Safetytech Logo

Enhancing Workplace Safety and Business Performance: AI and Computer Vision in MSD Prevention

 

Maurizio Pilu (00:00)

Good day from London. I’m Maurizio Pilu and I’m the Managing Director of Safetytech Accelerator. I would like to welcome you to this new episode of our Insight Series. Today, we are going to talk about muscular skeletal disorders in the workplace. MSD represents one of the most significant workplace health and safety challenges. It affects millions of workers globally. It affects the productivity of businesses and also whole economies.

 

Some people do estimate that as much as 5 % of global GDP may be affected by MSD, although the estimates vary considerably. In the USA, 30 % of all accidents at work and work -related illnesses are due to MSD. And in Europe, as much as 60 % of workers at some point of their lives have been affected by MSD, causing huge amounts of productivity losses.

 

and health care expenses each year. So with a problem so big, the question is, what is the role of innovative technologies in mitigating these risks? And what could and should businesses and regulators do to increase their use in the workplace? Today, we are speaking to Diwakar Ganesan, co -founder of Tumeke, an entrepreneur and a technology company that are at the forefront of this mission.

 

developing cutting -edge tech solutions to combat MSD. We had the pleasure working with Tumeke as part of our collaboration with the National Safety Council in the USA during the setup and running of the MSD Solutions Lab, which is funded by Amazon Inc. Diwakar, welcome and real pleasure to have you here today.

 

Diwakar (01:52)

Likewise Maurizio it’s great to be on the show.

 

Maurizio Pilu (01:56)

Fantastic. Thank you again. So can we start with your story? If you can tell us a little bit about you, your background, but also I think we are really interested in the story behind you co -founding Tumeke.

 

Diwakar (02:13)

Yeah, of course. Yeah. So I’m Diwakar. I’m one of the co -founders of Tumeke. What we do is make software to reduce and hopefully in the long -term eliminate musculoskeletal injuries. So I’m the technical co -founder on the team, which means my job in the early days involved building the initial version of the product and working with our early customers. But now it more transitions into managing our engineering team and setting direction for what we want to build next. My background is in artificial intelligence and computer vision.

 

But after learning more about the scope of this problem, we partnered with Dr. Alan Hedge, who’s a professor of industrial engineering from Cornell. And from him, we learned a lot about ergonomics, learned a lot about biomechanics, and his expertise helped us get the product off the ground in the early days.

 

Maurizio Pilu (03:00)

And I wanted to ask you when was the light bulb moment where you said, I want to fund this company.

 

Diwakar (03:05)

Yeah.

 

Yeah, so we started off by working with Ovo Fund in Palo Alto and we did a fellowship with that firm and we were speaking to different people in manufacturing and in insurance and then safety more broadly because in our own lives we had kind of seen how debilitating a musculoskeletal injury can be and we also knew that there was pretty significant advancements in computer vision and in machine learning that meant that we could now very

 

cost effectively understand human motion, which is like one of the biggest driving causes of musculoskeletal injuries is human motion not being done optimally or in an unsafe way. So from that initial learning with that fellowship, we developed our minimum viable product, the first version of our software that was able to automate ergonomic evaluations. We started with the RULA and Rebut evals because that’s what we saw was used most commonly in industry.

 

For me, the aha moment was first seeing how much time we were saving by using computer vision and performing these assessments. People who were previously spending 45 minutes to an hour to complete a full RULA or Reba eval were now spending less than like five minutes, which is a huge improvement. But the other light bulb moment was realizing what the, I guess, emergent properties of being able to do these assessments quickly was.

 

And the most important one was that now, because there was not as much of a barrier to doing the evals, customers were performing them a lot more frequently than they used to. And they were using them for things that previously they’d never intend, they’d never been doing. So we saw customers that were using our assessments for training purposes, because it was so easy to get an assessment through our platform. You could now show someone really quickly what they were doing wrong with the job. And it became part of a training loop and bringing the worker themselves into the safety equation.

 

and getting them excited to improve.

 

Maurizio Pilu (05:05)

That’s really amazing. So for you, the light bulb moment was essentially observing people doing this assessment and you said, haha, can do it better with tech. That’s really interesting. We have something in common. I also have a PhD in computer vision many, many years ago now. And it’s amazing to see how technology progressed since the time I did my PhD. And I think, I guess you will tell us in a moment.

 

Diwakar (05:14)

Exactly.

 

Maurizio Pilu (05:34)

probably 30 years ago this would have been difficult to do with the technology available at that time, certainly not on a mobile phone, because they didn’t exist. So I think the convergence of tech and the capability of the computer vision made these sort of things happen. But we will cover that in a moment. What I wanted to ask you now is if you can step back a little bit and tell us about MSD as a problem in the world.

 

So how big is the problem? And what are you hearing from your clients? I mean, you’re probably going to come to this speaking to other clients you’re presenting. What is the market telling you about this problem? And maybe even the role of tech.

 

Diwakar (06:24)

Well, there’s a couple of different ways of looking at that question. I guess first we could start by just looking at the numbers. So in America, based on information from the National Council on Compensation Insurance, we know that the average cost of a musculoskeletal injury for employers is about $36 ,000. So that’s for each time someone presents with a sprain and strain. And the aggregate cost across the US economy for just like handling the

 

medical and legal fees associated with these types of injuries is north of $60 billion annually. So this is definitely a very substantial problem. And musculoskeletal injuries on average cause about 14 days away from work. So that’s kind of the quantitative aspect, but there’s also this very human element, which is like we’ve all kind of experienced these types of injuries like sprains and strains. And I think where it gets, where it gets,

 

particularly painful is when you consider what the life cycle of care is for an injury like this. So yeah, we have the numbers, but what they don’t fully capture is the fact that let’s say you present with some shoulder pain because of some injury that happened in the workplace. So you might start with some physical therapy, it might not work. And then suppose you have to go into surgery to like ultimately fix that problem. If these types of surgeries have risk associated with them and

 

There’s people that we have spoken to that have spent the last 10 years in and out of surgery and physical therapy because of what started out as a minor shoulder problem caused by a lack of training or poor form when performing some kind of activity. And to me, what really pains me is just how avoidable all of this is. We have information technology now that would let us teach people quickly the right way of doing these types of movements.

 

Maurizio Pilu (08:01)

Thank you.

 

Diwakar (08:10)

we very much have the ability to go after these types of injuries.

 

Maurizio Pilu (08:14)

Thank you. And just speaking about technology, doing computer visual -based assessment is only one of the technologies that can be used for MSD. But maybe, could you give us a little bit of a picture of what other technologies are out there? And we heard about exoskeletons and so on. I mean, I guess some of them are for prevention, and your case is for prevention and maybe even training.

 

Diwakar (08:21)

Mm -hmm.

 

Yeah.

 

Maurizio Pilu (08:39)

And so could you give us a little landscape of the various solutions that are out there to tackle this challenge?

 

Diwakar (08:47)

Yeah, there’s a variety of technology that’s out there and like with anything new, it’s going to come, everything is going to come with its own sets of pros and cons. So we’ve already mentioned markerless motion capture technology, which is what we do. So understanding human movement without the need for wearables. And there’s also exoskeleton solutions and there’s wearable technology companies as well that…

 

It’s like minimally invasive things that you put on your body that measure biomechanical characteristics and then give feedback. And then there’s also, I guess, fixed camera tracking solutions that maybe are not as granular when it comes to ergonomics, but like loop into existing camera systems and provide continuous feedback on general safety related incidents.

 

Maurizio Pilu (09:36)

Our portfolio of technologies that are being used, I guess, in different use cases and also for different type of interventions. For our listeners, we have developed jointly with the National Safety Council a roadmap of MSD technology. So we will add the link to the notes that will accompany this interview.

 

So why, well, I guess it’s the perfect timing now to talk about Tumeke the product, You are one of the prominent emerging players in this space. And congratulations on the $10 million round in October. Yeah, welcome. Congratulations. And so can you bring your product?

 

Diwakar (10:12)

Mm -hmm. Thank you. Yep.

 

Maurizio Pilu (10:24)

to life for our listeners. What does it do in practice and what benefits does it bring to your client? So when a client is using it, you mentioned already the beginning, the faster assessment time, but perhaps can you elaborate a little bit more? But first of all, tell us exactly what the product does.

 

Diwakar (10:45)

Yeah. So there’s two parts to our system. There’s a mobile app and a web platform. The mobile app is meant to help collect ergonomics information and videos of job cycles. So if you were a user of our mobile product, what you’d see is an interface that makes it easy for you to film job cycles in your facility and then upload them to our backend for processing. And the web platform serves as a place where you can do analysis on those videos. So if you want it to understand

 

trends in ergonomic safety information. If you wanted to understand what interventions were being implemented to address the risks that our system has called out, then you would use the web platform. I guess one other thing to mention is that we really are agnostic to the source of the video. The mobile app is meant to be the most convenient way of getting the information. But if you have pre -existing footage that you’ve collected before in other surveys that you’ve done,

 

If you have footage from IP cameras that are installed, GoPros, we can kind of consume it all and do ergonomic evaluations on those. But I guess mobile app is meant to be the most convenient if you’re on site and you need to do an evaluation. I think there was another part of the question on what value the customer gets from using the tool. So I see really three distinct sources of value. One is time savings, which we’ve already touched on.

 

Maurizio Pilu (12:01)

Hmm. Yes.

 

Diwakar (12:09)

If you’ve been mandated to do large scale safety surveys of ergonomics, that can be very time consuming given the size of the facility. And we cut down on that pretty substantially. The other thing is injury reductions. Of course, as we do these evaluations and show individuals the risk results, train them on processes, even if we’re not implementing, like we’ve seen, even if we’re not implementing engineering controls, there’s pretty substantial reduction in injury rates. And maybe at a later point,

 

get into specifically what that looks like. And the third source of value is a little bit more intangible, but it’s the improved relations and conversations between safety specialists and workers themselves, where our technology kind of brings workers into the fold, where they can see why the best practices are what they are. And this makes everybody more excited to work in a safer way and work in a way that isn’t going to cause long -term harm.

 

Maurizio Pilu (13:08)

That’s very interesting, the third point you mentioned, the one that we could call continuous improvement. I think it’s a recurrent theme of these technologies. So they give a direct benefit. So functionally, you do one job or several jobs, and you get value directly. But it changes the culture completely. Dynamics of a

 

of managing the workforce changes. And it almost doesn’t matter what type of technology we see that effect. You called it earlier emerging properties. I really love that. It’s a very techie term for this, but it’s a beautiful thing that maybe we will expand in another podcast. Because they are material and also they’re long lasting. In other words, it’s quite difficult to go back once you have these emerging properties, once the type of way of working is

 

Diwakar (13:40)

Mm -hmm.

 

Exactly.

 

Maurizio Pilu (13:59)

is embedded, it’s difficult to go back and it’s great to see. You mentioned that in terms of time saved into the assessment, obviously you see that when it’s a large organization. Am I correct to say that your typical clients are large, medium to large, or it varies? So what’s your typical client at the moment?

 

Diwakar (14:24)

It really varies. So our customers run the gamut from small manufacturers that might only have 100 employees to global companies with a footprint all around the world. Really the common thread that ties them together is the fact that they have people doing some kind of manual work. As long as that’s the case, there’s going to be a very strong reason to use our system.

 

Maurizio Pilu (14:46)

So manual work could be warehouses, could be manufacturing, transport, logistics, et cetera. So it’s quite, it’s a fairly big chunk of the economy, actually, if you add it all up.

 

Diwakar (14:50)

Mm -hmm.

 

Exactly, exactly.

 

Exactly, exactly. And we actually recently finished a deployment with a company in the airline industry, a very prominent airline in America. And they’re using our system to train flight attendants on proper technique of doing various movements in aircraft. So this is a use case for our system that kind of goes beyond like typical manufacturing industrial where you’d expect to see ergonomics.

 

Maurizio Pilu (15:20)

that’s impressive. That’s really impressive. And having all of us flown many times, they do have a difficult job. Flight attendants in very narrow spaces. Thank you. So let’s maybe focus on the client, on the barrier now. So.

 

you have the solution. And on the other hand, there is somebody that needs to make that decision to buy the product. And this is a very common thing that comes up in the very many discussions we have with entrepreneurs. One of the common questions is how do I sell more and faster? I wish, you know, as an entrepreneur, but that would be a difficult question to answer in general, I suppose. But going back to your particular company,

 

If you put yourself in the shoes of a typical buyer, so the person decides to check in one of these companies.

 

What are the typical triggers that make them say, yeah, now I want to bite? What moves them across the line? In the future, it might be different, right? There might be peer pressure to bite. At the moment in time, it’s this level of adoption. What are the two, three scenarios where they say, I’m going to give a call to Tumeke and I’m going to do this?

 

Diwakar (16:43)

Yep. That’s a really good question. And it’s of course, something we spend a lot of time thinking about internally. I think ROI is incredibly important to the buyer as it should be when it comes to buying enterprise solutions like ours. So that, I think that purchase decision happens when they are able to see for themselves how much time that this tool will save them and see some early indications of how a system like ours

 

will help reduce injuries over time. And we find that usually takes about a month or two for the organization to like do their own trials, see how the system can be implemented in their own situation, and then put together the data to make that purchase decision.

 

Maurizio Pilu (17:27)

So you see, typically, clients will want to trial the technology in their context, maybe small scale, I don’t know, and then they do the full purchase. But even the trial, at some point, when they say, I want to trial this, it’s a big step to say, I want to trial the technology. So if we go back to that moment where they say, we want to trial it,

 

Diwakar (17:43)

Yeah. Yeah.

 

Maurizio Pilu (17:51)

Have they seen it at a conference? Have they looked you up online? So what this is actually something that comes up all the time in safety tech, the core safety tech, is that it differently from pure productivity solutions and so on, where you buy, you save. There is a very, very compelling need to buy. And also the buyers are different, like to be operations people and so forth. On this one, there is a slightly different

 

if you like, motivation to buy. And we’re always quite curious to know. So the typical person that is moving to a trial with you, how did they decide to do it? If you can give us two or three examples, maybe even one.

 

Diwakar (18:37)

Yeah. Well, I think one is that the product is very visually charismatic. So when you see a demo, you’re able to very quickly understand what it can do for you. Everybody speaks the language of red, yellow, green. And when you see that skeleton change colors in real time, that really clicks, especially if you’re in the biomechanics safety space. The other thing is that a lot of times our trials get kicked off when

 

folks are tasked with doing ergonomic safety surveys and then they see how to me he could save them a lot of time and improve the objectivity of having done such a survey.

 

Maurizio Pilu (19:19)

So they said we like, suppose you have, I don’t know, 5000 employees, you’re gonna do another round of surveys to say, well, it’s time we do something in a different way. Let’s use technology. And that’s where they say, let’s try this. That’s fascinating. I couldn’t agree more with the visual impact, your website and collateral. So you have, you know, you kind of, you see it and you get it. You get it, what it’s doing. And it’s, it’s, it’s not a small detail actually, because

 

Diwakar (19:28)

Exactly.

 

Yeah. Yeah.

 

Maurizio Pilu (19:47)

You almost have an emotional connection to the product, right? If you are one of those specialists that have to do these jobs.

 

Diwakar (19:54)

Yes, yes, yeah, because you see how much time it’s going to save and you see what it can do for the people you’re tasked with keeping safe.

 

Maurizio Pilu (19:56)

Yeah, yeah, yeah.

 

That’s interesting. The power of images, right? As they say. That’s really fascinating. Can we expand a little bit on the ROI? OK, so it’s coming up all the time. And it’s.

 

Diwakar (20:05)

Mm -hmm.

 

Maurizio Pilu (20:18)

is probably the main driver really. But sometimes it’s chicken and egg, right? Until you try it in your organization, you don’t know what is your ROI. Hence, this trial that people do and so forth. But I wanted to ask you about

 

things actually about ROI. The reason why I pose is that it’s actually very complicated. The first question is the productivity increase element, right? So you mentioned this continuous improvement and so on, but it might be odd. So for example, for exoskeletons, they actually have seen people being able to carry more boxes faster and heavier, right? So it’s like

 

you’re becoming a little bit of a cyborg with an exoskeleton. That’s direct and immediate improvement. And then there is less injuries and so forth. So I’ve seen some cases of exoskeletons where people have said, OK, we can shift stuff much faster. Now, in your case, you mentioned the ergonomic survey savings, which must be quite significant. But when else have you seen?

 

measurable productivity increases by using your technology.

 

Diwakar (21:35)

Well, this goes back to a conversation we were having earlier about what the impact of being able to do more of these evals is. So what we find is we reduce the time it takes to train new employees now. Whereas before, you would have to go and create training materials for safety and maybe have people go through an in -person safety training session. Now it’s as easy as filming them and showing them where they might be deviating from best practices.

 

So that wasn’t really possible before our system. So we’re saving time and the time it takes to onboard new employees. The other places that we see ROI are of course reducing injuries and reducing the risk of the different jobs that are happening at facility.

 

Maurizio Pilu (22:21)

So it’s a number of things. You add them up and then it’s a big increase. You have a… sorry. Please.

 

Diwakar (22:27)

Exactly.

 

I was going to say maybe to speak a little bit more specifically about the outcomes some of our customers have seen. So as an example, M3 insurance saw a 50 % reduction in the overall time it took to do an ergonomic safety process by using our system. So that not only includes doing the assessment, which was much faster, but also coming up with solutions, guiding people through the implementation of those solutions, and then making sure that the solutions.

 

actually reduced the risk and it was being followed through it. So that entire cycle, they saw a 50 % reduction in the time it took. And we’ve also observed reductions in the risk of injury associated with jobs. So Hitachi S. Demo, for example, saw a 57 % reduction in the MSU risk of some of the workstations that they had redesigned using our tool. And of course, finally, there’s like the reducing injuries themselves. And we’ve been working with, we had been working with a large

 

distribution center that saw a 68 % reduction in MSD injuries from using our tool for training and onboarding over a six month time period.

 

Maurizio Pilu (23:39)

That’s really impressive. And now that you have given us these impressive numbers, let me switch to some of the barriers, maybe. There are always barriers to buying technology solutions. They always exist. But what I wonder is that the evidence of the benefits of these technologies is starting to grow.

 

case studies are being published and there’s more evidence. The health and safety professionals are starting to talk about the benefits. So what’s holding people back from buying generally in the economy on a sector? I assume that the level of penetration is still quite early, this technology in general, right? It’s not like everybody’s using them, which means there’s a lot of potential and that’s why maybe you’re so attractive to investors.

 

But in your experience, what’s holding them back in the face of these evidence? Maybe the top three reasons for not to move ahead yet with a technology such as this.

 

Diwakar (24:48)

Yeah.

 

Yeah, I think the biggest reason is timing where the company has internally set goals to improve safety outcomes, but they’ve slated it to start like six months down the line. So it doesn’t make sense to implement a solution like ours just yet because their goals for improving ergonomics, improving safety are maybe one or two quarters down the line. So timing is an important thing. I think another restricting factor is of course budget and whether or not

 

the company has already allocated the money for a solution like this. Like what we find is that because this is a new technology, right? This is a new source of value. Historically companies have not budgeted for something like this. So what we find is that sometimes our buyer has to wait for a budgeting cycle to begin so that a new item can be introduced into their budgeting system.

 

Maurizio Pilu (25:49)

so because the budget is not allocated or it’s allocated elsewhere, sometimes just simply not available and one has to wait for the cycle. Also interesting, very fascinating, the problem of timing. So what you said earlier about one of the main value being when they’re running this big ergonomic assessment. So you have almost a cycle.

 

Diwakar (25:59)

Mm -hmm.

 

Maurizio Pilu (26:15)

for each company, right? I don’t know whether it’s annual or six monthly, but that’s very interesting. And in terms of a culture, the culture of the buyer, so maybe they’re a bit risk averse in terms of adopting new technology. Have you also noticed some more structural, if you like, maybe lack of awareness of the potential and so on out there?

 

Diwakar (26:16)

Mm -hmm. Mm -hmm.

 

We definitely saw that in our early days when markerless motion capture was just getting started, but it’s been a couple of years now. So people, by the time they speak to us, have definitely been made aware of the potential of this kind of technology. So we don’t really see that sort of averseness anymore.

 

Maurizio Pilu (27:00)

So they’re aware of the potential, now it’s down to practicalities to decide to, that’s very interesting. And then so the first hurdle is people say, I think I might use that in the future. And then it’s about finding the right time. That’s great. I wanted to zoom out slightly from computer vision solutions for MSD assessment and maybe get your view on where else.

 

Diwakar (27:04)

Exactly. Yeah.

 

Maurizio Pilu (27:30)

in occupational health and safety, technology can make a big difference outside of MSD, that is. Have you seen your clients talking about holder technology solution, other use cases, what’s hot in the market? Maybe if you can share a couple of points on that, it would be very interesting to hear your perspective.

 

Diwakar (27:53)

Absolutely. We hear our customers talking a lot about safety as it relates to lockout, tagout, slips, trips and falls, chemical safety, electrical safety. So these are all like still top of mind for our customers and the stakeholders that we work with. I think the difference is that the leading cause is still MSDs as a bucket. And it’s the one that is the hardest to go after because the causes of these injuries are more subtle and less direct.

 

And that’s why I think like our system makes the most sense as like an emerging technology because it’s going after the, what we see as being the biggest problem.

 

Maurizio Pilu (28:33)

Thank you for that. Well, you almost went by the, if you like, by the probability of occurrence of incidents. So MSD first slip, trip and falls falling from height. And that’s actually how many professionals think they just go after the biggest risk and the biggest impact of that risk.

 

So talking about these classes of technology, we call them safety tech, but we have been analyzing the level of investment going into these classes of technologies. And we published a report about a year ago, which again, we might add into the notes to this interview. But I wanted to ask you whether you see the future being bright for these classes of technologies. I’m sure you will say yes, but.

 

Maybe you can elaborate more. And in particular, I’m interested on your perspective on the level of investment. Essentially, VCs, well, they do get something spectacularly wrong sometimes. But generally speaking, they follow the opportunities, right, future opportunities. Do you see increased activity around safety tech from investors?

 

Diwakar (29:45)

Absolutely. I’m incredibly optimistic about the potential that technology has to help people have healthy work lives and work to their fullest potential, like not just in computer vision, but also with all the advancements in large language models like GPT and different techniques like that. It’s going to be more and more possible to coach people at scale on safety best practices and how to do their jobs more effectively.

 

And on the VC side, there’s definitely growing interest in the built world, in the physical world, in manufacturing. And safety is a huge part of that. And I only see that going up over time, especially in America.

 

Maurizio Pilu (30:26)

Thank you. I mean, we are landing on a high, I guess on a high in terms of future outlook. One final question. If you had one recommendation each for clients and entrepreneurs in this space, what would that be?

 

Diwakar (30:52)

That’s a good question. I think for clients, my recommendation would be to continue to follow the trends in technology and continue to be at the forefront of adopting best practices because at the end of the day, it’s all about making sure people stay safe. And technology has been a huge part of that.

 

For entrepreneurs, I’d say it would be to listen carefully to customer needs and what the buyer is asking for, because sometimes it can be counterintuitive and it might not be what you expect.

 

Maurizio Pilu (31:31)

Brilliant. Thank you very much. Fantastic. We could talk for much longer on this topic, but we need to wrap up now. Diwakar thank you. Thank you very much for joining us today and for a very stimulating discussion. All the best to you and Tumeke

 

Diwakar (31:50)

Thank you very much. Yeah, thanks for having me on.

 

Maurizio Pilu (31:52)

Fantastic. Thank you. And thanks everyone for watching this episode of our Insight Series. If you would like to discuss this topic further or if you have suggestions for further topics in the future, please do get in touch with us. This is Maurizio Pilu from London. Goodbye.

Join Our Webinar
Innovating Fire Prevention in Maritime: Advancements in Detection and Risk Management