Oleg Danilchenko, PwC: AI already has a visible impact on business 20.04.2018
Oleg Danilchenko, PwC: AI already has a visible impact on business

Consulting companies are being transformed now: the list of basic services includes services using artificial intelligence technologies. AI already brings real income and create a tangible impact on business, thinks speaker and moderator of AI Conference 2018, Director of data analysis of Russia’s PwC office Oleg Danilchenko. We talked about the benefits of using AI technologies in business, the experience of using AI in PwC and trends on the world AI market.

Interviewer: AI Conference (AIC)
Respondent: Oleg Danilchenko (OD)

AIC: Oleg, tell me, what is the area of activities of the Competence Center you are heading?

OD: One of the areas is the application of data analysis technologies in various industries: in telecommunication companies, in the banking sector, in trade networks, in energy and others. What we are doing is the systematic application of statistical methods to large and complex data sets to study the relationship between key parameters that the business uses. It is a project work.

Another direction is the development of software and hardware products based on data analysis, machine learning and other artificial intelligence technologies.

The third is the development of a strategy for using data within the company and applying innovative technology solutions to customers. Here we are talking about the creation of road maps and business cases for the introduction of technologies.

Our team includes data engineers who are qualified enough to construct statistical models, including models of machine learning, these are experts with practical experience.

AIC: What methods and technologies does PwC use to collect data?

OD: We use a whole range of technologies when working with customer’s data. They are usually based on solutions such as Apache, Apache Hadoop, GridGain Ignite, and Microsoft Azure.

AIC: What is the role of AI in data analysis at PwC?

OD: The business of consulting companies around the world is transforming. And now the list of professional services of companies, including the Big Four companies, includes services for data analysis using artificial intelligence technologies. As an example: to improve the quality of our auditors' work we use the technology of recognition of scans and their translation into digital format.

Another example: when estimating the cost of commercial real estate, we apply the technology of analysis of big data (primarily geolocation). We are building a model of machine learning, which predicts both the value of the property and rental rates.

AIC: What was the most difficult task for you in PwC? How did you cope with it?

OD: We had to make a prediction of the failure of electric submersible equipment for the oil and gas producing company. It is called forecasting and maintenance. The difficulty was that the amount of data was significant: there was a lot of information about the usual state of the equipment, but there is very little information about breakdowns. I must say, this was a difficult mathematical task, for which a lot of methods were tested, in particular, the Dirichlet method was applied.

A large number of failures in oil and gas production facilities occur on gas pumps or on submersible equipment. In our case, we used data from sensors. There were about 1134 pumps. Information was provided from 68 sensors with a discreteness of two events per second.

Data preparation was carried out primarily by filling in missing or blank values. The emission was cleared in the data statistics. The primary study of regularities was conducted: in particular, the templates for the operation of equipment in the regular and not regular modes were compared. Based on this, a forecast model was constructed that described the key parameters of equipment failure. The number of precedents, values ​​in the objective function was low, so we began to doubt whether this problem can be solved. But in the end it was done.

AIC: How would you rate the market of developments in the field of artificial intelligence in Russia? How competitive is it at the international level?

OD: From the point of view of technical implementation, it is very competitive due to the powerful scientific base that is still present in the post-Soviet space.

And from the point of view of introduction in business process there is still no successes. Business in Russia is often afraid and does not understand how to use AI technologies. Plus, not all decisions that are born in incubators and accelerators get to market.

This direction needs to be developed and invested in at the national level. Since much will be based on AI in the fourth industrial revolution.

AIC: What are the promising AI developments in the Russian Federation?

OD: Many enterprises in various industries are striving to achieve digital leadership by introducing innovative technologies. There is a number of techniques and algorithms in AI technologies that increase the operational efficiency of companies in the whole world (including in Russia) today.

There is a growing tendency to automate the most time-consuming routine work. The scientific and technical base in Russia allows a number of companies to take first steps in this direction, and some have already made significant progress. There are already a number of excellent examples in the agro-industrial complex of the country, metallurgy, oil and gas production and, of course, among financial institutions and telecommunications companies.

The key things that are now being introduced and are beginning to be used in the industry and metallurgy are:

  • monitoring of systems using sensors (processing, analysis and visualization of data received from sensors in real time, to increase the transparency of information about the current state of the equipment, notification and start of processes in the event of exceeding thresholds);
  • quality forecasting (identification of factors affecting product quality, construction of analytical models for identifying or predicting quality problems at the initial stage of the production process);
  • improving the efficiency of processes (identifying opportunities to improve the efficiency of processes based on the analysis of data from sensors, adjusting technological parameters to reflect the current state of the equipment).

PwC helps such companies to implement these solutions both in terms of strategic vision and growth potential, and from the technical and mathematical side.

AIC: You moderated the session at the AI Conference 2018. Share your impressions. What was your presentation about?

OD: I made a presentation on the strategy for using AI in industry, in the real sector and in general in commercial structures. I told how to properly implement AI in various processes, how to prioritize business cases, how to build the organizational structure of the company and transform it using these technologies.

There was a discussion on trends in the field of artificial intelligence that exist in Russia and in the world. According to the results of the PwC study “Artificial Intelligence: Do not miss the benefits”, global GDP will grow by 14%, or $15.7 trillion, due to the active use of AI in 2030.

AI technologies are not just the collection and analysis of large data, but an instrument for making real money and creating a tangible impact on business.

AI-industry news and cases for you