Cutting edge technology

Expanding seven techniques for implementing an artificial in

Introduction: Pilot projects for artificial intelligence (AI) technology surged in 2018 as many companies tested machine learning (ML) algorithms and a range of automated tools to consolidate relationships with customers and improve network operations. Or enhance their network security environment.

Now that your company has experimented with artificial intelligence, it's time to consider how to extend these efforts. According to Andrew Ng, an expert in artificial intelligence forecasting, and experts from PricewaterhouseCoopers (PwC) and Deloitte, this should be done in this way.

Pilot projects for artificial intelligence (AI) technology surged in 2018 as many companies tested machine learning (ML) algorithms and a range of automated tools to consolidate relationships with customers and improve network operations Work or enhance their network security environment.

Inspired by initial results, CIOs are preparing for the next challenge: expanding the use of artificial intelligence across the enterprise. According to a recent study by PricewaterhouseCoopers (PwC), 20% of 1,000 US corporate executives said their company plans to deploy artificial intelligence throughout the enterprise by 2019.

Business aspirations are soaring. Companies are investing more in these emerging technologies, and according to the Internet Data Center (IDC), spending on cognitive and artificial intelligence systems will reach $77.6 billion by 2022, the $24 billion forecast for 2018. More than one.

But no matter how big the company's ambitions are, the road to expanding the use of artificial intelligence is full of dangers, such as strategic conflicts and shifts in business priorities, which can stifle cross-sectoral cooperation. The lack of talent to handle technical work complicates the issue.

Here, artificial intelligence experts from PricewaterhouseCoopers, Deloitte and some start-ups provide some important recommendations that should be considered when CIOs are expanding their implementation of artificial intelligence initiatives.

Establishing an internal artificial intelligence team

Andrew Ng, founder and CEO of startup Landing AI, said that to ensure the support of senior management, companies should establish aThe artificial intelligence team, which will help the project to be carried out within the company. An artificial intelligence team can be led by a chief technology officer, chief information officer, or development director (chief digital officer or chief data officer), or even a lead artificial intelligence officer. Forming such a team will help recruit and retain talent.

"With a new artificial intelligence team, you can connect artificial intelligence talent to different departments to drive cross-functional projects," Wu Enda said in an artificial intelligence transformation guide published last December. “There will be new job descriptions and new team organizations.” Wu Enda has led the artificial intelligence team at Google and Baidu, where he recruited machine learning engineers, data engineers, data scientists and artificial intelligence product managers. However, Wu Enda admits that the current competition for artificial intelligence talents is “zero sum in the short term”. Companies must work with recruiters to find talent in key positions.

Guiding “citizen artificial intelligence” employees and artificial intelligence experts to work together

The lack of artificial intelligence talent should not stifle artificial intelligence plans. Instead, companies should take advantage of tools that democratize artificial intelligence and data science, including providing user-friendly interfaces for artificial intelligence developers and educational programs for non-technical experts.

Companies can divide employees into three levels: citizen users who will learn how to use artificial intelligence-enhanced applications; citizen developers or advanced users who can identify use cases and data sets, and work with people Smart experts work closely together to develop new artificial intelligence applications; data scientists who will take on the responsibility of developing, deploying, and managing artificial intelligence applications, Scott Liken, PwC's new services and emerging technology lead Scott Likens said that he is also the co-author of the entire enterprise's report on extending the use of artificial intelligence. This work will require skills improvement to narrow the talent gap.

Establishing a Center of Excellence

In a report by PricewaterhouseCoopers, Likens said that one of the best ways to create an artificial intelligence foundation is to build an Center of Excellence for Artificial Intelligence (CoE). The organization will identify technical standards, architecture, tools, technologies, suppliers and intellectual property management, and will determine how to identify use cases and how to develop accountability and governance.

For example, energy giant Shell has established a data science center of excellence that uses artificial intelligence, machine learning and analytics to solve project issues such as predictive maintenance of oil rig parts. According to PricewaterhouseCoopers, 24% of respondents have established some form of AI Center of Excellence.

Adding Artificial Intelligence Strategies Through Experiments

Although it may be attractive to develop an artificial intelligence strategy immediately, Wu Enda said that most companies can only develop comprehensive knowledge after having some technical experience. Artificial intelligence strategy.

Wu Enda suggested building several difficult artificial intelligence assets. These assets generally have a consistent strategy, but they need to be tailored to form an advantage in the industry, which makes it difficult for competitors to replicate. . This requires sophisticated data analysis strategies to foster business insight.

For example, Keller Williams relies on thousands of carefully crafted data points and machine learning software for homes to improve its listings, the company's chief product officer. Neil Dholakia said. The real estate agent uses the Keller Williams app to record a video of the house on a smartphone, and the app connects to Google's Cloud AutoML software. The software instantly identifies and marks hardware features such as floor or granite countertops.

Dholakia told reporters, “This process can take anywhere from a few days to a few minutes, and our agents can use it for free.” He highly valued the potential of machine learning technology to provide competitive advantage in the industry. And said he plans to expand the company's use of artificial intelligence technology in 2019.

"Artificial intelligence strategies will guide your company to create value while also building a defensive moat," Wu Enda said. “Once the team begins to see the success of the initial artificial intelligence project and develops a deeper understanding of artificial intelligence, you can determine where artificial intelligence can create the most value and focus resources on these areas.”

[ 123] Building Responsible Artificial Intelligence

One of the main obstacles to using artificial intelligence is to explain how artificial intelligence models make their decisions, which is a prominent problem in regulated markets such as finance.Cathy Bessant, chief operating and technical officer at Bank of America, said at the recent New York Artificial Intelligence Summit that this is an important reason for creating a transparent artificial intelligence model.

Organizations can address these "black box" questions by answering the following questions: Can the organization ensure that these decisions are accurate? Who is responsible for the artificial intelligence system? Is there proper compliance control?

Successful artificial intelligence deployment establishes accountability for all of these factors to build "responsible artificial intelligence."

Practice participatory design or people-oriented design

How to build responsible artificial intelligence? According to Deloitte's latest report on the state of artificial intelligence in the enterprise, the first step that stakeholders should take is to adopt a practical approach to designing complex artificial intelligence implementations.

Participatory design – a people-centered design – embeds the needs of the user “group” directly into the design process to design a more sustainable solution. This allows designers to be aware of and avoid unpredictable conditions due to problems with context or imagination.

For example, if the call center deploys a chat bot to reduce employee workload, the participatory process involves customer service center employees, leadership team members, and customers who may interact with the chat bot.

To ensure that artificial intelligence is based on ethics, companies should conduct participatory design through “periodic review and evaluation of algorithms to ensure that algorithms operate correctly”, Deloitte's head of global data risk and Analysis director Vic Katyal said. Finally, companies should allow third parties to independently verify artificial intelligence, which will help fill gaps and overcome work blind spots.

Is it good news? Katyal said that CIOs are the most common senior executives who are responsible for managing the use of artificial intelligence in the enterprise on behalf of the company's board of directors.

Develop communication strategies

Since artificial intelligence will have a significant impact on the business, companies should develop communication plans to ensure consistency across all aspects. This work will cover investor relations (a theory that explains the value created by artificial intelligence); government relations(if necessary); customers (think strategic marketing); talent (brands are essential for attracting fresh blood); and internal communication.

"Because people still know very little about artificial intelligence, especially the general artificial intelligence is over-hyped, people have fear, uncertainty and doubt," Wu Enda said. “Many employees are also worried that their work will be automated by artificial intelligence. Clear internal communication can explain both artificial intelligence and employee concerns, which will reduce internal reluctance to use artificial intelligence.”

Corporate net income

Most executives are optimistic about the prospects of artificial intelligence. 56% of the 1,100 IT and business executives surveyed by Deloitte said that artificial intelligence will change their business within three years.

"The arms race in artificial intelligence/analytical technology will continue as companies need to be lean, agile, and focused on growth," said CIO Consultant and strategic advisor to Fractal Analytics, Andy Walter ( Andy Walter) said. “The use of artificial intelligence leaders in targeted business processes will increase the value-driven opportunities for the entire enterprise. ‘Intelligent companies’ will beat competitors in both total sales and net income.”