Although people are talking about artificial intelligence ( AI ), it may still be difficult to find out how technologies such as machine learning can really benefit your business.
A new Microsoft report entitled ” Maximising the AI opportunity” shows that nearly two-thirds of business leaders do not understand the potential rewards of using ” AI”.
” By default, artificial intelligence is often associated with high or unreasonable expenditures,” the report said. The report surveyed 1,000 business leaders and 4,000 employees in Britain. This suspicion has been supported by some industry observers, who have emphasized the gap between the prospects and reality of artificial intelligence-related technologies, from chat robots to self-driving cars.
Despite these doubts, the report said that its survey found that ” enterprises that have embarked on the journey of artificial intelligence have increased their productivity, performance and business results by 5% over those that have not.”
Michael Wignall, Microsoft’s UK chief technology officer, said: ” The companies we are investigating have already done some things, even small things with low costs, and they have seen real benefits from them.”
According to Wignall, the technologies the company is implementing related to artificial intelligence include everything from web chat robots to custom machine learning models that predict based on manufacturing data.
The following are the five steps pointed out in the report to promote artificial intelligence ( AI ) in enterprises.
1. Identify business issues that need to be addressed
First, identify the business problems that need to be solved, and then evaluate each problem to see if they are applicable to existing AI – related technologies, such as chat robots for handling simple customer queries, or Robotic Process Automation, RPA ) for some back office roles ( there is a dispute over whether RPA’s rule-based approach should be classified as a form of AI, but this report mentions RPA as an AI – related technology ).
Cindy Rose, chief executive officer of Microsoft UK, wrote in the report: ” Like many other business problems, overcoming this inertia requires first finding the business problems that need to be solved.”
For example, is it necessary to improve the efficiency of management tasks such as wages and invoices? If so, then the RPA solution may be the correct answer. Is it through chat robots or automated telephone systems to improve customer experience? Is it necessary to use machine learning to deal with more mundane or direct parts of work, thus freeing up staff’s time and allowing them to engage in creative work? Are the above business problems to be solved?
In the report, Centrica, a British energy company, said that it was trying to solve the business problems of fast inquiry of the target customers. For this reason, it set up a natural language robot to help call center personnel to provide them with the best customer support information and update it to our back-end system.
2. Determine whether your enterprise is ready to build, manage and support systems related to artificial intelligence
Once appropriate business problems are found, companies need to check whether they are ready to build and manage the system of their choice.
There are various machine learning services from major cloud service providers, such as on-demand image and voice recognition services, building custom models and tools, and systems using GPU and machine learning frameworks.
The report said that one of the key problems in machine learning is whether the correct data are captured and whether the data are properly processed so that the training model can make useful predictions.
Microsoft’s Wignall said: ” What we mean by good data is that it has both quality and quantity.”
” You need a lot of data to better train artificial intelligence models. The more data you can get, the better, whether it’s customer interaction, Internet of Things or sensor data.”
” But in fact, it is also a matter of quality. Make sure the data meets the purpose you want.”
” Before implementing any technology, you need to think about why you want to collect these data and what you do with them.” Capturing every piece of data indiscriminately is no longer applicable to AI. The report found that many companies have begun to collect and clean up data, and more than one-third of the companies said that their enterprises are already using tools such as predictive data analysis and data integration.
3. Mark the core skills and those missing skills in your enterprise.
Once the scope of the business problem is determined and the potential technologies and data required are identified, the next step is to determine which internal skills you have to implement the project.
The report recommends that a road map of available skills and skills needed in the medium to long term be drawn up before finding ways to build any missing skills, including using in-house skills programs, recruiting and working with partners.
These are not only the skills needed to realize the project, but also those needed by employees whose jobs have changed due to the use of artificial intelligence-related technologies such as chat robots in enterprises.
Microsoft has found that about one-third of business leaders admit that they are not sure how to provide employees with the necessary skills to cope with the chaos caused by artificial intelligence reshaping its role.
Although nearly half of the employees feel that they can master new skills that are important to their jobs, only 15% of the employees said that their company is helping to master new skills. In addition, only 18% said they were actively learning new skills to help themselves keep up with the future changes brought about by artificial intelligence.
This unprepared discovery coincides with a report by the British Human Resources Association ( CIPD ). The report found last year that the UK’s spending on training in 2005 has been declining, and in recent years, the participation of job-related adults in learning has dropped significantly.
4. Cultivate a culture in which employees can test and evaluate artificial intelligence.
Microsoft found that employees and business leaders are willing to try technologies related to artificial intelligence to help them finish their work. 67% of leaders and 59% of employees said they were open to the idea.
However, employees may need some encouragement before using these new technologies. Newcastle City Council has been experimenting with various robots to handle simple interactions with the public, recommending some simple steps.
Jenny Nelson, manager of the Digital Transformation Project, wrote in the report: ” Part of the reason is to choose the right project to test.”
” Starting from small things and expanding the scale will help the team build trust, gain feedback, learn lessons and build confidence.”
Microsoft’s Wignall said: ” The employees we surveyed are more and more open to the use of artificial intelligence. From the perspective of employees, there is no deep-rooted resistance. However, they may not have the skills to take advantage of this change, and if their existing skills are not the right ones for the future, they may also worry about what will happen to their jobs.
He suggested ” cultivating a learning and development environment from bottom to top, allowing people to think about how their work may change with the support of artificial intelligence”, ” investing in skills” and encouraging ” continuous learning and development”.
Don’t forget prejudice and morality.
As mentioned above, the quality of machine learning systems depends on the data they train.
The report pointed out: ” If these data are not representative, biased or completely wrong, the way machine learning uses these data will be fundamentally flawed.”
” Only by helping engineers eliminate data blind spots around factors such as gender, race, ethnicity and socio-economic background can we ensure that artificial intelligence technology brings the just and responsible social results we want.”
The report recommends that a declaration on artificial intelligence be drawn up, setting a framework for the ethical use of the technology, protecting data privacy, preventing malicious abuse of artificial intelligence, and setting clear guidelines on inherent bias, automation, and where to take responsibility in case of errors.