Consumer-focused deliveries have seen a revolution over the past few years, but change is coming more slowly to business-to-business logistics.
That’s the area Reefknot Investments is targeting with its $50 million fund, which was closed during the first quarter of last year. Reefknot is a Singapore-based 50:50 joint venture between state-owned investment firm Temasek Holdings and transport and logistics firm Kuehne + Nagel.
The fund is focusing on Series A and B investments in startups with transformative technology for supply chains and logistics.
“We bucket it into three different areas: AI and deep tech, logistics and digitization, and trade finance,” said Marc Dragon, managing director at Reefknot Investments. “Some business model elements are also considered. It might not be a core AI company; it might be a startup that’s looking to disintermediate a certain part of the industry, but it might use some AI to support its business model.”
So far, Reefknot has clocked one investment, in Prowler.io, a UK-based artificial intelligence company targeting improved forecasting for the sector, including truck scheduling and pallet collection. “Their core is AI and deep tech around probabilistic modelling and reinforcement learning,” said Dragon.
The firm is currently talking to a slew of startups. “We are particularly looking at those with differentiated technology, especially if it’s scalable, preferably across a region or a certain type of market, or globally,” said Dragon.
Edited excerpts of an interview:
What is Prowler’s business?
Their core is AI and deep tech around probabilistic modelling and reinforcement learning. The way the supply chain and logistics industry has approached forecasting and planning was relatively old-school antiquated. The first step is generally demand forecasting and then you plan what kind of inventory is required, how much you need to produce and how you need to move goods across the globe. But the moment the forecasting is off, it’s a bullwhip effect. It amplifies itself across the industry.
How Prowler does it is slightly different. They use probabilistic modelling, which uses much less data and comes up with a feasible result, typically giving improvements of 15 per cent onwards, which is very significant from an industry perspective. That’s an example of the kind of companies and startups we are looking for: Those that not only have potentially a huge return on investment, but also a transformational impact.
What type of technology do you consider transformational?
We bucket it into three different areas: AI and deep tech, logistics and digitization, and trade finance. Some business model elements are also considered. It might not be a core AI company; it might be a startup that’s looking to disintermediate a certain part of the industry, but it might use some AI to support its business model.
With new regulations coming in on our industry, such as IMO2020 in the maritime space, there are lots of opportunities for startups to play. Also, a lot of traditional organisations would want to build digital capabilities and partner with startups and companies to improve their digitization. Logistics digitization can be transformative all the way from shippers to freight forwarders to carriers as well.
Trade finance is an interesting area because it has been traditionally looked at as fintech. But with the digitization of the supply chain and logistics industry and the evolution of technology, there’s greater visibility of goods movement and trade flows; there’s greater certainty of information and therefore, greater trade finance opportunities will emerge. Those trade finance opportunities would need to leverage supply chain logistics information to create new business models and new capabilities. We see many firms in this space. There are some new emerging technologies, especially in the AI space, around credit risk assessment for suppliers.
Where do you see the pain points in the logistics industry? What’s still done with pen and paper?
Across the entire supply chain, there are still a lot of legacy pain points slowly being smoothened out. Everything from documents being in a pen-and-paper format to emails and the data not being captured or rationalised. Of course, there’s cross-border customs and not only the standardisation of data but how cross-border players work with each other, from those that orchestrate cross-border movements to those who move the goods and store the goods and agents on the customs side within each of the countries. How all the players work together and in an optimum manner is still yet to be realized.
How big a problem is it that processes have not been digitized or are still on paper?
It is a multi-faceted problem. The first problem is that some things are still on paper. Another problem is that things are in digital format, but in different repositories or in different standards. Those are areas startups are trying to address. It needs a very heavy industry contextualisation for solutions to emerge. It means understanding parts of the industry very well, such as trade documents. How do one country’s customs documents relate to another country’s customs documents? There are also human beings in the mix, such as customs agents. You see some interesting technologies around, for example, leveraging game theory with the understanding it’s not just digital machines behind the scenes, but also humans making decisions.
Are you looking at startups addressing industry fragmentation, particularly in the trucking industry regionally?
In certain markets, the trucking industry is extremely fragmented. For example, you might have a majority 90 per cent of operators with less than five trucks. But other markets are a bit more organised where it’s slightly less of those and more of 100 trucks and above. Whether it is very fragmented or pseudo-fragmented, there’s always that opportunity of overlaying a digital aspect to it.
Are you looking at any startups that might address that like a ‘Grab for trucks’?
We have been talking to a lot of startups in that space in the emerging markets. And we are particularly looking at those with differentiated technology, especially if it’s scalable, preferably across a region or a certain type of market, or globally.
Are you close to another investment?
We are close to a few investments; get back to me in the next couple of months. We very actively scan on a global basis and we work with partners across the globe, everybody from fellow VC partners and more early-stage VC and incubator partners.
Have you seen startups trying to address issues caused by the trade war? Is it becoming a real pain point?
I have not seen startups emerging specifically because of the trade war. But I’ve seen startup solutions that become obvious to be deployed because of the trade war. Planning and forecasting solutions and yield-capacity planning tend to come up a little bit more in terms of addressing some trade-war related problem areas. There might be cases where startups, because they already play in a certain space, such as cross-border or logistics fulfilment, would see it as more of an ancillary benefit. It’ll be interesting to see solutions that come out of Brexit. There are a lot of questions and uncertainties around trade and customs.
Are there particular areas that a lot of startups are focusing on?
I see a lot more startups addressing niche areas that formerly were not addressed. It can be as niche as digitizing customs documents and digitizing EBLs (electronic bills of lading) and facilitating trade finance workflows and decisions. Part of that is digitization and part of that is some rules-based heuristics method to facilitate decision making. And, of course, there’s how do you deal with planning and optimization of capacity and goods. That’s been an age-old problem as well. So, more and more startups are starting to address niche areas pertaining to these generally larger problems.
What types of last-mile issues might be addressed, particularly in high-traffic areas such as Jakarta?
From a B2B perspective, there’s moving stuff from the warehouse to the retail shops. They will hit those kinds of traffic problems [similar to B2C issues]. Then the question goes back to my forecasting problem again: Are they actually placing inventory at the right location to distribute it optimally? If that inventory is at the wrong location, they’ll hit more of these kinds of problems. If you have more optimal inventory placements, then you hit fewer of these problems.
For example, in the US, in mature markets and some of the mature markets in Europe, they understand this is important. That’s why Amazon has their hub-and-spoke strategy. If I’m going to do delivery within two hours, like in China, I need to design my supply chain accordingly. In the emerging markets, there are some concepts like that, but not as mature.