In a recent Design World webinar (now available on demand) three IoT experts answered several questions about how manufacturers and machine builders can benefit from IIoT with more efficient, productive, and sustainable industrial enterprises … and best practices for approaching cybersecurity, ramping up a digital workforce, and appropriately scaling the digital transformation. Here’s what they had to say.
What’s the best way to get started with IoT?
MacGillivray: IoT is shifting from being a technology solution to a key enabler of digital transformations. Companies on a digital transformation or DX journey are leveraging IoT by bringing together their stakeholders to execute their organization’s collective IoT vision.
Research we did with Schneider Electric shows that more than 40% of IoT projects are funded upfront by the line-of-business … and that’s impressive, as it indicates line-of-business understands challenges faced in R&D, operations, engineering, and even marketing. One caveat: IT is still heavily involved in longer-term budgeting for IoT … and must understand the necessary networks, connectivity, and storage as IoT projects are developed and deployed.
Regardless of the stakeholders at the table, IoT promises to drive efficiencies with intra-company efficiencies, advanced analytics, and artificial intelligences … as it’s a key data source for AI applications. No wonder that in manufacturing, 49% of all digital transformation (DX) efforts are led by IoT.
Until recently, most IoT deployments were internally (operationally) focused — and less focused on driving end customer value. But now, improving customer experience and controlling costs. So according to our research, there are three reasons to invest in IoT now:
- Improve internal efficiencies and trim operation costs — even for maintenance
- Increase revenue opportunities
- Differentiate from the competition
What are some ways to cut through the IoT hype?
Perducat: Key is regarding on an IoT solution as a business decision — not a technology decision. I think it’s easy to get lost in all the technology bells and whistles … but, ultimately, IoT should solve a concrete problem. So first define that problem and then identify the technology and services to solve it. At Schneider Electric, which is leading the digital transformation of energy management and automation in homes, buildings, data centers, Infrastructure, and industries, we leverage IoT to advance solutions for our customers to improve efficiency, productivity, and sustainability.
MacGillivray: Recognize that many organizations aren’t even thinking about their solutions as IoT … but just enabling physical assets to boost efficiency, inform product design, and contain costs.
My advice to manufacturers is to keep the focus on the business problems and involve line-of-business, IT, and operations folks. I think that’s where the IoT strategies become real.
Cabanes: Also consider the industrial company’s level of maturity … and ability of engineering and manufacturing to change business models with smart products and services leveraging data analytics.
One concrete benefit of IoT is the way it can let manufacturers deliver customized products. Our research from a report we did with the World Economic Forum indicates that 70% of industrial companies aren’t leveraging IIoT. You can download a PDF of Unleashing the Value of Connected Products. It also quantifies the potential savings yet to be realized by companies worldwide from IIoT’s predictive maintenance and analytics.
What questions should be asked of a potential IoT vendor?
MacGillivray: When considering an IoT vendor, ask them these five questions:
- What makes this IoT offer worth the investment from a business standpoint? Does it solve a specific problem? Or is it just technology that layers on top without necessarily solving any core issues?
- Is the solution operable across systems? This is really important as the IoT market continues to evolve.
- What cybersecurity certifications and measures are in place? Ensure your potential vendor has sufficient security credentials to meet your organization’s cybersecurity requirements.
- Do you have a comprehensive partner system? No one vendor can provide an end-to-end IoT solution, and anyone who claims they can probably isn’t being honest. IoT solutions are partner-driven solutions … and those vendors who have a best-in-class partner program should be strongly considered.
- Do you offer digital services to accompany the IoT element? Digital services can supplement IoT projects.
When choosing an IoT vendor to support an IoT investment, look for a proven track record. Prospective vendors must have a strong portfolio of proof points from an existing customer base when making solution proposals to yours. Again, look for IoT vendors with a robust partner ecosystem. Any IoT solutions need technology elements from myriad vendors … as no one goes out today and buys an IoT solution off the shelf. After all, IoT solutions are amalgamations of different technology elements — connectivity, hardware, software, and services and security.
MacGillivray: Involve your IT and vested stakeholders across your operations in the early stages of planning. Adding them later risks failure to recognize key elements necessary for the IoT buildout.
Also — start small and then look to scale. Finally, look for a partner that’s really invested in your success … and with the industry experience for insight into your business pain points.
Note that your potential IoT vendor must have industry-specific expertise as well. These vendors will have familiarity with the pain points experienced in specific business processes. Plus their solutions must be interoperable … using open technologies that are flexible and future-proofed — necessities for longer-term IoT solutions.
How can companies create new digital business models with IoT applications?
Cabanes: There are two dimensions to consider here — and I like to use the EQ and IQ human-intelligence models as an analogy. The EQ of a product would be its intelligence or connectivity, even to full autonomous operation. The IQ would be its service, platform, and ecosystem. And the IoT enables both. Consider a hypothetical autonomous product offered with standard after-sale services. We’re seeing that such products serve customers who are not looking for direct ownership … which for example is why Tesla may soon offer ride-sharing products. Robots as a service is a related trend here and indeed a new business model.
IoT also allows autonomously operating products through ecosystems so it’s not a single company but rather an ecosystem of companies selling a design. For example, French company Faurecia collaborates with German car companies on intelligent car cockpits; FANUC collaborates with Cisco and Rockwell on artificial intelligence for industrial robots.
Perducat: We design new offerings and their business models at the same time. IoT transforms data into value … largely using connectivity and the transformation of data into information with context. The technology to do that includes analytics or AI.
Case in point: We help set up machine monitoring for OEMs selling to machine manufacturers. Some such services are subscription-based. Consider those for packaging and conveying equipment. Customers here must be able to track those machines to monitor those machines. Sometimes that’s through distributors or with end users — who define the operating context of that machine. Monitoring the machine relies on connectivity and gathering information for that specific machine for performance optimization … and in some cases, remotely applying fixes to problems. After all, many facilities have precious few specialty technicians … so traditional troubleshooting often involves flying a technician to the machine’s location. It’s much more efficient if problems can get fixed remotely.
What are some pitfalls to avoid when embarking on a digital transformation?
MacGillivray: I don’t think anyone should be solving for a mission-critical problem at the get-go. Engineers should start by looking at something that’s more manageable — and bite off a piece of a larger process to improve. I think it’s important to start small and work out the kinks during the deployment of that first solution. In fact, deciding on IoT project scale is paramount. On one IoT project in which we were involved for improving distribution, we started with one country. Then we scaled in a deliberately paced way to include more geographies and facility types using an iterative design process and permanent communication with customers.
On a related note, scalability is key to making IoT solutions economically viable. Let’s say a manufacturer wants to ultimately deploy in multiple factories. Well, subsequent factories should cost less than the first to upgrade.
What are some ways design engineers can leverage artificial intelligence (AI) and analytics to boost throughput?
Cabanes: One book by Paul Daugherty from Accenture — Human + Machine: Reimagining Work in the Age of Artificial Intelligence — details three imperatives related to AI. There’s a fear that AI will destroy jobs and not create new positions for work. But even today, much production takes a classical approach — that is, very linear — with ultimately unbearable inefficiencies. Plus as of June 2017, there were six million open jobs in North America and over the next decade, three to four million workers will need to fill manufacturing jobs to replace retiring workers. AI will help industries address these inefficiencies and resource issues though will also require the training of workers with new factory skills to support modern automation.
In short, AI complements human tasks and functions. Daugherty calls it the Missing Middle. More work needs to be done in this arena, as it holds a lot of potential. Responsible AI will also need measures of accountability regarding transparency, honesty, business, and human centricity. Artificial intelligence isn’t going to replace humans, it must be controlled by humans. I think this is also something very important.
On the topic of worker skills for AI-supplemented work, we’ll need to teach and train people and those efforts will need to be huge. I believe it will even come to change the way we are teaching young students. Another challenge is data sifting for facilitating actual decisions. Ensuring good AI logic will rely on the context and the integrity of the data being perfect. We should keep in mind that there’s no finish line here. I think it’s going to be a never-ending process to evolve and develop machine learning with AI prompting waves of improvement and enabling things once thought impossible.
How might machine learning add a layer of cybersecurity?
Perducat: First let me underscore the benefits of AI. In our experiments deploying AI in more than 100 of our factories worldwide, we’ve gotten very positive feedback from our plant managers — especially for how AI distills data into usable summaries. Of course, AI in machine learning for IoT can make connectivity a nightmare if we don’t have a way to process the tremendous amount of data produced.
When it comes to cybersecurity, it’s a fact that IoT connectivity poses risks … so IoT solutions must be compliant with all relevant standards before they come to market. What’s more, vendors much work continually with manufacturers on cybersecurity monitoring — to identify abnormal equipment behavior, compromised IT, or unauthorized data extraction.
But we believe strongly that not being connect poses greater cybersecurity risk. Connected equipment is best. With unconnected equipment, we don’t know what’s happening with it; we have no visibility of potential or existing problems. To illustrate, we won’t know if a subcontractor has left monitoring equipment on a machine to push its information to competitors. We won’t know of planted viruses to be activated later.
Of industrial applications compromised by malware in 2018, the equipment most affected was unmonitored and only marginally connected — part of the shadow IT lacking the right patches and software updates. Ultimately, we can only work to be better equipped to prevent security issues in the future … as nobody can guarantee the dodging of all past, present, and future attacks. But the best protection is quick isolation of problems and immediate action — not discovering issues after three months of monitoring that a malicious force launched an attack or effort to take manufacturing data.
How can co-innovation help companies keep pace with digital transformation (DX)?
Cabanes: Co-innovation helps vendors and IoT users keep pace with DX in industry because let’s face it, customer expectations are such that what once took industrial suppliers many years to deliver must now be available in months. So companies must leverage digital capabilities as never before. Evolving to go beyond a product-driven business core (primarily focused on customer usage and expected product attributes) to offering products that include support services and operating platforms that customers say they want. Design approaches must also go beyond the linear step-by-step processes of the past to more adaptive processes for active design and prototyping. That speeds up test-fail-test-fail-test-success cycles which in fact go beyond siloed in-house designs to leverage open ecosystems for more holistic innovation as well. That in turn makes for competitive advantage.
Perducat: Co-innovation has been key for us over the last several years — via an expanding network of collaborating companies and Schneider Electric’s IoT-enabled EcoStruxure™ Platform. We’ve also worked with other companies including Accenture on the Digital Services Factory — an internal, “virtual” factory that enables Schneider Electric to rapidly build and scale new offerings in areas such as predictive maintenance, asset monitoring, and energy optimization and cut time to market. Part of the work on this front is examining specific business problems driven by customer needs and identifying early DX adopters interested in updating their operations — and then trying and iterating only solutions that are relevant to their business. It’s key for these projects that the co-innovation is rooted in a concrete business context.
Schneider Electric uses similar approaches in its involvement in other ecosystems — with thousands of partner companies. Consider how in the industrial field, we have more than 3,000 system integrators. We design software with and for those partners — and not just for end users. Close relationships with such customers let us analyze their business problems and deploy relevant solutions. Innovation here focuses on making solutions that are easy to sell, deploy, and install. That extends to our end-to-end IoT EcoStruxure architecture. With it, we’re not trying to be everything for everybody. We’ve only integrated the best solutions for top partners … with most work on improving integration, customer experience, and ease of use.
We’re also not going it alone — and are partnering with multiple companies … including a range of organizations such as Microsoft and other large companies to small startups.
The aim has been to get the right pieces of technology and architecture to ensure the excellent customer experience with sophisticated deployments. Read more in “The Value of IoT for Manufacturers” eBook we developed in conjunction with IDC.
Also go to designworldonline.com/webinar-iiot-manufacturing-start-small-big-business-results to watch an on-demand archived webinar on this topic.
You may also like:
Filed Under: ALL INDUSTRY NEWS • PROFILES • COMMENTARIES, ALL INDUSTRIES, AI • machine learning, Factory automation, WEBINARS, Manufacturing, Cybersecurity, IoT • IIoT • internet of things • Industry 4.0