Supply Chain: “The right AI model in the right place could be a massive advantage”
Our team recently met Ernesto Rodríguez, Co-Founder & COO at Tryolabs, a global company specializing in designing and implementing AI-based solutions for enterprises since 2010. Ernesto is an engineer and a tech entrepreneur who has been working in applied AI before it was trendy. He is passionate about connecting the dots between academic knowledge, industry opportunities, and creating the future. He tells us more about the implementation of AI into the supply chain and its consequences.
Can you give a simple definition of AI? What are its main principles?
Ernesto Rodríguez : Here's a non-technical definition that I like: AI is about making software learn from experience. That "experience" appears in the form of data and how we use it. Data is the fuel of all the processes.
So, typically, in AI, we build systems that receive input (that could be unknown beforehand) and that are trained (with known data) to decide on an "intelligent" output.
How can AI help a business, especially with regard to its supply chain? What are the practical advances brought by AI?
E. R. : We see all kinds of applications. In warehouses, there are different types of robots now that rely on sensors and the intelligent use of their input to speed up processes, do inventory management or even reduce physical risk. On the algorithmic side, there are applications for predictions and forecasting of demand, pricing and ticketing, routing, and logistics. Machine Learning models are allowing to make many of these decisions more informed and optimal.
In general, there are many applications where AI is helping scale processes, use the correct information and automate highly repetitive tasks, with the end goals of productivity, cost reduction, and efficiency
How can a company with no experience embrace AI?
E. R. : Some companies are aiming to become AI-first companies. That's a significant change, and it's challenging. Companies don't have to change what they are to become AI-powered companies; they should maintain their domain expertise and add new tools that improve their processes and allow more data-driven decisions.
In most cases, the recommendation would be: approach AI through the right partner. When approaching AI projects, there are many challenges, so the right skills and the right tools are required to have a successful first experience.
I always advise companies to start with a small PoC (proof of concept), find the right problem, and show its positive impact. Once that's done there will be easier to have more and more people from the organization on board to try more ambitious things.
Start small but start. Leadership should be 100% on board and opening the door to experimenting. The right AI model in the right place could be a massive advantage. With the right partner and an innovative mindset, the benefits could be incredible.
Tryolabs helps companies build custom AI solutions. Could you share with us a specific study case where you implemented AI and the results for the company?
E. R. : At Tryolabs, we have developed demand forecasting and price optimization solutions, proving our approach with leading retail players worldwide. In particular, for a leading travel retailer of luxury products, we helped boost their gross margins by 28%.
Product experts and market analysts used to struggle to define an optimal price strategy for their entire catalog. As the number of items increased, optimally pricing each one becomes a cumbersome task. Along with the client, we defined a universe of over 100 SKU's (stock keeping unit) to perform the pricing optimization. Our approach integrated different data sources ranging from historical prices to macroeconomic data and competition price strategies.
We produced a demand forecaster model to build dynamic SKU demand curves and optimize prices weekly. This reactive model automatically adapts to new ongoing real-world data, including built-in support for out-of-stock and inventory replenishment recommendations. The outcome was a gross margin boosted by 28% in the stores where the solution was applied.
According to you, in which economic sector will AI be the most disruptive?
E. R. : AI has for years proved to be useful in many industries and sectors. It's difficult to predict where it's going to impact the most. Typically, where there's a bigger need, technology finds its way quicker.
Many companies are accelerating their AI investment as a result of Covid. So, I'll bet for some AI-based disruption in communications, travel, transportation, commercial real estate. And other sectors that have been impacted heavily by covid.
In our daily lives, I expect to see more and more IoT devices with artificial intelligence flying, rolling, and even walking around pretty soon!
Ernesto Rodríguez will be a speaker at the Digital Supply Chain Europe Summit next 14th September at European Convention Center Luxembourg. Among other things, he will deliver a keynote speech entitled “How to embrace AI: demystifying the technical challenge”. Digital Supply Chain Europe is part of the renowned tech summit ICT Spring Europe, attended by more than a hundred experts.