• Founders Project
  • Posts
  • Revolutionizing Maritime Operations: Greywing Journey to Optimize Vessel Efficiency with AI

Revolutionizing Maritime Operations: Greywing Journey to Optimize Vessel Efficiency with AI

Hrishi Olickel @ Greywing (YC W'21)

Interview with Hrishi Olickel, the founder of grey-wing.com. Greywing is a Y Combinator startup from the W21 batch, specializing in AI for B2B services in the maritime, aerospace, transportation, and logistics industries.

Greywing raised $2.62M in total funding from Flexport, Transmedia Capital, Signal Ventures, Motion Ventures, Rebel Ventures, Entrepreneur First, Y Combinator.

The Screenshot of the Greywing startup profile in YC database

Screenshot from YC presentation of Greywing startup

Read Full Interview Here:

Who are you and what your startup is?

My name's Hrishi, and I'm the CTO of Greywing. Having worked in robotics, defense, and insurance, I have a strong foundation in AI and deep tech.

At the company that I'm in now, Greywing, we focus on vessel operations. For example, large ship operations like container vessels, product tankers, and oil tankers. We apply the right amount of machine intelligence and data processing to help streamline their operations and save costs. That's the gist of what we do.

Can you share any KPIs or metrics?

Sure, so we work with over 500 ships at the moment. We've got a pipeline for about 2,000 to 3,000 ships. We've managed tens of thousands of crew members that have gone through the platform, sort of thousands of flights that have been booked on the platform. We've helped vessels route over a million kilometers in terms of where they need to go, and what decisions need to be made. Those are really the metrics that we measure ourselves by.

Do you or your co-founder have personal expertise or a relevant background in the problem your startup is solving?

I would say we have that from both sides. From my perspective, I knew nothing about shipping before I got in, and from his perspective, he knew everything that you could know about shipping, because he'd had about 12 to 14 years of experience at that point, having done all of the jobs from the bottom up. When we first got into shipping, my background was largely in data, in working with industries that are not early adopters of technology. That came in really handy with shipping because it's not that different from insurance or other parts of different industries. So there was quite a bit of founder-market fit. In ‘traditional’ industries like shipping, you cannot survive without founder-market fit.

How did you come up with the idea of Greywing?

It depends on which one you're talking about. If you're talking about what we do now, we got it from what we call hitting the market continuously, just hitting the market. What we started out with is very, very far away from what we do today. We started out by looking at security on board ships, and that was very close to what my co-founder and were proficient in, and so it was a merge of our interests. That worked relatively well for about 6 months until COVID-19 happened, and then the bottom fell out of that market.

Then, we focused entirely on the biggest problem with shipping, which was crewing. We built software and tools for crewing instead of moving crew on and off the ships. And we were quite successful. But as COVID went away and the problem started changing, we started adding more tools for managing crew travel, for routing instead of fuel optimization.

I would say that we didn't come up with what we do today. Most of it was customer-driven. Most of it was success-driven. We pointed in the direction that works.

We validated that problem by building small versions of what we wanted to do. I think we've always been a technology-first company. And what that means is demo first. We end up building the best we can, putting it in front of people, and then asking our users, "Do you want to use it?" That initial use and that initial spike of users will tell you enough about whether something needs to exist and be built. Even today, that's what we do when we have an idea, or when we want to build something, but we don't necessarily know how to validate it. The best way we've found is to build the thinnest version of that, put it out there, and then, once it goes out there, you'll get an understanding of what your users want.

What did your MVP look like?

It was essentially a map, just a map with ships on it. And when you clicked on the ships, it would tell you the information we knew about the ships.

The screenshot of Greywing MVP

Screenshot from Greywings MVP

And that was partly because we spoke to a few customers who said, "Look, we don't have a way to know where our ships are at any point in time. You have to go to all these different sites." So, just presenting that information in a better way was good for them. We also picked the map because we wanted to make sure that the product was also impressive to look at, not just to use. Because, you know, a startup founder is always selling to two different marketplaces, right? First, you're selling to customers. But you're also selling to investors.

Time is very limited when you're an early-stage company. So we wanted to build something that would do both those things: work for investors, but also work for customers.

Did you organize any customer interviews?

Yes, customer interviews are still very much just how we do things. We make sure we have as many as we can per week. In the beginning, it was largely trying to figure out what part of the organization we would fit into. So, we set up customer interviews with people that are all the way close to the ships, and all the way up at the top, like upper management and everything else. I think when it comes to customers, it's useful to have the distinction. And for us, in the early days, that distinction was "What is the distance between the buyer and the user?"

When you sell B2C, when you just sell directly to customers, your buyer is the user. When you buy a Spotify subscription, you're the buyer and the user. But when it comes to large enterprises, your buyer is almost never the user.

So, almost no customer interview matters unless you get the same signal from both the buyer and the user. For us, the first step was to identify who those were inside the companies that we were talking to, and then make sure that almost anything that was said was validated by both sides.

Did you leverage your networks in the early days?

We very much leveraged my cofounder’s network for the first couple of months. But I think, post that point, especially because it's a very distributed world, content marketing and inbound have been the drivers of our sales.

Now, we call it content marketing. But at the time, we mostly wanted to write as much as we could. We didn't necessarily write for keyword optimization or SEO, we basically just wrote as much as we could on LinkedIn, on our blog, and everywhere else. Who we were, what we were trying to build, and where we see this going. The good thing about putting that out there is, you will find people that resonate with you. Even now, 5 years into building this company, someone will find an article from 4 years ago, and be like, "Well, I agree with that." That's been really helpful.

You didn’t optimize for keywords?

Right. I invest in companies, and I've seen enough companies do the opposite of what we did. I also have quite a few friends who run companies. One thing that happens if you do it too early is you become a keyword-driven company.

A screenshot of the Greywing blog

Screenshot of Greywings website

You optimize your marketing for keywords, and then you get responses because you optimized for keywords. Then, your idea and your company start moving towards the keyword, which can be problematic. We didn't optimize that early. Once we knew what the product was, there were enough customers and our revenue was growing, we started optimizing for keywords. We started optimizing when we were going from 1 to 10, not when we growing from 0 to 1.

Do you have your own version of doing things that don’t scale at Greywing?

Oh, 100%. We do a few things that scale, but that's about 20%. The other 80% are the things that don't scale, where, you know, we simply did those things and said, "We will scale when we need to," and it still hasn't happened. Even today, most of our customers are served from a technical perspective, from a single server that we pay about a hundred dollars a month.

At the beginning, we said, "Look, let's build the simplest stack we can of product. And the day that that looks like it's getting a little overwhelmed, we'll scale." That still hasn't happened, right? And we started going out to customers. We started building specific software for customers, specific UIs, and things for customers because that doesn't scale, but even today, those ones are still the ones that people end up using.

We have several large clients. That's about 20 to 30% of the market or large clients, and there's quite a bit of revenue there. But 80% are really small clients. It's a long tail of shipping that goes down.

But we focus on the large players because that's a good place to set the pace, sort of get the data that you need as a machine intelligence company to sort of figure out what you need to do.

How long did it take you to acquire your first paying users?

On the early product, the first thing that we launched? I think we got paying users almost immediately because at that time it was a marketplace. It was more about connecting buyers and sellers, and the demand was high. We found a customer for what we were offering within a couple of weeks.

Now, after we changed, and sort of moved further into crewing, I think it was about a month, month and a half before we found our first paying customer. It was really just for us about finding early adopters. We found our second customer in about another month or month and a half.

I want to say, we had an idea. But I don't think we did. We had an idea about what problem we wanted to solve. The early users were mostly buying their seriousness about the problem and their trust in our ability to solve it.

I think that's where building that MVP very quickly made a difference because the problems in crewing just started happening during COVID-19. We were the only company that had any solution that was built even partly right, in a month.

Mostly, our users came to us and they went, "Look, this is gonna solve some problems. It's not gonna solve the big ones yet, but we're happy to sign you and sort of bring you on because you've gotten this far in a month. We can believe that you can solve this in two months." They were buying the solution that we were going to build.

How did feedback drive your product?

We've continually added more standards to this process over time as we've identified patterns. There have been a few adjustments made in different areas. Generally, the majority of the company attended customer meetings in the early days to gather diverse viewpoints. When four of us participated in a customer meeting and discussed their problems and feedback, we often heard different perspectives, all of which could be valid. We documented all of this information and primarily managed it through user requests, which could range from significant to minor issues. These requests were categorized accordingly.

We sought validation from multiple customers because sometimes the problems were temporary and would surface sporadically. After the initial MVP was developed, most of the product's features and improvements came from customer feedback regarding their issues with the product or their broader life problems, rather than specific solutions.

Why did you decide to apply to YC?

Well, we encountered some challenges with fundraising in the early days. Actually, it's a bit complicated because we didn't face much trouble initially. We received many term sheets. However, our difficulty arose when an investor signed a term sheet, went through three months of due diligence, and then abruptly withdrew on the last day, without providing any reason. This left us feeling rejected. So, when we approached funding again, we wanted to ensure that we were properly prepared. YC played a significant role in this regard, and it was also because both of us admired figures like Paul Graham, Sam Altman, and Michael Seibel, who were associated with YC. We applied because we wanted to be part of that community and to seek advice and learning opportunities.

Moreover, as a B2B enterprise company operating in the shipping industry, having the endorsement of YC has been highly beneficial for attracting venture backing.

Did you get in on your first try? 

Actually, we didn't. It's quite common for companies to succeed on their second attempt. For us, the second attempt was easier because it served as a signal. In our first application, we made promises about what we intended to do, but in the second attempt, we actually executed on those promises. This made it clear that there was a viable path forward and that we followed through on our commitments.

There might be a misconception about the stage at which YC typically invests. They usually come in after a company has already achieved some validation, and that initial growth surge is often considered the sweet spot.

What changed after YC?

There are a few answers there. From a personal perspective, and I emphasize this because it's crucial but often overlooked: therapy.

Many of us, especially founders, especially in the B2B space, go through this challenging journey. It can be difficult for people to connect or share what's happening. In some cases, all it takes is someone to tell you, "Hey, you're actually overreacting," or "You're on the right track."

Having that supportive community has made a significant difference. I say this from experience, having observed numerous companies. Initially, we were part of Entrepreneur First, and we witnessed many companies emerging from Singapore. Surprisingly, the primary reason most companies from our batch didn't survive wasn't financial issues or a lack of revenue; it was co-founder mismatches or personal problems stemming from running a startup. YC genuinely helped by providing a large community of individuals at various stages of building a company. Some were generating $50 million in recurring revenue, while others were just starting with their first user. This diversity and support network have been incredibly valuable.

On a professional level, I'd like to highlight two things. First, regarding fundraising, YC has the most extensive database and one of the most effective fundraising support systems I've ever encountered.

Secondly, it's the expertise they offer because they address various significant challenges you encounter along the way. Growing a company from $0 to $100, from $100 to a million, and from a million to $10 million all present different sets of problems. Thanks to YC's extensive network, we could connect with companies that had experience in B2B SaaS within industries like trucking and shipping. This allowed us to receive specific, targeted advice, which was immensely helpful. The YC partners are exceptional, and the people who run YC provide invaluable guidance.

That definitely boosts your confidence. It's also beneficial to have respected individuals occasionally give you straightforward startup advice when needed. Sometimes, founders can become stuck in a loop and may not recognize what they should be doing.

Have you found the PMF?

I don't believe you ever truly discover product-market fit. However, you do observe indicators of it. Yeah, I think that's what you pursue, and most of the time, you notice signs of product-market fit. And I believe we're getting closer if nothing else. For us, these signs of product-market fit manifest as shorter sales cycles. It's almost like you can skip some meetings. Previously, you'd have meetings to discuss the problem, the solution, contract negotiations, and iron out details. Maybe now, you can bypass all of that and jump straight to discussing pricing.

You initiate the meeting, and by the end of it, they're inquiring about the cost, perhaps because they're interested in making a purchase. You also see signs of user adoption. People begin discussing it elsewhere, and they come to you about it. Achieving product-market fit in the realm of enterprise B2B is always somewhat complex, especially when dealing with significant order sizes. There might be periods of inactivity, and then suddenly, seven-figure revenues appear from a single source. So, as much as possible, I think there are certainly indications of product-market fit.

Exactly. You never really know because a multitude of factors come into play. The buying company might have faced certain circumstances. Their financial situation could have changed. The internal champion who supported your product might have left, which has happened to us several times, or the opposite could have occurred. It's challenging to discern.

Along your journey, was there a thing or a hack that delivered far more user growth than you expected?

That's a great question. There are actually quite a few things that we didn't anticipate. We didn't expect it to have the impact it did. One of them, I think, pertains to the product, which involved breaking up the product into separate control surfaces instead of consolidating it in one place. Our product encompasses flight planning and various other aspects scattered all over the place. They all seamlessly work together in the same process.

However, breaking it up and providing users with different entry points made it much easier for some users to focus on what they wanted to do. If you think about any product in your life where you enter with one intention, but the product bombards you with ten other things you don't care about, it is somewhat similar to that.

Another aspect was our effort to share more about our information and what we do instead of developing everything in secret. We've been open-sourcing certain components that had a significant impact, and I believe it enhanced our credibility.

As a startup, you don't have many advantages in a large industry. One of your key advantages is your ability to move swiftly, deliver products quickly, and pivot rapidly. Your agility sets you apart. So, the more you can leverage these strengths, the better.

Many startups we've interviewed have experienced moments of resourcefulness, like the famous Airbnb Obama O's cereal story. Did you have any similar experiences?

For us, I don't think there was a single defining moment like that. To be fair, we did experiment with building various products in the early days as part of the discovery process. We attempted to repurpose certain aspects of our tools and operational expertise to see where they could be of assistance. During the COVID-19 pandemic, we engaged in various activities, such as running initiatives here in Singapore and in other countries to help stranded seafarers return home, in collaboration with crewing unions. We also worked with travel agents and other industry bodies that were not part of our customer base, partly for visibility, partly to be helpful, and partly to explore new product possibilities. We even developed products for government use to see where that might lead. So, while we had several initiatives, none of them were as remarkable as the Airbnb story.

I've seen that you've launched the SeaGPT recently. Congrats on that. What's your next big milestone or goal for Greywing?

I believe the next significant change is in shipping, both as an industry as a whole and in our specific areas of crewing and chartering. Not enough attention is being given to it. Perhaps because of us, more people are looking into it now. Shipping, as an industry, heavily relies on human language and unstructured data, which includes emails, SMS, and WhatsApp. This is the foundation of the entire shipping industry. Operational knowledge, including deals worth seven or eight figures, is conducted through these means and often remains unstructured.

Our current goal is to take the areas we operate in and structure all of this data. We process hundreds of emails, and our next significant milestone is to process millions of emails within the next 3 to 6 months. We aim to automate conversations that don't require human involvement, extract structured data from these messages, and make a significant portion of ship operations as autonomous as possible. This was challenging until recently, and I believe large language models were the missing piece needed to make it feasible. That's our next focus.

Would you say that opportunities in shipping are currently overlooked by founders?

Absolutely, I think so. This sentiment has been echoed by nearly every startup in our space. The opportunities are genuinely overlooked. People often see a daunting domain of things they don't understand and think they need to master it all to make a difference in the industry.

However, I believe that the shipping industry is becoming more accessible. It's an essential industry, and we now have a team of engineers, most of whom have no prior shipping experience. Within just a few weeks to a month, they've been able to start building incredibly useful and innovative solutions within the shipping sector. But I think many people, like myself, are intimidated by the seemingly impenetrable "domain wall" of not knowing what goes on inside the industry, which hides the valuable opportunities within.

How has being a founder changed you personally?

I think it's made me more resilient as a human being. We all have our comfort zones, but startups essentially force you out of them. You might feel comfortable for a short period after raising money, although it's not advisable to get too comfortable during that time. However, beyond those initial two months post-funding, you're mostly venturing into the unknown. So, being a founder has made me more resilient. When something challenging happens, I've learned to say, "That's alright, I can handle it," and move forward.

If you were to reflect on your journey, is there anything you would do differently now with the knowledge you've gained?

I believe I would take more risks. I say this as someone who has taken significant risks along the way, and I'm saying it with the benefit of hindsight and accumulated knowledge. But taking more risks would be a key change. Additionally, I would build thinner products. Typically, when people start, they aim to provide as much value as possible to the customer by creating comprehensive products. However, if I were to start again, I would focus on building leaner products that are easier to integrate into a customer's life or the lives of the people we serve.

Learn More About the Founders:

A photo of Nick Clarke and Hrishi Olickel

Nick Clarke and Hrishi Olickel

Former Royal Marine Commando, bootstrapped entrepreneur. Now turning data into Crew Changes in the maritime industry.

Ex-researcher, with papers in applied cryptography, formal verification, human-computer interaction, genetic algorithms and learning, continuous glucose monitoring and baseline estimation for qPCR and X-Ray crystallography. Some work in robotics and optical flow modeling, as well as 3 years in Insurtech and reinsurance in Southeast Asia, then a few years in fintech in Singapore.

Hrishi’s Personal Blog: olickel.com

If you enjoyed the interview, consider supporting Founders Project:

Join the conversation

or to participate.