Intention-Driven Design

Intention-driven design starts with self-aware sentience. As organizations build digital models in their physical world, everything from sight, sound, smell, taste, humidity, and touch are coming through sensors and analytical ecosystems. Digital equivalents of analog systems and even humanity are being delivered.

Cameras play a role today in our visual senses. We have listening devices that are more than audio—they can detect vibration. Other instruments from the speedometer to the durometer, the accelerometer inside your phone, take down acceleration and position. Even climate is being integrated into digital business systems to determine if humidity, temperature, or weather could impact a decision. These sensing networks and their related sensor and analytic ecosystems are the front lines for intention-driven design. And these networks that were once isolated are now set up in a connected world of ecosystems that take these data points and broker them to both the consumer and enterprise worlds. Smart machines and wearables are providing new types of sensors that add to the mix of data that’s creating insights. Constellation Research estimates that at least 200 million smart wearables will ship by 2018. These are bracelets. These are fabrics. These are watches. These are eyewear. These are any sensor-embedded devices worn by a human. They can even be connected to humans through human APIs and neural networks. Data from automobiles and medical devices and household appliances, as well as from power generators and building systems, are providing opportunities to improve operational efficiencies, create new business models, and identify new usage patterns.

These systems will not only communicate with one another, but also interface with people, overtly and covertly. The Internet of Things moves from an abstract concept to a living and breathing machine-to-machine meshed network interfaced with humanity. These could be human APIs. These could be new ways to connect through sensor neural networks.

By 2020, the global market for a few billion cell phone SIMs will swell to 100 billion MIMs—machine ID modules. Technologies will include a 100-gigabit optical network—and that’s robotics and manufacturing, building management systems, MIMs, geolocation drones, self-driving cars, smart grids, and software-defined networks. There’s a quantum leap in the quantity and quality of information coursing through these digital businesses.

In fact, digital business disruption is happening right in front of us, and data is the foundation of these digital businesses. As a result, manufacturers are remodeling their factories’ shop floors. They’re trying to figure out how to mimic and how to model and how to forecast what could happen. Retailers are remodeling the shopping experience to see where and why customers engage. Logistics companies are tracking routes and warehouse movements. In the consumer world, home automation and wearables are just the beginning of the sensored data connected to power these new models. Wearables such as Fitbit, Nike, Jawbone, Apple Watch, and Samsung are just the beginning. Nest and Apple iHome modules are just a glimpse of what’s next. The car is also one of the next battlegrounds as we’re looking at sentience in machines. This is also happening in advanced fiber technology and electronics. There’s convergence to create smart textiles—where material science meets the digital age. E-textiles can sense and respond to a range of environmental stimuli. We’re talking everything from monitoring capabilities like internal signals such as heart rate, perspiration, and skin pH to external elements such as CO₂, temperature, humidity, and sunlight. The textiles can then respond through an alert or even a change in color or pattern.

Surfaces such as smart glass also play a role in how information will be captured, presented, shared, and interacted with. In fact, advances in material sciences are playing a critical role in enabling intention. So when we combine these data and signals, we enable improved context.

How do we take all this data and bring out intent? It requires significant orchestration. As these sensor and analytical ecosystems mature, digital or electronic or interface message processors (IMPs) emerge to bring embedded hardware and software platforms together, to connect networks of devices and wearables to work with one another. And they’re working in a similar way to how interface message processors worked in the early days of the internet to connect computers.

Now we’re connecting every end point. Almost every individual is an IP address. Devices allow manufacturers and developers and networks and machines and people to manage and scale with one another and to create networks that support intent. These are where we’re getting our data, and this is why we’re seeing self-aware sentience. This self-aware sentience is creating sense-and-response capabilities.

Probabilistic Models

The second part of being intention driven is following predictive or probabilistic, not deterministic, models.

I know, that sounds like a mouthful, but what’s been happening is that organizations have trained themselves to identify their best business processes, to standardize them in order to improve quality or meet regulatory requirements. The results have moved organizations forward. When they were first introduced, these deterministic rules were great. They led to guided selling techniques, service checklists, marketing funnels, and loyalty journey maps. Unfortunately, these were the best practices of the 1980s, carried over into the 1990s, codified in the 2000s, and now we’re stuck with these fixed processes. That’s what we call deterministic. These approaches worked brilliantly for companies then, but today we are saddled with a set of fixed rules that no longer apply to how an organization works or responds. Especially in the digital world.

Now we’re moving away from deterministic areas toward probabilistic ones. And what’s happening here is really a war for algorithms. This is the death of deterministic models. Those fixed rules worked for systems of transaction, which were nice, neat little systems, and they worked for some systems of engagement. But at the digital scale, there is no right pathway. There is no correct journey map. In a world of mass personalization at scale, there’s only the probability that what lies ahead might be the best path. At times it’s going to seem undefined.

Given the current conditions, the probabilistic model is probably going to be the right way to go. Digital brings all these probable outcomes together and surfaces them based on what’s most likely to happen. So probabilistic models have grown in importance. These are coming from machines learning algorithms, artificial intelligence. They’re sensing. They’re predicting. They’re inferring. And in some ways, they’re even thinking. Systems are gleaning insights from past behaviors on current conditions and translating them into predictions of future models.

What these systems are trying to figure out is, will something happen? Is it going to happen again? Why does it happen? The predictions aren’t 100 percent accurate, but we’re trying to make those decisions precise. We want to accurately guess what the customer may want to do next or what a system needs to do next.

So it’s no longer black and white. Everything is fifty shades of digital gray (i.e., #d3d3d3 for the geeks in the house). Why? Well, armed with the perfect context, probabilistic models can generate precision decisions. To achieve mass personalization at scale, these models have to suggest the best set of decisions at a point in time. That’s the key here—a point in time. The logical conclusion is, then, that the next battleground in digital business will be math—yes, math. The best algorithm’s going to win.

These algorithms are basically small programs and sets of instructions. In a digital world, we need the ability to quickly assemble small units or composable sets of code and algorithms. It’s where the battle lines begin. How these composable bits of code work with algorithms is the secret sauce inside organizations. That’s the core intellectual property. These exist. They’re happening. We already see early versions of them in high-speed trading networks that involve massive changes in volume trading based on thousands—even, in some cases, tens of thousands—of factors in market conditions. This is going to change the way we buy. This is going to change the way we place an order. This is going to change the way our warehouse restocks itself. This is going to change the way we look at fraud and security. This is going to change the way all kinds of things work—from boarding an airplane to walking into a store.

Today, most of these algorithms are focusing on finding the right match or optimizing behavior. But in the future, they’re going to be more sentient. They’re going to ask and answer questions. They’re going to formulate hypotheses. And this is why we talk about this model being predictive and probabilistic, not deterministic. What we’re actually moving toward are models that drive precision decisions.

Self-Learning

The last part of the intention-driven system requires self-learning and cognitive capabilities. What we’re talking about here are powerful systems that augment humanity. Digital businesses have cognition. They learn from previous experiences. They analyze failure. Their goal is to augment human decisions and, at some point, automate the mundane and routine but also surface the exceptions that allow for intention-driven systems. This is a convergence of artificial intelligence, dynamic learning, hypothesis generation, and natural language processing.

What we are really doing is trying to figure out how to take all these volumes of data and make them intelligible. If we want to make better decisions, we need this ability to self-learn. Self-learning and cognitive capabilities enable continuous reprogramming. These advances are really a new class of technology that enables human- and machine-guided decisions. This cognition is part of the foundation for what we call augmented humanity, where we’re taking our collected insights and data, we’re surfacing them in the right time and the right context, and making those decisions. A range of these technologies are now popping up, from facial recognition and human APIs to machine learning, self-learning algorithms.

And that is the promise. We’re basically trying to interact more effectively with computing. We’re trying to help these machines actually self-learn and understand and interact in a way that makes sense. We’re also trying to build deep levels of expertise here. These deep levels of expertise are actually not just in one area. What we’re really trying to do is help machines build knowledge, gain a domain, really change the way expert systems move from a hard-coded model to something that’s always learning, something that’s taking experiences and building on them. Instead of sticking to hard-coded rules, these systems might change over time. This is about the promise of these cognitive systems.

A great example is what’s happening at IBM. The company is looking at a world of cognitive computing and thinking about how the human brain and the mind senses and reasons and then making that work inside a computer. It’s trying to create these cognitive-computing applications that adjust to experiences and change the way they learn, build knowledge, and ultimately engage with other systems. This allows the systems to act as a decision support system to improve decision making, to improve information distribution based on a limited set of data (or a lot of data), and build those into deep industry expertise.

The challenge here is, we want to get to augmented humanity. We want humans and machines working together. And digital businesses are building these models. Traditionally, the idea behind artificial intelligence is that humans are not in the equation. In fact, humans are completely ruled out. The machines are making their own decisions, doing whatever they’re doing. But in cognitive computing, this class of machines is learning from humans and human behavior. Augmented humanity is a part of that. It is creating this level of natural interaction so that we can create feedback loops between machines and humans and sensors. It will allow us to use new techniques to help humans make better decisions. Because while machines can make better automated decisions, humans and machines still need to connect with each other to be most effective.

So the three elements that are most important for delivering on what we’re calling intention driven are self-aware sentience, which is the sense and response; predictive probabilistic and not deterministic, which is building smart algorithms; and self-learning and augmented humanity, which is really about our ability to take cognition and bring it to life. This intention-driven mind-set is one of the five key elements for disrupting digital business.