Smart Connected Products with AIoT and Digital Twins

Revision as of 07:16, 20 May 2022 by Admin (talk | contribs) (Created page with "Products are becoming increasingly intelligent and networked, thus combining AI and IoT in the form of "AIoT". The "digital twin" serves as a mirror image of reality, makes co...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

Products are becoming increasingly intelligent and networked, thus combining AI and IoT in the form of "AIoT". The "digital twin" serves as a mirror image of reality, makes complexity manageable and provides the foundation for intelligent decisions and actions. Products thus become "Smart Connected Products". This development is driven by the markets and opens up attractive prospects for the companies offering them. However, the transformation to a successful digital company is by no means trivial. Smart connected products are highly complex and present new challenges for the organization. How can companies adapt to this and successfully master the change?

Introduction

Artificial Intelligence (=AI), the Internet of Things (=IoT) and Digital Twins are three technological approaches that we can hardly imagine our lives without. Increasingly, the combination of these approaches - AIoT & Digital Twins - can be observed in the form of intelligent networked products and solutions. This is referred to as "Smart Connected Products": networked trains whose tachographs provide data on driving behavior and fuel consumption, so that train drivers receive advice on energy-saving driving methods. Positioning aids in the networked CNC milling machine, which use AI technology to ensure that workers do not have to laboriously reposition components on the machine. Networked electricity meters whose data, in combination with data from other sources (weather data, event data or similar), enable more precise forecasts of energy load peaks.

AIoT and digital twins are also on the rise in cities - in so-called "smart cities." For example, the city of Vienna is pursuing three major key objectives as a result: Resources, innovation and quality of life. It is relying on networked sensor technology to make buildings more energy-efficient. Intelligent traffic lights make it possible to detect approaching pedestrians at an early stage and thus increase safety. A digital twin of the city created by surveying public space will make it possible to approve events more quickly and optimize "street furniture" from bicycle stands to kiosks. With the spread of wireless standards such as LoRaWAN, NB-IoT and 5G, the vision of the smart village is now within reach. Not far from Vienna, IoT applications in agriculture, infrastructure and tourism are therefore being tested in Lower Austria as part of the Dataskop research project.

External drivers and great potentials lead companies to implement AIoT and Digital Twin in the form of Smart Connected Products

What drives companies to launch products based on AIoT and Digital Twin? Why are these companies changing their business model? The basic answer to this is: to secure and expand their own competitiveness. A more detailed rationale for this can be answered on the basis of three exogenous drivers observable in the market: increasing competitive pressure, changing customer needs, and ongoing technological development.

Increasing competitive pressure: As a result of globalization, which has been increasing for a long time, new competitors (e.g. from Asia) are increasingly opening up existing markets, which is reflected in increasing price pressure and decreasing margins. This can be observed above all in the material goods business. In the service business, on the other hand, rising margins can be seen, as a result of which service-based business models / offerings are becoming increasingly attractive for all types of companies.

Changing customer needs: Customers no longer buy off the shelf (and not every customer wants their "Ford Model T" to be black). Demand is becoming more customized and comprehensive in terms of the solution customers expect. Figuratively speaking, customers want the clean, correctly placed drill hole and not the drill. This requires more complex products / solutions, which must also ensure overall integration into the respective system. In addition, customers increasingly no longer want to make advance payments and make large one-off payments in advance. As a result, the demand for flexible and predictable / service-based payment models is increasing (see Spotify, Netflix, etc. in the B2C market).

Technological development: What used to be costly and complex or could only be anchored in the product architecture with special internal know-how is now much more accessible: components for networking, sensors, actuators or computing components. These things are now significantly cheaper to procure, easier to integrate into existing systems, and controllable by a growing number of available experts. Applications in the field of analytics and AI are also available on the market in scalable form (e.g., on cloud-based platforms) and can be deployed at short notice.

These external drivers are not the only ones motivating companies to launch smart connected products on the market. In addition, positive effects can be observed within successful digital companies through the introduction of AIoT and Digital Twins in the form of Smart Connected Products. These serve as an incentive for companies that do not yet have a digital product portfolio.

Forklift AIoT

Smart Connected Products enable better and individualized customer interaction based on a deeper understanding of the customer. This results from the stream of usage, process or fleet data that has been brought into focus. Based on this, essential questions for product development can be answered, such as: On which input screen does the customer have to search too long for the right element? Which tasks keep the customer busy most of the day, even though they really shouldn't?

The business case is positively influenced by effects such as increased productivity, reduced costs or the development of new revenue sources. This can be attributed, for example, to improved utilization of production facilities, maintenance optimization or data-based additional services.

Smart connected products also offer the opportunity to realize new high-margin, service-based business models with recurring revenues. For example, the use of a machine can be tracked precisely and the business model can change from the classic one-time purchase to usage-based billing (keyword: Everything as-a-service).

Smart Connected Products consist of physical components, cloud AI and integration into the customer environment.

So if there is a desire to evolve the product portfolio towards AIoT & Digital Twins, the basic structure of Smart Connected Products should be understood first. All Smart Connected Products combine three basic characteristics: they consist of a physical component (so-called "edge device"), have a digital representation in the IT backend (the "digital twin" in the "cloud") and are integrated into the customer environment. Together, the properties serve the overarching goal of providing the greatest possible added value to the customer. This concept can be vividly described using an intelligent, networked forklift truck.

AIoT Vision

As a physical component or edge device, the forklift has its own computing power and intelligence. It also records relevant status information and is clearly identifiable. It becomes a "networked" product because it has a communication module to ensure data exchange with the IT back end.

The forklift truck becomes a "smart" networked product through its representation in the IT back end, for example in the form of a cloud platform. This is the home for the virtual image, the so-called "digital twin". This is permanently analyzed using machine learning and artificial intelligence methods. From this, derivations can be made for control, process optimization or interaction with the fleet.

The third perspective is integration into operational processes: For example, the autonomously operating forklift can optimally support logistical processes. In intralogistics, the centrally supported assistance system can ensure that the correct shelf location is automatically approached. Or that the routes in the warehouse are optimized through analytics-supported allocation of shelf locations. Last but not least, safety functions (such as collision avoidance) can be implemented via AIoT features. The context-sensitive exchange of data with customer systems solves user problems in a targeted manner, thus fully exploiting the great potential of AIoT and Digital Twin.

To master AIoT & Digital Twins, product development must be fundamentally adapted to the complexity of different product development cycles

It quickly becomes apparent that Smart Connected Products have a high degree of complexity due to the integration of the three elements - networked physical component, cloud backend with analytics component and connection to customer systems and processes. This means major changes for the product development of a company that has not previously developed Smart Connected Products. Success factors in two dimensions in particular can be identified in practice: First, the different development cycles of the components must be coordinated. Secondly, new competencies must be built up or otherwise made available for mastering all the required disciplines.

In the traditional development of products that are not networked and do not have an AI component, development is often based on the V model. This means that at the beginning of development, requirements are defined for the entire product and broken down to all subsystems and components. The entire product is then developed. At the end of the development the verification of the individual components takes place the integration to subsystems and to the total product, which is finally validated. From requirements definition through to market launch, all elements of the overall product therefore pass through one and the same product development cycle. The organizational structure behind this is also structured accordingly, for example according to functional separation such as mechanics, software or electroplating.

In the case of Smart Connected Products, on the other hand, the challenge is that the combination of AI and IoT changes the value chain and the product architecture components required for this: In addition to the hardware component, connectivity components, messaging/bus systems, IoT platforms, analytics/AI components, user applications, and additional service and integration services are required.

These components have different product development cycles: Hardware and connectivity tend to go through one cycle, messaging/bus systems and IoT platforms already go through several cycles, and analytics/AI components and applications go through many cycles. This is due to the long- or short-lived nature of the components. Due to physical processes and dependencies, a hardware component is usually defined once in the form of requirements and developed and produced based on these. An application, on the other hand, can be tested and further developed continuously and very quickly with test users - even before being rolled out to the end customer. AI components, in which AI models and their training play a central role, are constantly evolving due to continuous machine learning.

Accordingly, the interfaces between these components must be defined in such a way that they remain constant and are coordinated with each other over the entire development period and beyond. This requires interdisciplinary teams and permanent exchange across component boundaries.

Essential for the success of this type of development, in addition to the product architecture and mastery of the interfaces, is the coverage of the required disciplines. In addition to traditional skills such as mechanical or material experts, these include new skills such as UI/UX experts, data scientists, solution designers, system architects and connectivity experts. Since these are not core competencies for many companies that were previously active in the traditional hardware product business, the development of competencies must be strategic. This means that, in addition to insourcing, other options such as outsourcing, partnering or coopetition must also be examined and implemented.

Successful business with Smart Connected Products means not only adapting development but also adapting the market approach, the development strategy and the entire organization

The transformation from the traditional hardware product business to the business with Smart Connected Products requires success even before the actual development: the market approach with a corresponding digital target image with customer-centric, flexible business models. This is the only way to achieve the major advantage of Smart Connected Products, which lies in providing the greatest possible added value to customers - for which they are prepared to pay high prices.

The high complexity of smart connected products also requires a rethink of the product strategy, in which the provider should focus above all on a scalable architecture and control and sovereignty over the valuable data stream. It is also important to think in terms of an ecosystem and, if necessary, even to outsource parts of the value creation process.

For the design of the organization, this means a different form of sourcing: co-creation. The company thus slips into more of an orchestrator role. For this, it is important to involve operations early in the development process, which is often reflected in interdisciplinary DevOps teams in practice. The organization must also adapt at the end of the value chain, in that sales must reorient itself from pure product sales to customer success management.

These points are described in more detail in the context of ten paradigms for successful Smart Connected Products:

SCP Paradigms

Holistic approaches with a focus on the greatest possible customer added value are the key to success with Smart Connected Products.

So to be successful in the market with Smart Connected Products and make the transition to a digital company, companies need to work on many areas of their strategy, organization, processes and competencies. Accordingly, holistic approaches that encompass the entire product lifecycle are recommended wherever possible. The AIoT User Group shows an approach that has proven itself in practice. This is set out in the Digital Playbook and is constantly being further developed by the User Group in close exchange with numerous industrial companies. But regardless of which approach is chosen, it should be holistic. This is the only way to set the right focus points for the necessary transformation, strategically absorb the technical complexity and ensure customer-oriented solution design. This is the only way to generate the greatest possible added value for customers in the long term.

DT Quotes

This article was originally published in German by mm1 at Informatik Aktuell on 28.12.2021.; authors: David Monzel & Heiko Loeffler