Can AI Benefit?
Recently, artificial intelligence (AI) has been more discussed in the media than other technologies. At the same time, the technologies underlying it and the benefits that it can bring to business are not clear to anyone. And this is not only Russian, but a global trend. In 2019, analysts at the MIT Sloan Management Review and BCG conducted a study that surveyed over 2,500 CEOs in 27 industries around the world. It turned out that business is still doing badly to make money on AI: 70% of businessmen said that the introduction of AI did not affect their business in any way, and only 40% of respondents managed to get at least some profit.
According to the latest researchVTsIOM 69% of Russian companies note a shortage of skilled personnel in the field of AI. At the same time, large and high-tech corporations are realizing the benefits that can be obtained now by introducing AI-based solutions into their processes. But even big players lack their own expertise to work with these technologies.
The importance of the issue is also understood at the state level. In October 2019, the President issued a decree on the development of artificial intelligence in the Russian Federation, which proposes to approve the national strategy for the development of AI until 2030.
While officials are thinking about global strategies, businesses (even medium and small ones) have learned to use digital tools to automate their processes. One of these tools is RPA solutions, which have become widespread - they relieve people of regular routine processes. For example, RPA can fill out report forms or transfer data from one database to another. Unfortunately, while companies use these tools unsystematically: in a long end-to-end business process today, on average, only one third is automated. Robots are not as smart as we would like, because business is not taking full advantage of AI.
For most companies, AI is actually a very complex and obscure technology. IBM points out in its research that most executives believethat their companies do not have the necessary competencies in data science, machine learning and other AI-related technologies to automate processes.
Business is ready to pay money for specific benefits that can be obtained from automation and digitalization, but does not understand how this can be realized on its own.
Solutions began to appear on the market that help to easily integrate external developments into their business processes. Digital services have also begun to emerge that help combine AI with automated RPA solutions.
How AI helps robots
Today, using the synergy of AI and RPA, you can do what was previously impossible within the framework of the usual automation of routine business processes. RPA is committed to applying cutting-edge technologies, including artificial intelligence (AI) and machine learning, to increasingly automate processes and empower humans. We found several interesting fresh cases that illustrate the possibilities of using this approach in various areas of business.
In general, about 90% of the current cases of using AI are working with documents in various forms: recognition of passports, PTS, diplomas, checks and payments. Cases for recognizing useful content in emails are especially relevant. When receiving a letter from RPA, AI helps to select the main thing from the text, classify the letter appropriately and send it to the desired addressee. Today, robots, using AI, help the accounting department, human resources, sales, purchasing, logistics and other departments that deal with the collection and processing of information.
Merchandising in a new way
Among the trendsetters and innovators of merchandising, one can distinguish, for example, Walmart, which launched at the end of 2019a system based on artificial intelligence, which makes it possible to monitor goods on the shelves in real time. The system was installed in one of the stores of the future, operating in the Intelligent Retail Lab concept - or IRL for short. AI cameras monitor in real time the availability of goods on the shelves. The gadgets will track inventory levels to determine, for example, whether staff need to bring more meat from the warehouse refrigerators and restock the shelves, or give an alarm if some fresh food has been on the shelf for too long and needs to be pulled out.
Automatic resume scoring
Predictive behavioral analytics - a new wordin HR: the largest companies in the world try to retain valuable talent with its help, noticing dangerous changes in the way people feel about work in time. Some, for example Sberbank, go further and already at the start use scoring of candidates according to preliminary parameters to predict the likelihood of their dismissal. It is assumed that in this way the bank will be able to fight the high turnover of personnel in mass positions.
One of the most productive uses of AI in human resources management is associated with the search for burned out workers. So, the American company Ultimate Software Group, which develops software for personnel management, has created a safety indexemployees. This is an indicator by which, based on 50 indicators, you can predict whether a specialist is going to quit in the near future.
IBM used its AI supercomputer Watson to create a similar solution . To gauge a person's mood, the supercomputer analyzes their career history, length of service, salary, job responsibilities, distance from work to home, and other metrics. Now, the HR department of a company can predict who is going to quit with a 95% probability six to nine months before it happens, and take measures to prevent layoffs.
Sentiment analysis in blogs
Sentiment analysis or analysis of the sentiment of information flows has great potential for application for monitoring, analytical and signaling systems, for document flow systems and advertising platforms targeted on the subject of web pages. The direction is considered one of the most attractive, which encourages the study and application of AI in various industries.
The authors of an already classical studyused sentiment analysis to study people's opinions and feedback about three car companies: Mercedes, Audi and BMW. The robot fetched all tweets with brand mentions, after which they were processed using text mining methods. All tweets were divided into three categories: positive, negative, and neutral. The results of this study provided insight into the importance of analyzing consumer reviews and opinions in this industry. The authors managed to get very valuable information for the marketing of these brands.
The analysis of AI tweets showed that Audi received the most positive reviews (83%). At the same time, Audi received less negative feedback (16%) than other studied competitors. It is obvious that the advertising offers on the Audi website will reach more loyal users than those on the BMW and Mercedes websites. There is something to think about for both manufacturers and marketers of these cars.
How to implement RPA and AI synergy
There are several solutions on the market that are helping to harness the power of AI for robotic processes with varying success. If you believe the specialized ratings , then the leader in creating services for robotization is UiPath. According to statistics, on average, about 30-40% of the end-to-end business process is automated today. Using the UiPath platform, which includes solutions for Process Mining, AI Fabric and other products, you can increase the percentage of automation of such processes to 70.
The vendor recently released the AI Fabric platform that helps to get synergies from RPA and AI. AI Fabric is the link between artificial intelligence and automated processes. The platform is designed to take ML application in business processes beyond small, highly skilled development teams and pass it on to business users. In other words, with the help of this solution, even a junior developer can implement AI in a company - you no longer need to delve into technical subtleties and independently deploy the necessary infrastructure. For business, the platform is useful in that it helps to understand in practice the benefits and opportunities from using AI in real business processes.
The percentage of automation of various business processes can be further increased and increased, but only with the help of RPA this will no longer work - the tasks have natural limitations on their formal algorithmization. But in conjunction with AI, this can be done.
The AI Fabric platform allows you to use your own machine learning models or models purchased from third parties with robots. Using the results of their work, you can automatically obtain data to improve the performance of models. Thus, you get the opportunity to seamlessly integrate AI into the business processes of the company and at the same time convenient tools for managing your models.
Getting started with the platform is easy. First you need to define the category of your user-case. Then choose the appropriate model that suits your request, for example, from those that are supplied "in a box", or developed in your company. As you know, machine learning models are quite gluttonous in terms of CPU and GPU resources, and therefore AI robots are automatically created for the models to work, which in fact are special containers that allow you to flexibly manage resource consumption.
As an example, consider the case of predicting customer churn.
UiPath platform
- Retrieves information about users from CRM
- Receives customer churn prediction from AI Fabric
- Updates CRM based on this information
- Sends information about such users to employees
Scenario implementation allows you to proactively prevent customer churn, eliminate the human factor due to late response, and optimize customer retention and acquisition costs.
Services that leverage the full power of AI in conjunction with traditional automation tools help businesses lower the threshold for AI adoption. They allow companies to use out-of-the-box solutions and save their own resources. Today, this is becoming a new trend in hyper-automation, which will become widespread in the near future.