Innovation is such a generic and overused term. It seems that every organization claims that they are, or aspire to be, innovative. But what does it mean to innovate with all the artificial intelligence (AI) capabilities coming to the edge? What kind of approach should you have to take this expansive and complex thing that is edge computing plus AI to discover opportunities for creating value that was not there before for your business or your customers?
What Does Innovation at the Edge Look Like?
Innovation starts with an idea. However, good ideas don’t come about without awareness of the frontier of what is possible and practical. If we look at the “Edge” primed with a layer of AI as a canvas for innovation, there is no avoiding grasping the concepts and the state of the art and technology that I have covered in the Cutting Edge series thus far.
Firstly, without this foundational knowledge, there really can’t be meaningful ideation. Secondly, you wouldn’t have a reference for determining feasibility. After all, not all ideas are good ideas. Not all innovations are valuable. Certainly, ideas that are neither can’t impact your business positively; they can actually be detrimental.
One of the things that I try to establish with my clients, when advising them on innovation or facilitating ideation sessions, is a sense of what the net new is, in terms of the state of art and technologies. I also help them understand that innovation is not about technology in most cases. It’s about the application of technologies into novel forms of products, services, or capabilities that help us do things better. If we arrive at a great idea, hopefully we can anticipate significant impact and benefit in its implementation for our business and our customers.
In this sense, innovating at the edge with AI is not a unique exercise. The challenge many organizations will face is getting a real feel for where technologies are today. How mature are they? Where are they on their commercialization path? Can we rely on vendors, service providers, and consultants who claim they can support our ambitions to innovate and reinvent our businesses at the edge with AI?
How Should You Approach AI Innovation at the Edge?
Let’s face it, innovating is not as easy as we might like it to be or as we might claim it to be. It is especially difficult to arrive at innovations that have an impact. It is also difficult for business folks to grasp the implications and potential of technologies to properly inform their imagination and creativity.
So, is there a simple way of approaching or looking at Edge AI that is digestible for non-tech-savvy constituents of an innovation initiative? Yes.
But let’s start with the net new. Edge AI is making the edge, as we knew it and know it, more cognitive and intelligent. What do I mean by cognitive versus intelligent?
Cognitive capabilities are those AI-enabled applications that help a system perceive and understand the context.
A great example is speech recognition which uses NLP (Natural Language Processing) AI methods to help machines understand human speech. A classic Edge AI application is Alexa or Siri. Another good example is the various use of on-device inference in many IoT applications that help a system make sense of data streaming off a multitude of sensors.
While the concept of intelligence in humans and machines is debatable, I like to think of intelligent capabilities as higher-order AI applications that help a system learn, understand situations and concepts, and adapt (make decisions) to changes in its environment.
While some neuroscientists will argue problem-solving is a basic cognitive function, in the machine world, we tend to see it in more intelligent systems. We see this in decision engines that apply predictive ML models to determine the optimal “next best action” based on observable customer behavior.
Edge AI is bringing unprecedented cognitive capabilities to new categories of devices that are not only able to execute increasingly robust AI operations, but also, more commonly, can connect with other devices, edge infrastructure, or the cloud. This enables these connected, intelligent devices to participate in higher-order intelligent functions or applications by tapping into the cloud.
In the future, many of these intelligent edge systems will be able to reside more locally across the edge.
Extracting Value Out of Unstructured Data
The edge today, any edge, is a trove of unstructured data. There are also huge gaps in structured and unstructured data sets that Edge AI can help enterprises capture and convert into machine-readable form. This is a big deal for many enterprises that have holes in their operational data or simply don’t have visibility to what they would like to have at the edge of their business.
This is the essence of the cognitive opportunities that Edge AI presents to an enterprise—sensing more about your business to know more.
Not only is it important to make sense of data captured across the edge, but it’s also important to be able to form a complete and quality corpora of data that can be used as inputs into building efficacious and valuable AI applications. This is that critical first step that many enterprises must take, especially those wondering why their AI strategies are ending up in POC limbo or as data science projects that seem to go nowhere. Get that cognitive capability layered on top of your sensor network, your Internet of Things.
As you start to develop a solid foundation of knowledge for your AI applications, you can then start to consider higher-order intelligent and maybe even autonomous functions for your business enabled by Edge AI and automation.
For the moment, these advanced AI solutions are constrained by the cloud, where much of the model training and heavy-duty inference operations still reside. Fortunately, that won’t be for long, as we see the introduction of new models of distributed, federated learning emerge on top of novel edge computing architectures and infrastructure implementations.
The CXO’s Edge AI Imperative
For the moment, the key question that CXOs should be asking themselves, as they contemplate innovating at the edge, is what their organizations could do if they could see their company better across the very fringe of the business. What enhancements to their products and services could be possible if they were contextually aware?
Edge AI Fills a Gap in the Retail Industry
A great allegory is what we have seen in the retail industry. For the longest time, online retailers have had a huge blind spot. It’s called offline. Ecommerce companies, such as Amazon and Alibaba, had no idea what their customers’ offline behaviors were. Hence, they suffered a huge gap in their data set that got in the way of their 360-view of their customers.
What did they do? Well, Amazon introduced Alexa Echo and a whole line of smart speakers and devices to get a better sense of your life. They developed cashier-less shopping experience with Amazon Go, with highly instrumented shops enhanced with cognitive AI to monitor floor traffic and track inventory. Their Edge AI system helped to see better and see what they couldn’t before.
How could better and more expansive visibility drive impactful innovation for your business?
Now, start ideating. Good luck!
Looking for real-world insights into hyperautomation? Subscribe to the Hyperautomation channel: