Some of the most promising themes within digital transformation revolve around new possibilities with artificial intelligence (AI) and machine learning (ML) and the rise of the Internet of Things (IoT).
The organizations that can best incorporate these emerging technologies into their operations, business models and services will be able to out-perform their competitors and will ultimately be the ones that survive in this future technology-powered landscape.
However, there are several potential setbacks on this journey. If you do not have a firm grip on data quality, you will run into several obstacles that will significantly inhibit your ability to harness the power of these nascent technologies.
Thankfully, there is a solution to the data-quality issues that can plague these and several other IT initiatives: master data management (MDM).
Why Master Data Management Is Critical to Success with AI, ML And IoT
The lifeblood of AI, ML and IoT is data, and the circulation of data must run automatically. This is only possible with data of high quality.
Master data management is the technology infrastructure and processes that put data-quality standards to work, ensuring consistent and reliable information — and relationships across enterprise systems.
While MDM forms the foundation of enterprise data strategy, it is most directly concerned with master data, the core, non-transactional data used across the enterprise, including customers, products, suppliers, and locations — to name a few.
Master data that is used throughout several business processes — and therefore must be fit for use in many different scenarios — must meet a range of data-quality dimensions to underpin automated business processes.
Data-quality efforts must be sustainable
If data quality is not continuously maintained, the data that was of high quality at a given time for a given purpose will very quickly decay. Such unmaintained data will not be suitable for use in AI-supported business processes or be suitable for operating IoT environments and analyzing the data gathered in such environments.
The data that is most compromised in this challenge is the master data that describes the core entities involved in these business processes and environments. MDM is the right solution to use to onboard reusable master data and control the lifecycle of that data.
Complexity must be captured and encapsulated to power AI processes
As humans, we have a natural-born capability to understand the complexity of the who, what and where of the core entities involved in business processes and data gathering. Machines, however, must have a digital digestible way of getting that picture.
MDM is the ideal solution for providing AI with an encapsulated description of the related core entities involved in business processes and (the same) core entities involved in connecting smart devices in IoT environments.
Why Artificial Intelligence Relies on Data Quality
The business advantage of using AI is to automate business processes and to arrive at faster, more reliable business decisions.
The risk of unwanted business outcomes
However, if the AI processes run on top of data that is not unique, accurate, consistent and timely, these processes will not produce reliable results and therefore lead to unwanted business outcomes.
Examples of such unwanted business outcomes include:
- Making different decisions for two customers or supplier master data records that describe the same real-world party.
- Recommending a product to a customer where a remarkably similar product has previously been returned.
- Accepting a purchase from a vendor where another company family member has previously been ruled out.
Such results can have an extremely negative impact on business outcomes, the reputation of your business and make your business reluctant to embark on new AI initiative, which can inadvertently cause you to slip behind your competition.
Avoiding endless data cleansing
A knee-jerk reaction to data quality issues in AI processing will be to start ad-hoc cleansing the data that goes into that process. This kind of symptom relief will, unfortunately, be extremely costly and unhealthy overall and will only become unmanageable AI continues to support additional business processes and use cases.
The better way is to sustainably cure the data-quality issues at the source by using a capable MDM solution. A robust, multidomain MDM solution can connect disparate enterprise systems, merge and match data (potentially using its own ML engine) to build and maintain a trusted golden record of customer, product and other data.
When the entire organization is working from a ‘single source of truth,’ practitioners of AI and ML programs know they are working with accurate, timely and accurate data.
Rationalizing data used to power machine learning
Machine learning (ML) is the discipline used to ignite AI. While you support the machine with training data for ML, it may be tempting to do a little data cleansing as the training data will not be part of the continuous future AI processing.
The risk associated with taking this approach is that you will probably cleanse each training dataset a little differently. This means that when more AI-supported business processes start to interlink, they will have a different “way of thinking.”
The answer is to have your training datasets derived from already-rationalized production data having the same master data foundation.
Using rationalized data in artificial intelligence
Having your AI-supported business processes running on top of master data that is unique, accurate, consistent and timely will make a huge positive difference for the business outcomes achieved from applying AI.
The results will be reliable. The processes will be repeatable over time. The concept will be reusable in other scenarios.
Why Internet of Things Will Extend the Scope of Master Data Management
The potential of the Internet of Things (IoT) theme is enormous. We will increasingly use smart devices that are connectable in our daily life.
The smart devices that will become an increasingly large role in home security, frictionless purchases and more, so will the importance of the underlying data that powers these initiatives.
Master data management will similarly become more critical as data volumes increase and the market for IoT devices and services matures and more companies get involved.