How AIOT is Driving Digital Transformation for Industrial Organizations

The Internet of Things (IoT) paired with Artificial Intelligence (AI) , is also known as AIoT, and it is digitally transforming how industrial companies operate. Industrial organizations utilizing AIoT have collected data from places they’ve never been able to before, and in turn, drive new insights, create new user experiences, and go to market with new business models.

This is not a hypothetical, “virtual reality will change the world” scenario either. AIoT is real, tangible, and making an impact today. We are at the forefront of a monumental shift in how industrial companies operate in the new digital world.

From the forests of Australia to the deserts in Ethiopia, AIoT is making a positive impact on people’s livelihoods, health, and businesses. But first, let’s talk about what AIoT really means so we can separate the buzzwords from reality.

Combining AI + IoT is very similar to how the nervous system + brain operate.

One way of describing AI + IoT is that IoT is the nervous system (delivers the data), and AI is the brain (acts on the sensor data). We can extend this to include our senses (touch, feel, taste, see, and hear), which are the remote sensors that perceive the world around us. That sensor data must be automatically processed for the human body to make decisions and react.

There is a new wave of intelligent devices coming. IoT ushered in the era of “smart devices”—devices that can provide data and be controlled remotely. For example, wearable devices, connected appliances, digital scales, robotic vacuums, etc.

It is no longer enough to show people data from their remote devices. End users are looking for and expecting higher levels of insight. Adding AI to IoT (AIoT) gets you Artificial Intelligence of Things, and it allows devices to make decisions and act in real-time on data without relying on humans to interpret it.

IoT allows the analysis of data, generation of data, and human reactions to the data. AIoT adds action (or automation)—analysis turns to action, data generation turns to insight generation, and reactive turns to proactive.

Devices are getting more powerful and smarter (without being connected to the cloud 24/7)

Connecting things via hardware+software isn’t something new. It’s been happening for ages, and we called it IoT. Algorithms have also been added to embedded devices for ages. AIoT is taking IoT to a whole new level by adding intelligence, which creates new user experiences while keeping the value of traditional IoT (data collection and remote control).

The biggest difference now is the AI capabilities that can be added to traditional IoT devices. Instead of just collecting data and presenting it for analysis, new AI capabilities help businesses make better decisions and create new value, with its ability to process high dimensional data, identify trends, and perform predictive analysis.

Reasons include:

• Accessibility of computing power
• Availability of deep neural networks
• Accessibility of data to train algorithms

So the data collection part hasn’t changed, but now what we can do with that data has become much more accessible.

AIoT also includes the notion of “putting the intelligence and processing where it is needed most.” Traditional IoT delivers data to the cloud, and any processing takes place there. This does not work in certain scenarios where the data pipeline is narrow (cellular connection), the data pipeline is intermittent, quick reaction time is required, or where there needs to be long battery life (power conservation).

In these scenarios, the data processing needs to happen before the data gets to the cloud. This is called edge computing. This is possible today because there is a real effort to create processors that can process raw sensor data on the edge in a cost-effective and efficient way.

The difference between AIoT and IoT

The biggest difference between AIoT and IoT is mainly where the data is processed. Processing on the device enables this and enables it to be done in real-time, on its own. Without this, the action would be initiated from the cloud, either automatically after processing the data in the cloud or from a human after looking at the data presented from the cloud through a web app dashboard. In this case, real-time action would not necessarily be achievable.

IoT looks like this:

For AIoT, the majority of sensor information is processed on the “edge” of the IoT system, which means that information is actually analyzed and processed on the device itself using AI algorithms.

Speed matters, and processing on the edge ensures real-time notifications as soon as the event happens.

In our charity: water example below, if we had to send the sensor data to the cloud for processing, the business model would fall apart simply because of the cost of sending raw sensor data to the cloud. The raw sensor data is being transformed by the algorithms on the edge into higher level insights or metrics (mainly pump strokes per hour and water flow per hour). This transformation allows the business model to work. And for our Flexco example, a “real-time” event notification feature would not be possible if processing didn’t happen on the device itself.

“AI needs to accept human behavior the way it is, not the way we wish it would be.”

-Wired Magazine

Digital sensors don’t need to be added to everything, some things bring value without digital connectivity. The key is understanding how adding digital sensors to analog things can bring value to the customer.

There are several reasons why a business would consider investing in digital sensors and AIoT:

• It Solves a specific stakeholder need or problem.
• The need for products to provide more automation. (cost savings, better user experiences, new value creation)
• Problems getting a return on their existing connected product. Adding AI provides new value.
• AI applied to IoT opens up new possibilities—those that haven’t been feasible before.
• The need to get more critical information to accelerate a business’s digital transformation journey.
• The ability to utilize AI to enable real-time action and allow the system to act immediately.

So, how do you know if AIoT is right for your product or service?

It has everything to do with human-centered design. Human-centered design is a creative process used to solve problems for your users or customers. It starts and ends with the people you’re designing solutions for.

Here is the way we think about Human-Centered Design (HCD). We look at it through three lenses:

User Desirability:

Is this solution needed? Does it solve a big problem that users have? Do users desire this solution? A beautiful user interface or experience is great, but if the user doesn’t desire it, then it violates the principles of human-centered design.

Technical Feasibility:

Is this solution technically possible? If it is possible, how long will it take to build? Can and will the organization support this technical solution?

Business Viability: 

How does this bring value to our business and generate revenue? Does it align with business goals? Is it within the company’s budget requirements?

If you’re able to make a digital sensor work for a product that is highly desirable to your customers, is technically feasible, and adds value to your business, then you’ve got a perfect use case for investing or experimenting with AIoT.

If done right, AIOT can open up the possibility of new business models.

Smart companies understand their customers’ needs, then align new technology to support those needs while creating a new business model. HCD must be used to identify the right data and the right successful product to deliver the data and identify where AI must live within the AIoT system.

How human-centered design + AIoT transform customer experiences and business model (with real case studies)

When is a digital sensor not just a digital sensor? When it’s the foundation for creating new customer experiences and new business model innovations. Here are two real-life (before and after) AIoT scenarios that do just this.

What happens when you connect a hand water pump in Ethiopia to the cloud?

Through an HCD process, charity: water uncovered a key pain point: Donors of the non-profit organization were also interested in understanding exactly where their donations were going. Charity: water realized this was a major pain point for people when they found in their research that 42% of Americans don’t trust charities.
Our goal while working with charity: water was to take a water pump in the middle of Ethiopia and connect it to the cloud, then donors could see how much clean water was flowing remotely and AI could predict when a hand pump needed to be serviced before it broke down.

The sensor was initially a way to meet these goals and had the added value of providing insights to the well and pump maintenance team so they could plan to repair the pump before it would break down.

By creating a custom sensor connected to the cloud, we were able to make a standard hand pump into a digital AI-connected device accessible from anywhere in the world. This fulfilled charity: water’s mission of providing clean and safe water to people in developing countries.

That’s the power of considering humans first, instead of thinking purely about the technical solution. You get real solutions to real problems for key stakeholders. We could easily connect a digital sensor to a hand water pump and call it a day. But that’s not what this is about.

User Desirability for Flexco:

This solution solves one big problem: It minimizes pump down-time resulting in more clean water.

At the locations where the sensors are used, 30% of wells are non-operational, and on average, it takes 11 months before a hand pump breaks while taking as long as 4 years before it is fixed. In this case, using a sensor to gain insights on which wells need to be fixed before they break down is highly valuable and desirable to the end-users. We only knew this because our teams spent time in Africa, interacting with communities and hand pumps to understand how they’re used, and the technical details behind the hand pumps themselves.

Technical Feasibility for Flexco:

Not only was it technically feasible, but we were also able to create a great user experience along with it.

As a digitally-connected edge device, the sensor needed to be integrated or affixed to the pump itself. Through a series of prototypes and gauging the user experience, it was determined that the sensor housing was best integrated as part of the core pump to ensure a weather- and tamper-proof design that was easy to install.

Business Viability for Flexco:

Charity: water is a non-profit organization, so by default, this solution serves the purpose of its mission to provide drinking water to people in developing nations.

Donors can clearly see that their donations went to a pump that is still working, which will likely bring in new donors. The cost of the devices is low enough to appeal to most donors, allowing them to provide a long-term impact.

Local maintenance teams now know and can plan when to repair the pump.

What else could charity: water (hypothetically) do with this information? Remember, this is collected data from places we’ve not been able to collect data from before. Digital transformation involves using the data from the AIoT system to create new value for new stakeholders. How else could we expand on the business viability?

Here are a few ideas:

• Charity: water could potentially partner with local municipalities and governments (a new stakeholder) to give them the information they need to distribute more clean water.

• The pump manufacturer (another new stakeholder) could now get insights into how long their pumps last and how they are being used. In this case, the data can potentially be sold back to them.

Conveyor belts for the mining industry that go beyond cloud-based maintenance tracking tools

We helped Flexco, a belt conveyer solutions company, create a critical and innovative belt cleaner monitoring system that uses predictive insights to allow operations to move from traditional methods of monitoring cleaner performance to gathering data remotely and in real-time.

User Desirability for charity: water:

Through our HCD approach, this was a human-first, technology second solution:

We designed the sensor in a way that allows battery replacement while the belt is still in operation.

We developed a compatible app and cloud-based connection that provides access to data at any time for logging maintenance, updating schedules, and spotting data patterns and trends for future planning.

It made end users’ jobs easier by providing insights into how their machinery is operating and informing them of possible next steps.

Technical Feasibility for charity: water:

Not only is it technically feasible, it’s operating in mines all over the world, generating valuable data.

As a digitally-connected edge device, the sensor needed to be integrated or affixed to the pump itself. Through a series of prototypes and gauging the user experience, it was determined that the sensor housing was best integrated as part of the core pump to ensure a weather- and tamper-proof design that was easy to install.

Business Viability for charity: water:

This one is straightforward: Fewer conveyor belts shutdown equals increased productivity, production, and overall uptime. With predictive maintenance, more work is being done, with little to no interruptions. How could a business say no to that?

Additionally, Flexco sales teams now have insights on which clients need replacement parts.

And that’s just the starting point.

AIOT is cutting edge, but an essential part of every growing industrial business

Organizations need to get on the digital operating model to digitally transform. To do this, they need to understand the value of data and how the right digital and connected products can get them there.

To find the right digital product, a new and unique user experience may be required. Adding AI to IoT may be just the thing that allows for new user experiences—user experiences that may otherwise not be feasible. At this point, this is cutting edge stuff, but if you want to ride the exponential digital operating model value curve, it may take cutting edge technology to propel you and your company forward.