't built for those sorts of facilities.
Those facilities need automation and robotics that can deal with the level of variability that those facilities have to deal with, with the level of dynamic environments, different orders coming in from different customers, layouts changing. And that's I think been elusive on the technology side.
And all of these new advances in technology with AI, the really exciting thing in my opinion, it lets you start tackling those types of facilities, those types of challenges using automation. You can build robots that can, for example, read a build sheet for a pallet and build that pallet instead of needing an engineer to go custom program a specific routine to go build that pallet or or similar for a bunch of other different jobs.
So that's creating this, this sort of inversion where you're starting to see in the last few years a huge adoption of automation among mid size manufacturers, contract manufacturers, co-packers where they can start to use this technology for the first time.
[8:07]
JS: Yeah, it's pretty exciting. It's kind of like the perfect storm of the market really needing this and being ready for this based on labor issues and supply chain issues, and also just the explosion of exponential growth of what AI can do that's just started to occur over the last few years, I feel like. Can you give me some examples from within what you guys are doing at Tutor Intelligence? Take some of these concepts we're talking about here: more complex or situations where there's more variability that are being handled now by robots.
[8:45]
JG: Yeah. So I'll give examples from my own experience. One facility I think about a lot is a co-packer for a Fortune 500 packaged food brands and they're operating out of 1,000,000 square foot warehouse and handling a huge amount of volume. And this is a perfect situation where the demands of the customer are a ton of variability, right? If we want to, let's say do special pack this month to promote this particular product or marketing, besides we want to package something this way or Walmart needs, you know, a certain build or ALDI needs a certain build. Somebody's got to make that happen. And today, you know, you're seeing a transformation where these sorts of sites used to be entirely human operated. In a million square foot warehouse you would have just a ton of people taking things out of boxes, putting them into shippers, building pallets entirely manually because that's sort of the only way to get it done, right?
You can't go buy a robot that will handle that degree of variability. And now that's starting to change. You know, we work with this site and they can have a robot for example, that's packing a display shipper full of finished goods and the instructions on what exactly you're building that's going to come from marketing.
That's going to change. You're going to do changeover daily or multiple times daily. And you can have robots that can react to that and start to automate huge fractions of the work that you were previously doing. Similarly for other sorts of secondary packaging operations, these are really classically high changeover human intensive jobs that for the first time manufacturers are able to get efficiencies on by using automation.
[10:25]
JG: So that one stands out to me as like- I think there's an assumption that this new technology is like, wow, it's really fancy and it only works for people who already work with robots, and it's sort of exactly the opposite. I think a lot of the time facilities that have no robots today are in some sense the best setup for this next generation of technology.
10:49
JG: They can bring in this new type of robot system that can build a pallet or a packet shipper, and they can really take advantage of this technology for the first time.
[11:05]
JS: So what kind of programming knowledge is required in a situation like this?
Or is it, you know, I mean, my understanding is because I have one client in a very niche corner of manufacturing who has robots that are, it's like they describe it as the brain. It's learning with every motion it makes and everything it does and, and building this massive database of information. So it knows how to react next time it sees something like that or, or whatever. But like, in the situations that you see and that you guys work in, is there some level of programming required? And then where does machine learning kick in? Like put this in perspective for somebody who doesn't really understand the nuances of automation?
[11:42]
JG: Yeah. It's a really big shift. So traditionally the way that you would put in a robotic system is you need a robotic engineer, right? Who is going to go teach the appliance or or write a PLC program or, you know, there's a whole massive amount of work involved in in building these incredibly impressive but complicated systems.
[12:06]
JG: With AI, you can move to a world of no, there's no programming. The system is not going to run your program. It's instead you just tell it, what is the thing I want? Right. So we work in manufacturing and packaging. Usually that's like, what is the finished good going to look like? Like what's the pallet build going to look like? What's the shipper going to look like? You know, if I'm unloading a conveyor, show me the conveyor I'm going to unload and then the robot just figures that out. And that's like, I think alien and unexpected to a lot of people in this industry that that's now possible.
[12:39]
JG: And not only is it possible, it's a lot better. It means that you can use robots for the first time because let's say you have engineers on your site, they are so busy handling all of the operational complexity in your site, you probably don't want to hire more engineers. It's really hard to do. Automation engineers are really hard to come by today. It's a real challenge to adopt A robot if you know you're going to need to program it and spend a lot of time on it. But there are also a lot of sites that have no engineers and how are they going to adopt automation?
[13:10]
JG: I forget the statistics, but you know, there is like orders of magnitude more unfilled positions in manufacturing than there are automation engineers who can program robots specifically to do those processes that need to be done.
[13:27]
JG: So the only way we're going to get out of that situation is if you have robots that don't need to be programmed. That can receive instructions in the same formats you're already distributing your instructions. You know, if you're a contract manufacturer, you're getting a build sheet and specifications from your customer. With these new robot systems you can literally just upload those same specs and the robot will go do that job.
[13:49]
JS: It's pretty wild. It's just pretty cool to see how all of this is evolving.
JG: And I'll say it's not just for facilities which have no automation. I think there are a lot of facilities that they've been able to automate a lot of what they do. You know, maybe you can automate your core manufacturing, but for example, your pack out isn't automated because that's the part that's really changing. Let's say you're changing labels or, or whatever it is. That's much more flexible.
Again, you're able to plug in these new intelligence systems and automate fully these processes, which used to be really manual. I'm thinking of a contract manufacturer and CPG that we work with where they had mostly automated lines, but they were doing pack up by hand. They put in palletizers at the end of their lines and that's half a million dollars in savings a year, less than $100,000 of upfront spend. They didn't have to hire more engineers to do it. They didn't need to staff up. It was able to integrate into their existing workflows without engineering or further spend, and that's a little bit wild to me. It's crazy that that's possible.
[14:49]
JS: OK, we're going to take a quick break from the conversation here. Picture a world one year from now where the thousands or 10s of thousands of design engineers, plant managers, project managers, operations folks, CEOs, or whoever you need to reach and influence already knew who your company was, understood your value proposition, considered
In this world, your prospects would come to you with their guards down, already informed and genuinely excited to talk about how you could help them. And your sales team could reallocate the time they spend slinging product into facilitating deeper, more engaged conversations with right fit prospects who actually have buying intent.
[15:38]
JS: So how do you get from here to there? Well, I can assure you of this. It doesn't start by dabbling an SEO, or making your website look prettier or posting more social media updates. It starts with a strategic marketing plan rooted in the most important business outcomes that your manufacturing organization needs to achieve.
[15:59]
JS: I'm talking about outcomes like pipeline and revenue growth targets, or new market penetration or customer diversification. Start with the outcome, then reverse engineer the marketing program that will help you achieve it. For well over a decade, our team of industrial marketers at Gorilla 76 has been helping OEMs, machine builders, contract manufacturers, robotic systems integrators, and Industry 4 Point O technology companies strategize and implement revenue focused marketing programs.
[16:29]
JS: If you feel like your company could use some guidance, please visit gorilla76.com and consider requesting a strategy call. We'd love to have a conversation. That's gorilla like the animal 76.com. Now let's get back to the episode.
[16:45]
JS: Josh I've had a lot of conversations about automation on this show over the 4 1/2 years I've been running this podcast. A lot of them have come back around at some point in the conversation to the fact that we need robots because there aren't enough human beings to do the jobs to get done. There's a major push for re-shoring that's happening right now. There's with the new administration talk of tariffs and as a result, more domestic manufacturing happening. And then there's labor issues already going on. Have you found that the value proposition for automation lies more around labor cost reduction or filling jobs that nobody can find people to do? Or is it you know, some of both?
[17:34]
JG: Yeah, I think it's, it's holistic. It's a little bit of everything. It's funny you mentioned the onshoring because a lot of our customers are facilities that are doing work that's recently been reshored. And yeah, they have the exact challenge you're talking about, which is we're spinning up this operation, we're increasing our volume, where do we get the people to go do that? So certainly I think it's true that automation is a big part of that story. And that's part of how we wull get it solved. One thing that I think I disagree with maybe some people is I think that's not enough.
I think that there is a reason that we don't have 10s of millions of robots in factories across the US and it's because you want robots to improve your processes in every element. You want to drive operational savings, you want to fill headcount that you can't fill, you want to increase worker safety, you want to upskill your staff, and you want to run a generally smarter, more intelligent factory that's more digital and more intelligent. And I think until we have all of the above, it's more of a challenge.
[18:37]
JG: I think people underestimate how complex these facilities are, how complex the work they do is, and what the standard is to actually deliver systems that are operational improvements. So I, I do think in addition to filling headcount, you do also need to be able to deliver cost reductions. You also need to be able to improve worker safety, right. So those are definitely areas of emphasis for us.
JS: So a bigger story to tell.
JG: Yeah, definitely. And you know, I think it's really interesting to work with different types of facilities and interact with them and understand how robots are changing their business.
[19:15]
JG: You know, I was just in a facility where I was talking with some of the workers and previously they had to lift 40 LB cases all day long, right? And that is back breaking work. It's unsafe. It's physically just very strenuous.
[19:32]
JG: And it's hard to fill those positions for sure. And if you can go from, I'm lifting 40 LB boxes to I am collaborating with a robot and the robot does the hard work and I make sure that the line is functioning correctly, that's a win in everybody's book, right?
[19:48]
JG: The facility is generating operational savings, your workers are getting upskilled and exposed to better workplace conditions. It's a win holistically, not just in any one category.
JS: 100%, Josh, I've talked to a few companies recently that have sort of brought this idea to the surface of robots being offered in a similar way to how software may be offered or SaaS service, software as a service business. And my understanding is you guys are, you know, your model is like Robot as a service. Talk about what that means, how it's different from a traditional way of thinking about deploying robots.
JG: Yeah. No, it's a good question. I think there are a lot of different elements to it.
[20:36]
JG: There's certainly an accessibility side to robots as a service where if you're not putting down a massive capital expense, it's a lot easier to trial and adopt automation because you can get it in your facility very quickly. It can begin working and you don't have to go through capital processes or getting debt or whatever else it is in order to get these facilities on site. It's much lower risk.
[20:58]
JG: I think another element which is talked about less is that robots are hard and complicated. Thinking about something that one of our customers once said to me, having worked with robots for a lot of years, is “I wish I knew less about robots than I do.”
[21:16]
JG: And I think when a site adopts robotics, they don't want to become roboticists, right? They want a machine that's going to do their job and it's just going to work always from forever. And one element of a robot as a service engagement is that the vendor that's providing that robot as a service agreement, they're in charge of the robots, right? They're responsible.
If you have downtime or maintenance or any of these things which you know are going to happen in an industrial facility, it will be the vendor's problem. And especially as the technology gets better, as the software and AI enables these robots to be smarter, more intelligent, more productive, and the hardware gets better, you can be exposed to all of those improvements and you can get them.
[22:00]
JG: We deploy robots to our customers, and every robot in the fleet gets better. And that's partially because of the AI and the software of the robots are constantly learning and improving. But it's also because of the robots as a service model, because it's our job to maintain up time and deliver productivity, and that's what we get paid for.
[22:18]
JG: So we're really incentivized to make sure that robots are doing a good job, that they're able to work, because if the robots aren't working, we don't get paid. So I think there's this really beautiful alignment of incentives where under a robots as a service agreements, both the customer, the manufacturer and your vendor are really incentivized to drive productivity.
[22:43]
JS: Yeah, it makes a lot of sense. Other things that I hear just talking to so many manufacturing leaders is there's been, think over the last year or so, there's been some hesitation to, you know, make CapEx expenditures and to be able to have an option where you're turning a CapEx spend into an OpEx spend is, is appealing to some. For others, it's like so foreign, they can't really wrap their head around it. And there's a there's a mindset shift that I'm sure will continue to play out as this becomes more of a common thing.
[23:06]
JG: Well, I think it's also changing. I think one relatively recent change is it used to be that robots as a service meant a long term contract. And what that means is there's still this underlying reality of somebody financing a system and they're getting a debt on some terms. So you know, there's a lock in and especially with AI and you could have a general robot where you know, maybe the robot working in a 3PL doing some job is the same robot hardware that's working in a contract manufacturer and a Co Packer. These contracts don't need to necessarily work that way. You know, there are vendors that are starting to provide robots as a service agreements that don't lock you in for multiple years where really there is 0 risk to get involved. And in that world, I think the benefits over CapEx are obvious. There's no lock in, there's no major expenditure and there's no risk.
[24:05]
JS: Yeah. Well, that all makes a lot of sense, Josh, and really great conversation today for anybody who's listening now and, you know, they're interested in learning more about what you guys are doing at Tutor Intelligence or just this idea of robots as a service or are trying to figure out how to get started. Like where would you point them?
[24:25]
JS: Yeah, I'm a little bit conflicted, but I know my company, Tutor Intelligence, would love to talk to you. You're welcome to come to our website, tudorintelligence.com and request an intro meeting via our forum. We always love to meet new manufacturers, co-packers, 3PLS, who are thinking through how do they navigate this new technology, how do they begin their automation journey. And hopefully we can help you with robots. And even if we can't, we're always happy to provide advice, so we'd love to speak with you.
[25:00]
JS: Awesome. Well, Josh, really appreciate you taking the time to do this today. I think what you guys are doing is super interesting and there's obviously a lot of need for it out there. So it's going to be interesting to see how things unfold in the years ahead.
JG: Yeah, we're very excited.
JS: Awesome. Well, As for the rest of you, I hope to catch you on the next episode of The Manufacturing Executive.
[25:22]
You've been listening to the Manufacturing Executive podcast. To ensure that you never miss an episode, subscribe to the show in your favorite podcast player. If you'd like to learn more about industrial marketing and sales strategy, you'll find an ever expanding collection of articles, videos, guides, and tools specifically for B2B manufacturers at gorilla76.com/learn.
[25:45]
Thank you so much for listening. Until next time.