CONNECT WITH US

Connecting AI and supply chain: Q&A with Scale AI CEO Julien Billot

Judy Lin, DIGITIMES, Taipei 0

Scale AI is Canada's artificial intelligence (AI) platform dedicated to building the next-generation supply chain and boosting industry performance by leveraging AI technologies.

Digitimes recently talked to Scale AI CEO Julien Billot, who is also an adjunct professor of HEC Montreal and the Montreal lead for two transformational programs - NextAI and the CreativeDestructionLab (CDL) - aiming to launch and grow startups in AI leveraging Montreal tech and business ecosystem.

Q: For people who are not familiar with Scale AI, could you give a brief introduction to your organization? When was it founded, and what are its mission and goals?

A: It came from a Federal Initiative of the Federal Government of Canada. Back in 2017, they foresaw that the technological development of Canada was lagging behind, so instead of taking a top-down approach to fund research to companies, they decided to create five Superclusters to do the reverse and to fund industrial projects from bottom up, in support of researchers. So with the big auctions, Montreal was fortunate to win one of the Superclusters. Scale AI is an interesting project based on two things - Montreal is very strong in AI. For a very long time, we have devoted in fundamental research, and professor Yoshua Bengio of University of Montreal won the Turing Prize along with University of Toronto's Geoffrey Hinton and Yann LeCun of New York University in 2019. We have a long history of fundamental research in machine learning and AI.

Montreal is located on the St Lawrence River, at the door for Great Lakes. It is a strategic part of Canada in terms of supply chains and logistics. A lot of industries involved in supply chains are based on Montreal. As more and more supply chains are connected now, you can efficiently apply AI to every supply chains.

So back in 2018 we decided to apply for a Supercluster, to leverage AI in applications related to supply chain activities, and Scale AI was born. We went live two years ago, in May 2019. Our mandate is specifically to fund and facilitate initiatives that are helping to introduce AI into supply chains. Companies which use supply chains can be transportation companies, or companies that use supply chain for their production activities, such as Bombardier, or other industrial companies. Our mandate is going to last till March 2023. We hope to help introduce AI to industries here in Canada, in a nutshell.

Q: Scale AI as the Supercluster of AI for Canada, as I can imagine, is an upper-level coordinator that pulls in research institutes, companies, startups, and talents and create synergies among them by generating exchanges and project collaborations. What was the reason for Canada to see the necessity for such an organization to be created, since there are already all sorts of startup accelerators, VCs, to invest in AI startups?

A: That is a great question. It is exactly the core of the creation of the Supercluster Initiative in Canada. You have lots of players across the various fields, but there was no coordination, and more importantly, no fundings at the time. Introducing AI into supply chain, of course, can help AI startups. I try to explain this with a triangle. The three dots of the triangles are: (1) Fundamental research, you need to have researchers who build algorithm and find solutions; but if they just work by themselves, they only produce papers, and that is not very productive for the country. (2) Startups, many of them were created by work with researchers, so you need to create connection between researchers and startups. (3) And then the venture capital and accelerator programs. They could help facilitate the creation and fund the growth of startups. But obviously, you need to help large corporations and small and medium enterprises to transform into AI. To do that you need to create the links between fundamental research and industries, as well as links between industries and startups. And Scale AI facilitates all these links in the triangle by subsidizing programs.

Typically, we have three types of programs. One is to help accelerators and incubators create and develop startups in AI, leveraging fundamental research. That's to help linkage between research and startups. We also fund industrial projects in which companies leveraging AI or fundamental research by researchers to transform their operations. We fund programs in which startups are incentivized to collaborate with big enterprises or organizations to innovate with AI. We create the links, but more importantly, we give the money. Before Scale AI, all the parts of the triangle existed in Canada, but they were loans, and not public money fundings specifically to create those three links to pull them together. We focus on funding projects between corporation, research and startups related to the supply chain transformation leveraging AI technologies.

Q: Since many of our readers are supply chain manufacturers and are keen to look out for AI solutions or applications, are there examples of cross-industry collaborations for supply chains you can share?

A: Sure, we are funding two types of industrial projects. We are helping companies with better AI program designed supply chains. We have, for example, funded a company called Optel, which is based in Quebec. They work on traceability with AI for the aluminum industry, which is very important between the US and Canada. In that case, we are leveraging AI to improve existing solutions for the aluminum industry. We also help horizonal suppliers to design better products leveraging AI available for all industries.

In the same category, we are helping a company with virtual reality, a company called OVA. They work on VR, AR and XR for the maintenance of engines. The operators wear VR goggles, and because of AI, it is easier for them to identify issues in the engine. We also have companies design products including AI, but serve different industries.

The other type of projects that we fund are big companies. Some of them are transportation firms, some are industrial companies, and some are retailers. And also hospitals. Those companies are the ones with logistical issues in their supply chains. They are heavy users of supply chains. Examples such as Loblaws, one of the big retailers in Canada, or Bombardier, the business jets manufacturer, all need to improve their supply chains. They mainly use AI startup software to introduce to their supply chains. We pay half of the expense to this kind of projects. We are helping horizontal solution providers develop better supply chain design solutions embedding AI. On the other hand, we are also funding vertical-side industries to help manufacturers and retailers improve their logistics, leveraging AI.

Q: Since AI is applicable to all kinds of industries, do you have in-house experts from various disciplines to help provide domain knowledge needed for cross-industry collaborations?

A: The answer is no, because it is not our mandate, which is to create links and to fund, in order to create an ecosystem. Although we don't have in-house experts, we do have a long list of experts working in the ecosystem. If people have request or demand about finding experts in certain area or for certain subject, since we know the experts, we can put them in connection. If they agree to do the projects, we can fund the projects. We are a small organization with only 13 people, and our mandate is to create networking of partners and take care of funding. We also work on intellectual property (IP) by directing people who want to share IP, or to protect their IP, to the right people who can help them. In a word, we are a facilitator and a funding organization.

Q: How does Scale AI introduce AI startups to work with traditional industries? How do you know they are good matches? Do the traditional industry clients tell you their pain points and then you try to find AI firms suitable for them? How do they know how AI can solve their problems? How does ScaleAI assist in that thinking process?

A: We are not perfect every time. Up to today, we have funded 40 projects. So we begin to have good knowledge of the system, and maintained a network of people in the ecosystem, including researchers and experts. We also fund chairs in the university ecosystems in Montreal and also accross Canada. We know all the best researchers. And as we oversee the projects, we know which players are doing great jobs in the areas of their specialty. We also support incubators and accelerators across the country. Every year, we support 100-200 startups. Indirectly, we know them and what they do. For example, we hosted an event today at H.E.M called "Supply Chain thinktank," in which more than 250 people attended, mixing corporate people and startups. In Scale AI, we facilitate and organize these events where startups can present and hear industry's demands and needs. And industries can learn about what startups are doing. It is a great way of creating connections between industries and startups.

We have a unique database, the only one in Canada, which takes in 100-200 startups every year doing AI and supply chain. In three or four years, we will have close to 1,000 startups designing AI product for supply chains. That is a treasure for industries. If they have needs, we can match them.

Every year we host a big conference called "AI in Action" in Montreal, and we are also going to do that in Toronto and Vancouver next year. The goal is to give concrete examples of applications and insights of AI introduction. We also do conferences by verticals, such as human resource, transportation, retailers, etc. Every time we do this, we make sure to pull in corporates, researchers, and startups in all our vertical or general conferences to create opportunity for exchanges and connections.

Q: When deploying AI solutions, many companies would experience challenges such as a conflict with their original business process. How do you assist them overcome those challenges?

A: We have another program which I have not talked about yet, which is all about training. We devote 30% of our funding to training. The reason is because we recognize that introducing AI is not only about using a software, but it is about transforming a company. More importantly, it is more about the culture, working around the change in management, processes, and education.

We have a bunch of training available, which covers all steps of maturity for companies. For example, a training called AI 101 is for people who don't know anything about AI, or don't know how to introduce AI in their verticals. We also have programs to help you how to think about AI usage by reviewing your processes. We have done a program with H.E.M for Ericsson to train 275 people. The interesting things is, they trained 200 engineers, as well as 75 managers. They want to make sure the managers understand the disruption created by AI, and to understand engineers using AI. Scale AI has funded 50% of the expense of the program, pleased to see they are taking full account of the scope of introducing AI.

We really believe, for you to succeed in an AI introduction, you have to follow all the steps. First of all, you need to understand why you want to introduce AI, and what to be done. And if we receive a project, we also look carefully to see if the work has been well done, so that no public money will be wasted. When companies come to us, we want to carefully assess what they want is going to solve their business problems. Then you need to have the right sets of data and the right AI technology. Finally, we want to make sure post integration or implementation of software works smoothly.

How do we assess whether a project will be funded? We look at how the business needs are explained: what are the technologies? How can the data issues be solved? We encourage people to have proof of concept and sound strategy plan or blue print and data before coming to us. And finally, what is the plan for integration? We will review the progress of the projects, and if the quality of implementation is not meeting our expectations, we can always stop remitting, and let them come back to us to readjust and put the project back on track. We supervise and take care of the projects. Most of the work is done before acceptance, to make sure the technical quality of the project and all angles are covered, and have a great project from A to Z. In the end of the day, we invest with public money, so we want to make sure every dollar we invest is every dollar returned.

Q: What are the areas that Taiwan can work with Canada's AI companies in non-supply chain areas? Energy management? We already had two blackouts in May.

A: Researchers and consultants in Canada have very good knowledge in introducing AI software for supply chains. Energy management is a supply chain issue. The way we define supply chain is the way we move goods and services, but also digital bits or power, as well as human resources. When you think of AI, it is actually a mix between AI technology and operational research, which are more pragmatic and more able to define clear solutions for real problems. Over the past two years we have accumulated knowledge and experiences from the projects we have done, and I am sure there are also experts with the knowledge in Taiwan. We believe there will be opportunities to share and discuss between Canada and Taiwan.

Scale AI CEO Julien Billiot

Scale AI CEO Julien Billiot
Photo: Scale AI