"At the critical moment, Wang Jinxi threw away crutches and jumped into the mud pool regardless of his leg injury. He used his body to mix the mud and finally suppressed the blowout."

The dedication spirit of the Iron Man Wang Jinxi has been the motivation of our generation. However, such scene may not appear today because we have the power of IoT + AI.

At Shengli Oilfield in the Yellow River Delta, the exploitation equipment has been equipped with sensors to monitor the production in a real-time manner. However, these sensors are vulnerable in harsh environment, and any single failure may affect production safety. With the assistance of iSoftStone, Shengli Oilfield applies AI technology for intelligent analysis of the operation of oil well sensors, which gives pre-alerting of the possibly damaged sensors and significantly improves the operation and maintenance efficiency.

ISoftStone has many similar cases of intelligent application covering many sectors and application fields. The company is empowering digital transformation of all walks of life while AI is one of its core strategies.

AI implementation needs to be led by consultancy

In the past IT was an auxiliary tool of production to help producers do better and improve work efficiency. Therefore, only the IT department of an enterprise had explicit IT needs. The system integrators and software developers, although called solution providers, did not need to provide full and complete solutions, and all they did was hardware integration or software development as per customers' needs.

It was changed by the advent of digital transformation. IT becomes the core system of enterprise and assumes the major responsibility of business innovation of customer service. Now the IT needs come from the management or business departments who only care about the value creation ability of IT instead of the technology itself. In such a context, the solution provider is not only a capable executor but a consultant to provide advisory service.

Given the emerging new technologies and applications like cloud computing, IoT, big data and AI, the customer needs are about business rather than technology, without a clear indication of work. Hence the solution provider must have all-round digitalization capabilities to provide all-covering services from strategic consultancy to design and implementation, system migration, digital empowerment, technical support, etc.

Therefore iSoftStone's richness of technical capabilities and industry insight became one of its core competitiveness. Over the past 18 years, iSoftStone served over 1,000 domestic and foreign clients across a dozen of key industries, including over 90 world top 500 enterprises. Since the implementation of strategic transformation in 2013, iSoftStone enlarged investment in the emerging technologies of cloud computing, big data, mobiles internet, IoT, AI, etc. When technology meets demand, iSoftStone generates endless power to the business innovation of clients.

AI is more than a technology, but a capability.

As a new technology, is AI an application or a capability?

This seemingly philosophical question is a practical matter that decides the AI strategy direction of solution providers.

In our previous perception, we took AI as applications e.g. voice recognition, intelligent speaker, robot customer service, auto pilot, etc. because the previous AI applications were mostly for individual consumers.

Nowadays with the execution of more AI industrial applications, we start to find that AI is not only an explicit application, but a capability integrated in various applications.

For individual consumers, AI is indeed an application; but for enterprises, AI is a tool to realize intelligent business that becomes a capability of enterprises. Such implicit capability might not be so intuitive, but is of great importance to business decisions.

For example China National Petroleum Corporation (CNPC) used to apply traditional methods to locate failure points in the optical fiber communication network, and the scope might extend dozens of kilometers, indicating a huge workload of troubleshooting. The intelligent operation and maintenance platform of fiber communication developed by iSoftStone for CNPC utilizes a neural network algorithm to build a machine training model that extracts characteristics, analyzes and predicts the collected data. It reduces the locations scope of failure point within 20m with judgment of failure cause, which significantly improves the efficiency of patrol inspection and lowers the workload.

In the above case, AI technology is not explicitly shown but becomes a capability integrated in the application system to play a key role in the business of client.

ISoftStone not only creates AI capability for clients, but integrates AI into their solutions that ultimately empower the clients of all industries.

AI capability building based on greater middle-end strategy。

It has become a common understanding that enterprises must have their own middle-ends for digitalization. The middle-ends, including technology middle-end, business middle-end and organization middle-end, can act as a variable gear to achieve the balance between frontend and backend.

A complicated and oversized frontend system is unable to deal with changeable market demand, which has been a universal issue of enterprises. They can utilize the middle-end to streamline the frontend system and transfer some common capabilities to the middle-end so as to make the frontend more flexible. At the same time, some direct capabilities of backend may be uplifted to the middle-end to better support the frontend business.

The AI implementation also needs the middle-end. The bottom layer of corporate-level AI application is the basic layer dominated by computing resources, the middle layer is the technical layer in the form of algorithm, and the upper layer is the specific industrial applications.

AI is seemingly popular among all sectors but quite a number of AI applications are superficial—with poor user experience. The root cause is the absence of technical level or "middle-end", i.e. AI for AI's own sake.

Going straight to the heart of the matter, iSoftStone selects the technical level as the icebreaking point, and extends to the application layer. Based on years of accumulation in knowledge spectrum and deep learning, iSoftStone launches the smart brain platform including two major functions of data model training and knowledge computing that can perform the intelligent R&D tasks covering both perceptual intelligence and cognitive intelligence of enterprises. ISoftStone's smart brain is centered on greater middle-end system integration that combines the computing resources at downstream and complies with the business logic at upstream so as to create the AI capabilities for enterprises.

Popularizing AI in traditional industries

As the traditional industries accelerate the pace of digital transformation, AI is popularizing across industries. The users of traditional industries need AI as decision support but lack the relevant talents and abilities where iSoftStone can come into play.

The abundant accumulation of algorithms and integrations enable iSoftStone to implement and popularize AI in traditional industries. The company boasts at least three advantages in AI market: first the engineering capability gained over the past two decades guarantees the project deliverables; second the deep understanding of new technologies guarantees the competency of providing consultancy service to customers; third the serial AI tools are directly available to customers.

On account of AI middle-end integration and commercial intelligent service, iSoftStone begins to implement the AI application solutions in various sectors. The company categorizes these applications into four types, namely "production, cost reduction, troubleshooting and casualty avoidance". For "production", the intelligent operation and maintenance of production facilities can increase revenues of enterprises; for "cost reduction", the intelligent monitoring of energy conservation and emission reduction can reduce costs for enterprises; for "troubleshooting", the IT intelligent operation and maintenance can help improve production efficiency; for "casualty avoidance", the intelligent monitoring of production safety can prevent occurrence of safety incidents.

All these applications are built on a powerful AI middle-end, which better integrates AI capabilities into different industries to support all business aspects of strategic decision, production and operation, marketing and promotion, supply chain management, etc. for enterprises.