CASE STUDY: Ocado harnesses the power of AI
Friday October 5 2018
Online grocer Ocado has been tapping into the power of AI to navigate complex technology landscapes
Digital transformation is not simply about making your products available going online. Ocado is the world’s largest pure-play online grocery retailer and it’s never owned a shop and probably never will. Its stock is the technology to enable ordering, fulfilment and delivery. It employs over 13,000 people, and many of those are engineers. Compare that to Walmart’s 2.1 million.
At the heart of Ocado’s automated fulfilment are a patented array of robots and grids that pick and pack groceries in a warehouse that resembles a space station more than a supermarket.
This approach means that in technology terms, it is up against IBM, GE, Honeywell and ABB just as much as it is competing with supermarket chains who are offering their own ecommerce solutions.
That poses a very specific challenge as companies in this cohort have been patenting for decades and there are now well over a million patents in the area of on-line shopping, fulfilment and delivery. If you include Ocado’s trial with autonomous vehicles add another million.
It seems entirety fitting that Ocado have turned to AI to find a solution and are working with Cipher, the IP analytics platform developed by Aistemos. So how do you dodge a million bullets? Part of the answer lies in the way Ocado reframed the question.
This is how Lucy Wojcik, Head of IP at Ocado explained the challenge: “As Ocado moved in to the technology and hardware arena, it was clear the IP context would be challenging. The ability to allocate significant headcount to traditional IP landscaping was simply not an option. This provided the impetus we needed to think smarter. Cipher gave us the ability to understand the technology landscape and was an important foundation layer in creating a strategy for protecting our innovations whilst enabling us to understand how best to deal with others in potential conflict areas.”
To better understand the challenge requires a quick detour into patent wars across the ages. This includes Wright and Curtiss (powered flight), Edison and Westinghouse (light bulbs), Polaroid and Kodak (instant photography) and Nokia, Apple, Samsung and just about everyone in the smartphone wars.
The basic premise is that patents confer the right to exclude others from using your invention. Extrapolating from that position creates the myth that patents confer a quasi-monopoly.
While there may be some truth in a world where there are not many patents, a fresh approach is required when there are now 100 million patents owned by 1 million companies registered in over 100 countries.
This approach involves training algorithms to identify and sort those patents that may be relevant to a given technology or innovation.
So in a hypothetical example where you need to understand the sensors used by European OEMs in autonomous vehicles, AI now enables you to produce market maps of the sort included below.
This type of report is available to Ocado instantly which enables it to quickly figure out who’s doing what and the areas of Ocado’s technology horizon likely to be impacted by such competition.
“The alternative is to fly blind and to hope that nothing goes wrong. That’s about as crazy as allowing 1000s of our bots to move freely on our grid and hope that they don’t hit each other,” said Wozcik.
Conventional wisdom suggests that the expertise necessary to analyse patents is a task that can only be performed by humans. Not only is that no longer true, but it is no longer practical. Reading a patent can take well over an hour so even a cursory review of 1,000 patents would take months.
Cipher CEO Nigel Swycher, himself a former IP lawyer, suggests that this is not purely a matter of human vs machine: “Cipher’s role is to make patent information more accessible. The protection and exploitation of patented technologies is now a key commercial consideration and that means presenting data in a form that can facilitates communication between heads of IP and the technology and corporate strategy teams they support.”
AI is ubiquitous so it is perhaps no surprise that algorithms have a vital role to play in the innovation lifecycle that begins with the design of robots and end up with groceries on your doorstep.