- Notifiqation
- Posts
- How a venture studio is tackling the challenges of manufacturing companies.
How a venture studio is tackling the challenges of manufacturing companies.
OSS Ventures builds tech startups that develop digital products and solutions for manufacturing companies.
In our today’s heroes of series I discuss with Ashok Azhagarasan, Head of Growth at OSS Ventures, a venture studio out of France at the intersection of technology and manufacturing. In four years, it created 24 companies that deployed their digital products in more than 1200 factories, positioning itself as a thought and doer leader in Europe, recognized for being at the forefront of everything related to technology in tackling the challenges and problems of manufacturing companies by building and investing in startups.
Ashok, thank you for taking the time and agreeing to this interview. How are things going on your end? What are you currently working on?
Ashok: We're in the final stages of closing our second fund. Our dual approach continues to thrive; we're consistently launching four to five startups annually while making approximately ten investments per year.
With the new fund expected to close soon, we anticipate commencing the investment phase by Q4 this year. We've successfully launched four AI-focused startups, and we're seeing strong demand in this sector. We have just launched in the US market.
Nice, but when investing, how do you make sure you're not backing a startup that's a competitor to one of the companies you already built?
Ashok: Our evaluation process begins with the very first call, where we quickly assess whether there's an overlap with existing startups in our portfolio. Having launched 24 startups to date, we've developed a finely-tuned sense for market needs, current landscapes, and whether a new venture has true game-changing potential.
What sets our approach apart is our hands-on industry knowledge. We personally conduct weekly factory visits, which provides us with valuable "boots on the ground" intelligence. This direct engagement is feasible and scalable precisely because of our highly verticalized focus.
Even when evaluating startups in familiar sectors, we remain open to fresh approaches. We carefully consider how emerging technologies like AI agents might enable innovative solutions to persistent problems. This perspective also informs our investment strategy.
For pre-seed companies, our support extends far beyond capital. We offer comprehensive operational assistance, leveraging our network for strategic introductions, providing technical and product development expertise, and guiding go-to-market strategies. This hands-on involvement significantly increases their chances of success.
What inspired the creation of OSS Ventures, and how has this studio evolved since its inception?
Ashok: Our studio was founded five years ago by Renan Devillieres, who brings unique dual expertise to the table. With a decade of experience in the manufacturing sector, including roles as a factory director and supply chain director, he was essentially the right hand to a CEO in the luxury industry. His career then evolved through McKinsey before he launched an AI startup in Brazil, which was eventually acquired by a tech giant.
This distinctive blend of manufacturing and technology expertise shaped his vision. After his exit, Renan returned to France with a passion for manufacturing and the capital to invest in startups. What he discovered was troubling: pre-seed manufacturing startups faced significant challenges like limited access to capital, excessive focus on hardware at the expense of effective sales strategies, and a fundamental disconnect between traditional manufacturers and innovative startups during the sales cycle.
The most critical issue was market access. As Renan observed, "When you're a multinational industrial company, entrusting your production operations to a startup feels enormously risky; production stoppages can cost millions per day. The question becomes: can I trust this startup to be part of my core business?"
Finding no suitable startups to invest in France, Renan decided to establish his own studio. His approach was methodical and in the first year alone, we conducted approximately 100 factory visits across multiple sectors to identify pain points that could be addressed through technology.
These deep-dive visits involved comprehensive discussions with 7-8 people on each factory floor. By day's end, we consistently identified at least 10 pain points solvable through digital solutions. We discovered that manufacturing companies with 300+ engineers often had 300+ Excel spreadsheets, each engineer creating their own solution rather than implementing scalable approaches.
One particular challenge we identified was workforce management for blue-collar workers. While white-collar employees have structured career paths through platforms like Workday and clear career advancement opportunities, blue-collar workers often with 35 years on the factory floor lacked defined skill recognition and career progression. With the average age of European blue-collar workers being 55, manufacturers faced an urgent need for knowledge transfer and succession planning, especially in the luxury sector where experience and craftsmanship are paramount.
To address this, we created Mercateam, a solution focused on workforce planning & skills transfer enabled by AI. For instance, in France, where four-week summer holidays are standard, we found factories where only four senior employees knew how to operate certain machines. Our solution helps create knowledge transfer plans, ensuring operational continuity during absences and systematic skills transition before retirements.
Our approach focuses on three pillars: First, verticalized solutions targeting specific industry pain points that we then scale across multiple sectors. Second, horizontal assets like our no-code platform that connects to any machine. And third, sustainability initiatives including energy management solutions and waste management marketplaces. These core areas represent opportunities where AI can create significant additional value, both at the individual business and group levels.
How do you identify the operational problems in manufacturing companies?
Ashok: We go to the factory floor and identify pain points depending on use-cases, and ask the factory directors to vote on prioritizing the use-cases. If we can solve the top three problems, it will improve certain KPIs. Then we try to solve those specific problems which have the highest impact on EBITDA.
We identify what the main irritants are on a regular basis for a person working on the shop floor. Is it a system problem? Is it a people problem? Is it just change management that needs to be done? That's how we try to identify issues, and we co-build the solution with three companies in different sectors.
We then have an iterative period of around three to four months where we build the proof of concept. At the end of that period, we ask, "Do we provide enough value or not? If not, then we try to iterate." That's how we do it, value-based and proof-based.
Interesting. You also do a lot of field work, which is quite important.
Ashok: We deliberately provide these insights without charging for them. Unlike the traditional consulting model, which creates a transactional relationship where clients hesitate to reach out knowing they'll be billed, our approach fosters genuine intimacy and ongoing dialogue.
By keeping these services complimentary, we've successfully cultivated a community dynamic. Every quarter, we proactively contact our industrial partners to identify their three most pressing challenges. We explore whether these have evolved and how we might address them whether as investment opportunities or potential new ventures within our studio.
If we don't see an existing solution that fits, we evaluate whether there's potential to launch a new initiative in that sector. However, we remain pragmatic in our assessment. One key factor we consider is whether an American company has already secured significant funding, say $100 million to solve the same problem. If such a competitor exists and is likely to enter the European market, we factor that into our decision-making process.
Our methodology is fundamentally about being realistic and strategic, understanding the competitive landscape and ensuring we're addressing the right problems at the appropriate scale.
What methodologies do you use to scale these innovations effectively?
Ashok: One of the main things is that we have the same tech stack for pretty much all of our portfolio. It took us 10 months and 2 full-time tech people to launch the base tech stack, which is compatible with traditional systems like SAP.
It also passes cybersecurity requirements, which gives us an edge for launching MVPs within a few weeks rather than six months.
Second, we have UX design, so we try to do at least 16 interviews before launching a solution. We interview people on the shop floor, mid-level managers, and C-level executives. We have a baseline understanding of what the C-level is looking at in terms of business unit value. For a manager on the shop floor, we understand daily usage patterns and how they're going to use the solution. We can then build an experience based on that. We also spend a lot of time on the shop floor to understand the key areas and the returns needed to really fix the problem at the right level.
What are the main challenges you encounter when integrating these solutions in traditional manufacturing environments?
Ashok: The primary challenge we encounter is human resistance to change. When you approach someone who has performed their job the same way for two decades and introduce a new solution, their first reaction is often fear and concern that this tool will replace them. But that misunderstands our objective.
We're not looking to eliminate positions; we aim to automate routine tasks that don't require human judgment, freeing these experienced professionals to spend more time on the shop floor solving complex problems that truly demand their expertise. This is why change management is absolutely critical to our approach.
We've learned to be selective, choosing to work with only the top 20% of companies that demonstrate sufficient digital maturity to embrace transformation. Our benchmarking has revealed that roughly 40% of manufacturing companies simply aren't ready; their organizational culture isn't aligned with digital adoption. Implementing even the best SaaS solution in these environments is futile because people aren’t prepared for the change.
Instead, we focus on companies that have already established the necessary infrastructure, those with functioning data lakes and systems in place as they're inherently more adaptable to innovation. In some cases, we candidly recommend that companies first engage with traditional consulting firms to manage organizational change and potentially restructure their teams before attempting to work with startups.
This pragmatic approach protects both parties: the manufacturer avoids a failed implementation, and the startup isn't burdened with a 24-month deployment that should take only three months. These insights have been invaluable in refining our investment and incubation strategy.
Can you share one or two success stories where your portfolio companies transformed operations in factories or manufacturing companies overall?
Ashok: When we launched a startup four years ago called Fabriq to digitalize lean operations, we co-built the solution with an aerospace manufacturer. At that time, their processes were not digital and were not very effective in terms of lean planning because everything was on paper.
When we deployed the solution, within three months, they gained 10% productivity, and there was a direct impact on EBITDA. The solution is now deployed across the entire group at 45 factories, and they achieve about 20% gains per year.
The main game changer was that they were reinventing the wheel at every factory. When you have the lean process, you have standard daily meetings with a whiteboard and notes that's classic in every shop. Every time there's a problem, you put a note saying, "I need to solve this problem." Once the problem was solved, they put the note in the bin, and there was no knowledge gained.
What we did was help them map all these problems, and when a problem was solved, we documented how it was resolved. We did this at multiple factories and then at group level. Now they have an understanding of the main key problems that are occurring and how they can train people to prevent these problems altogether, rather than solve them each time.
We also added a layer of AI so it can identify what the recurring problems are and how they can be resolved even before the problem happens. That was the game changer at the group level.
They can identify if it's a cultural issue for example, is a business unit in France having the same problem as one in Germany? Is it a people issue? How can they put processes in place so that they can have gains across the board at group level rather than at a specific business unit?
It helped both on the shop floor and at the group level.
Speaking of AI, how did you adapt when it started to become more omnipresent? From the portfolio companies that you created, how many of them already had some AI capabilities, and how many didn't? Did you have to quickly implement something? Was it a challenge for you to do that?
Ashok: AI has become central to our innovation strategy. Our most recent ventures launched within the past year are all AI-native by design, with our latest startup leveraging AI agents representing the next frontier in our development approach.
For our established portfolio companies those operating for four years or more we're implementing AI as an enhancement layer that dramatically expands their capabilities. Take Fabriq, for instance, which manages comprehensive problem-solving ticketing systems for manufacturing operations across major enterprises. With vast datasets accumulated over years, they've now introduced what they call the "knowledge feature" an AI layer that delivers transformative results.
When a new issue ticket is created, the system now automatically identifies similar problems that were previously resolved in other facilities. It might determine, "This exact issue was solved at another factory six months ago," and immediately present the implemented solution. This capability fundamentally changes how knowledge propagates throughout manufacturing organizations.
Our talent management platform demonstrates similar AI-driven evolution. The system can now forecast that a specific employee will retire within the next two years, identify their five core competencies, analyze which skills have already been transferred to other team members, and identify which remain untaught. It then generates a comprehensive training roadmap for the coming year to ensure complete knowledge transfer before retirement.
The platform also predicts talent shortages in specific domains six months in advance, giving management ample time to prepare. Effectively, we're providing manufacturers with dedicated data science capabilities focused on their most critical challenge acquiring, developing, and retaining the right talent at the right time.
This targeted approach delivers specialized analytical power at scale, addressing one of manufacturing's most persistent challenges: maintaining an appropriately skilled workforce engaged in meaningful work.
Do you think that now all the solutions that you want to build have to be AI native?
Ashok: The market in general that we see, especially in the US, is that if your solution is not AI native and there are no AI agents, then there are about 10 other solutions in the market, and you have zero chance to win. Tomorrow, if a new company launches which is backed by Y Combinator, they will probably do a fundraising of $30 million because they are AI native, so they have a bigger chance to win the game.
Even though you have three years' advantage, they will probably wipe you out. That's what we're seeing, especially if you're fundraising with top-tier US companies. You need to be AI native; otherwise, it's not going to work due to the amount of competition.
When you build companies, do you focus more on going through different fundraising rounds, or do you focus more on building a sustainable and profitable business model from early stages?
Ashok: For us, the startup process is around 9 to 12 months to launch a startup. At the end of the process, we want the company to be profitable, to have at least three clients, and to be at around 500k signed ARR.
The objective is to raise from a VC for the seed round to replace the internal team. Our objective for them is to get them to $10 million ARR and EBITDA positive. Once you get there, then you can decide whether you want to take the private equity path or if you want to continue fundraising. But we want to build sustainable, profitable businesses first, and not do the VC game of throwing $5 million and expanding at a crazy rate when you're not profitable at all. It's not a good deal for the founders, and it's not a good deal for us either, because you're taking a huge risk when another AI native company comes into the game and you're blown out of the water.
As a venture studio, is your strategy to hold and grow, or grow the company to a certain level and then exit?
Ashok: Right now we want to invest more so that we can go even further, take them to year seven or eight, and then you can show outsized returns.
So we want to hold and grow since we have a portfolio which is very complementary in terms of our platform. It opens up a lot of possibilities.
Which areas do you think will see the most disruption in the manufacturing industries?
Ashok: Automotive is being disrupted like crazy. Luxury will be disrupted. The type of production and how you're going to produce is going to be completely different in the next 10 years.
But if we talk about the operations inside the manufacturing company, which parts do you think will be the most disrupted?
Ashok: Supply chain is already being disrupted. Shop floor operations are going to be the next disruption. Right now, it's already live at the top 5% of companies.
If you go to Tesla at the Fremont factory, you see the shop floor operator who is more of a developer trying to problem-solve and has 15 tablets around them trying to manage issues. That's the future that we are seeing operations headed toward, so you have multiple systems and you have multiple robots where you're taking care of operations, and you will need people with brains to actually fine-tune and be the orchestrator of the platform rather than doing manual operations like lifting. That's a huge area of disruption.
Another thing is that the way we produce is going to be completely disrupted, specifically with reshoring and with the current inflation, especially in Europe. They need to change or they will be out of competition.
Speaking of people, how do you select the founders and entrepreneurs that will lead the companies you create? Because not everyone has a background in manufacturing.
Ashok: It's pretty simple. We try not to get people from the manufacturing industry. We try to get people who already have some experience running a tech business or had a small exit and want to build something even bigger the second time.
So we try to look for second-time founders, and we try to look at people who already worked in a tech company, because we're building a SaaS tech company, so we don't need manufacturing expertise.
Ideally, it's better if you don't have manufacturing experience because otherwise you have 20 years of a single vision of how it's going to be. It's better for someone to have a fresh background while understanding very well the problem that we need to solve, because right now they're not solving it at the right level and there's a better way to do it.
Having unique ideas and a fresh point of view is probably better for us to build a rational company.
What advice would you give to manufacturing companies that are still hesitant about technology implementation and change?
Ashok: Right now, what we see in the market is that lots of people want to change, but they don't have the right level of expertise, or they are working with consultants who have done three or four projects and now have these scars and want to get out of that.
So it's better to start at a smaller scale because the main thing is they have a very engineering mindset; it's like, "Okay, I'm going to get this company or this SaaS and I'm going to do a group-level rollout," although the company is not really ready for it.
Try to do it at a smaller scale first, do a proof of concept, test and learn, and once you build that muscle, then it's easier to scale into the company because you already show wins. So if you deploy a solution in one country and you show gains, like, "Okay, this has business impact and it works in our facility," it's very easy to get everyone aligned across multiple business units. That's the first thing; start small and do iterations. You're probably going to fail, but at the second or third iteration, you will know how to work and you will know how to iterate as well.
Those are the key things because there's no one-size-fits-all. Every manufacturing issue is very specific and there's a lot of parameters that need to be adapted.
That's the thing that they need to get used to. And with the changing ecosystem with AI coming into place, probably a lot of stuff like admin, finance, and supply chain negotiation procurement is going to be disrupted, so be ready for that level of disruption and try to build your organization with that in mind.
When you validate the problem, how do you make sure that the problems you validate are not very different from one company to another so that you're able to build a solution that will solve the problems of at least three, four, or five manufacturing companies that you work with at the moment of the discovery phase?
Ashok: Initially, we try to do user research. We speak to a minimum of 60 to 100 people across the board. Internally, we have a database with around 14,000 conversations that we've had in the last four and a half years across every sector, every type of person, and every level shop floor, mid-level manager, as well as C-level. We already have that data.
There are at least companies in aeronautics or in luxury that we already spoke with three years ago, but three years ago, we didn't have AI agents, or the tech wasn't ready, or graph databases the tech was not there yet to solve this problem. Right now, the tech is ready. Maybe we can solve the problem at the right level, and we can disrupt that sector with the current instruments.
We go and show the Figma screens and prototypes to the users as well as the C-level to understand if it is something they're willing to pay for and if it can solve their problem. We try to narrow in on very specific sectors, and then keep in mind that we can expand the solution to multiple sectors from day one.
So we don't build something very specific just for the aeronautics sector or the luxury sector. From day one, we already try to see whether it can be agnostic across multiple sectors, and whether this company can be a hundred-million-dollar business in the future or not.
One last question: what advice would you give to startups to be revenue generating from day one and not depend on VC money?
Ashok: The co-building process—before, we used to do it for free. Right now, we ask the co-builder, the design partner, to pay from day one. So then they have financial commitment and they're much more serious about the project because if it's free, they're more likely to postpone the discussions.
We also ask for a time commitment of one full-time resource giving feedback and having at least one meeting every week so we can iterate fast. Right now, we've deployed in about 2,400 factories. So we know the stickiness is there once you deploy the software. It's just getting the problem and solving the problem at the right level; that's the main thing that we obsess over in terms of product.
So when co-building with manufacturing companies, they actually also pay for this?
Ashok: Yes. Initially, we did it for free, and then after six months when we asked them to pay, they're like, "Oh, it was a great project, but right now I'm not willing to pay."
I can give you a clear example of a project we closed. One startup was solving after-sales in the manufacturing sector, which is usually five people using email and forms to solve service issues and nothing of this is automated, nothing is systematized, and all the mental load is on people. We try to automate the process and have a centralized CRM-like system for after-sales.
The problem was there across multiple sectors, etc. We built the solution out, and when we came to the company, they weren't willing to pay because the people that they have there are minimum wage, and they were thinking, "Why would we pay 2,000 euros per month for something when we can just employ one more person, and it's not a big deal?"
So we need to test that at the right level on the shop level where there's a need, as well as at the C-level where there's willingness to pay. What are you building, and do they pay from the start or after a couple of months? They pay as soon as they have a product which is deployed on the shop floor.
Here we come to the end of our discussion with Ashok Azhagarasan from OSS Ventures. His insights helped us understand the challenges of manufacturing companies and how their venture studio is solving these challenges through the startups and companies they launch.
Thanks a lot Ashok for your time insights!
Follow Ashok Azhagarasan
Follow OSS Ventures
Subscribe to our newsletter https://notifiqation.beehiiv.com/subscribe
Follow us for latest updates and posts about upcoming articles on
https://www.linkedin.com/company/notifiqation/
https://x.com/Notifiqation_