The goal of applying Artificial Intelligence (AI) (*) to Internet of Things (IoT) systems is effectively placing additional layer of intelligence across the entire IoT stack – from infrastructure all the way to applications. Edy Liongosari, Global Managing Director and Chief Scientist of Accenture Labs, points out that if we focus on the lowest layer of the infrastructure, AI is used, primarily, to create sensors that self-calibrate or self-heal when the IoT network or an individual sensor fails, identifying them by proximity, or even creating new kinds of “virtual sensors”; or through the use of artificial MRI (Magnetic Resonance Imaging) vision, a non-invasive image visualisation technique which is used, for example, to detect cancer cells and reproduce them as three-dimensional images.

In the highest tier of the IoT framework, the application, Liongosari stresses that artificial intelligence can be applied to a wide variety of functions “from assessing the predictability of future events, to maintenance or security issues.

“In this regard, the Global Managing Director and Chief Scientist of Accenture Labs points out some examples: “Through the use of AI, we can identify who is authorised to use certain equipment and how qualified they are; offering context-based services (e.g. additional capacity during heavy usage), operational optimization (e.g. re-planning/re-scheduling due to supply chain interruption), to new user interaction (e.g. combination of gestures, voice and face to understand the user’s context and needs)”.

With the aim of understanding how AI helps all kinds of industries add more efficiency to their IoT infrastructures, it is beneficial to take a step back in time to see how the use of IT has evolved in the industrial environment. Wael William Diab, Senior Director of Huawei, member of the Industrial Internet Consortium (IIC), and an artificial intelligence expert who regularly speaks at the IoT Solutions World Congress of the Fira de Barcelona, recalls that, initially, information technology applied to industrial uses were seen as tools that increased efficiency within organisations.

The use of IT in industrial processes managed to improve the processes through the integration of (**) KPIs (Key Performance Indicators) to measure the performance of an activity through constant monitoring of the platform, comparing the results with the parameters established by the management team. If we limit ourselves to the scope of IoT, IT has been integrated into the management chain much further even reaching areas related to the decision-making process, and including more traditional industries that had no connection to IT in the past,” states Diab.

In his opinion, artificial intelligence today drives a new change in IT “providing knowledge that will serve to establish future objectives and KPI elements. In other words, AI has managed to get a seat at the management table, adding its voice to where the organisation should arrive through knowledge”, claimed the expert.


From a technological point of view, it is important to remember that artificial intelligence is composed of a set of technologies such as machine learning (ML) and deep learning (DL). “Artificial intelligence, the Internet of Things and analytics form three facets of the same reality”, said the head of Huawei, Wael William Diab, who also indicates that:

“While IoT focuses on networks of sensors that generate data, analytical processes are limited to the analysis of such data with the objective of creating value, whereas artificial intelligence enables the generation of knowledge and predictability using the valuable data that has been gained”.

Both Liongosari and Diab agree that the wide applicability of AI enables organisations to integrate analytical data mechanisms into practically all types of industrial sectors. However, Liongosari warns that the most important value that can be drawn from AI depends, to a great extent, on the applications themselves.

“AI has a very important role in the improvement of operational efficiency, which is a major part of the current implementations of the industrial Internet of Things from predictive maintenance, through to logistics until arriving at process optimisation.”

Studies on the impact of the AI in predictive maintenance, such as the one carried out by the firm Anodot, stress that this ability will save organisations between $240B to $630B by 2025 due to the reduction of downtime and costs associated with maintenance processes. At the other extreme, this set of technologies has the ability to generate new business models, new sales channels, better services and a superior user experience. In this regard, the Michelin ‘tyre-as-a-service’ strategy is an example of a transformation of the business model of a traditional industry through IoT and AI.

In aspects related to the application of predictive analytics in different industries, Diab recommends using the term “industrial” in a way similar to how the Industrial Internet Consortium (IIC) does when it comes to covering different sectors, instead of focusing only on the manufacturing processes.

“IIC has recently published the study Industrial IoT Analytics Framework (IIAF), which serves as a guide for detailed information and assistance on issues related to the development, documentation, communication and deployment of infrastructures of the Internet of Things and Analytic IIoT systems and industrial leaders who are committed to integrating analytical systems that contribute value to the business by being better informed when it comes to making decisions”.

In terms of standards, ISO/IEC JTC 1 recently established ISO/IEC JTC 1/SC 42. ISO/IEC JTC 1/SC 42 is a first of its kind standardization committee looking at the full AI IT ecosystem. Wael was appointed as the Chairman of this committee. In addition, there are other Standards Developing Organizations (SDOs) that are working on AI related projects and there a number of other partnerships and alliances looking at AI.

Standardizing AI will benefit the industry and lead to a new era of growth because companies are going to make important investments in AI solutions and should be sure that they can continue to work and evolve those systems for long.

(*) Cognitive systems and artificial intelligence are synonymous terms for the purpose of this document, while machine learning (ML) and deep learning (deep learning DL) constitute enabling technologies which form part of artificial intelligence.

(**) A KPI (key performance indicator) is a measurement of the level of performance of a process. The value of the indicator is directly related to a pre-set objective and it is usually expressed as a percentage. A KPI is designed to show how a particular aspect is progressing; therefore it is an indicator of performance. There are KPIs for various business areas: purchasing, logistics, sales, customer service, etc. Large companies have KPIs that show whether the actions taken are producing results or, on the contrary, they are not progressing as expected.

AUTHOR: Marga Verdú

SOURCE: IoT Solutions World Congress / Fira Barcelona

Pedro Mier

Pedro Mier holds a degree in Telecommunications Engineer ing from the Polytechnic University of Catalonia, MBA from ESADE and PADE from IESE. He is currently President of AMETIC (Association of Electronics, Information Technology and Telecommunications Companies of Spain), Shareholder and Chairman of the Board of Directors of TRYO Aerospace & Electronics, Board Member of the Premo Group and Committee of CTTC. member of Space Angels Network and Member of the Sc ientific Advisory