Read on to learn more about the basics of AI, Excubate’s AI player selection process, and a brief introduction to an example AI manufacturing solution.
Companies collect huge amounts of data on all kinds of business processes. Analyzing this data and turning it into tangible business value is uncertain, complex and time-consuming. A rapidly growing number of artificial intelligence-based software solutions aim to solve this challenge by using intelligent algorithms to analyze data sets, derive insights, and automatically recommend actions. Beneficial use cases can be found along all stages of the value chain; getting started is often easier than it seems. Examples show that ROI can be achieved as early as after just one day.
What is “AI”?
Artificial intelligence (AI) is a subfield of computer science that aims to enable machines to replicate human thinking and decision making. This is often achieved with machine learning (ML) algorithms, that continuously learn from datasets. In many cases the current discussion about AI is in fact a discussion about its subfield ML.
In comparison to classical systems (hardcoded and simple rule-based), ML algorithms can learn from existing data to react to previously unknown situations. Example: AI can learn what a person looks like from several existing images and then apply that rule to new images and recognize that person. This is possible due to the ML algorithms’ ability to recognize patterns in data and derive rules and statements from it. ML can be used in situations where large quantities of complex and unstructured data must be analysed and searched for patterns. It is faster and more accurate than human analyses of the same.
The potential use cases for AI are numerous but have yet to be fully understood. Here are some relevant business examples:
- Manufacturing process optimization achieved by sensor-controlled anomaly detection (e.g. prevention of unplanned downtimes, using machine vibration analysis)
- Recruitment improvement by matching talent supply and demand (AI-powered analysis to assess an applicant’s previous work experience and interests and match them with the most suitable job openings)
- Price optimization based on historic and real-time data to predict how customers are likely to react to different pricing and automatically adjust it accordingly (e.g., ML algorithms crawl the web to gather information about competitor prices and trends to adjust own prices while keeping in mind business goals)
- Product & process development, enhanced by design suggestions, consistency checks or predicted product properties for changed parameters
- Automatic lead identification based on similar existing customers or market data, including information on customer intent (e.g., “What will my customer buy next?”)
Realizing tangible value with AI
In many cases, companies are hesitant to introduce investment-intensive technologies – and often wait until best practices appear on the market and risks are foreseeable. However, when it comes to AI, hesitation might lead to a significant loss of opportunity. Experts go as far, as to call AI as an essential building block for the future viability of companies. Within the last two years we observed 5 major challenges, that companies are facing when implementing AI technology:
- Identification of impactful, measurable, and individually fitting use cases, including creation of transparency on their quantified benefits and risks.
- Knowledge about what and how data can be used, data quality, and availability of data relevant to the application area.
- Internal AI technology knowledge and hiring high-in-demand data science teams or training existing personnel.
- Complex and time-intense integration of AI solutions in legacy systems (e.g., SAP).
- Ambiguity of what applications are available on the product and process side, oversupply in emerging solutions, and unclarity in regards of make or buy benefits.
Despite these challenges, AI usage is on the rise. In a 2021 McKinsey study with over 1.800 international companies, 56% of all respondents reported AI adoption in at least one business function. Furthermore, 27% of respondents reported a minimum of 5% EBIT that is attributable to AI.
Today, the benefits of AI outweigh the costs in a wide range of use cases. Using the Digital Value Canvas® (read here more about the methodology) the benefits of AI usage in the example of industrial manufacturing can be mapped as followed: