Top 5 Trends in Artificial Intelligence and Machine Learning for 2022


In recent years, data science and advanced analytics have been among the most talked about topics in Fortune 100 strategy sessions and board meetings. The pandemic and remote working have only upped the urgency of these conversations. Not only are consumers getting more tech savvy and making educated buying decisions from the comfort of their living rooms – these also represent the remote workers who are gaining greater decision-making power in their place of employment.

It is no longer just the CXOs or business heads of B2B or B2C companies who are making business decisions. According to a Pew Research Survey, Gen Y’ers are increasingly getting involved in the decision-making process in their organizations. And these under-40 power hitters do not rely on word of mouth or old networking channels to decide on their technology service providers or business consultants. They are focused on advanced technologies and their computing prowess.

More than half of the surveyed Gen Y decision-makers and influencers said they rely on artificial intelligence (AI) and machine learning (ML) models and omni-channel auto-bots to inform their decisions. This number was even higher (approximately two-thirds) among Gen Z respondents.

Nearly every business leader in every Fortune 100 company is interested in knowing how to connect and influence these Gen Y and Gen Z professionals and consumers. We recommend starting with a clear understanding of the top trends in AI and ML.

Here are the top five trends in AI and ML leaders should be watching for in 2022.

1. Small and Wide Data Analytics

Small data is the collection and analysis of data sets sourced within an individual organization or based on individual problem-solving examples. It answers a very specific question. Wide data ties together disparate data sources across a wide range of sources to generate meaningful insights.

Hyper-local food delivery is a classic example of the effective use of small data analytics techniques. Consumer-facing e-commerce companies in the fintech, edutech, entertech and agritech sectors are also using small data to uncover localized customer behavior patterns, likes and dislikes. Retail companies can use small data to understand their customers’ preferences to design and promote localized sales, marketing or distribution strategies.

The beauty is that these companies don’t have to rely on heavy data storage and computing platforms. Their internal CRMs and ERPs can retrieve and store the local data on their data platforms. Auto ML models can then be designed by data scientists to come up with actionable business insights.  

Wide data analytics are often used in calculating stock market prices or designing systematic investment strategies and marketing campaigns across multiple channels.

Wide data needs more sophisticated data storage and computing platforms than small data. But the data size is much less than big data and, therefore, more cost effective. Business insights from wide data tend to be more regional, which is something big data has missed in the past.  

ISG expects 50% of Fortune 500 companies will invest in wide and small data analytics by 2023. Retail, CPG, QSR, Banks, Financial Services and Hospitality companies are most likely to benefit from wide and small data analytics by better understanding their consumers’ behavior patterns on a local level.

2. Edge Analytics

Edge analytics is an approach to data collection and analysis in which an automated analytical computation is performed on data by a sensor, network switch or other device instead of waiting for the data to be sent back to a centralized data store. Not only does it reduce needed storage space, but it significantly enhances the computational speed of a transaction.

Autonomous vehicles, in-hospital patient monitoring and smart home devices are examples of edge computing. Multi-channel customer communication is where edge analytics is finding value. Customers need to be continuously engaged by brands on various communication devices, including mobile, laptops, desktops and smart devices that depend on internet of things (IoT) technologies. Edge technology conducts channel-agnostic data storage and analysis to generate actionable business insights for consumer sector brands.

LinkedIn’s research shows that Technology and Research industries are among the top three fastest-growing sectors for two years in a row. Both these industries require high-speed data analysis and large data storage spaces. With lower storage costs and greater computing speeds, edge computing can serve these industries effectively. Initial tech infrastructure costs are still high, but with more research and usage, these costs are bound to go down. Other tangible advantages of edge computing include near real-time analysis of data, scalability and improved security.

3. Data Democratization

Anybody can use data at any time to make decisions on a unified analytics platform. Users can simply drag and drop automation blocks into a palette and start seeing insights instantly. Greater accessibility to an organization’s data creates data democratization. Of course, this comes with some implications for data privacy, governance and security. Data democratization platforms in the market like Alteryx and Macheye have built data sourcing, data accounting, data security and governance modules to address these issues. The goal is to have the user spend less time on data discovery and more time on data storytelling to drive business insights.

According to Equinix, 50% of an enterprise’s data will be created outside its data centers and cloud environment by 2023. This will drive demand for technologies like edge computing.

Consumer companies like those in Retail, CPG, Food and Beverage, Healthcare and Financial Services stand to benefit from data democratization. When companies design a new digital strategy or transition to a new CRM, business leaders should invite influencers from finance, marketing, operations and sales to help analyze the data (not the strategy), look at the data from the eyes of the stakeholder and design simple and effective data solutions that enhance the customer experience.  

Researchers from Harvard and Cambridge Universities believe that potential scientific breakthroughs for critical diseases like cancer and HIV – and even COVID-19 – can be accelerated using data democratization because it facilitates discovery and advanced research across the globe.

4. Data Fabric

A data fabric is a data management architecture that optimizes access to distributed data and intelligently curates and orchestrates it for self-service delivery to data consumers. It provides a single environment for accessing and collecting data, no matter where it is located and no matter how it is stored. A data fabric delivers greater scalability to handle increasing data volumes and data sources to make it easier to leverage the cloud by supporting on-premises, hybrid and multi-cloud environments and faster migration between these environments.

A data fabric can complement an AI framework for fraud detection, cloud security applications, real-time sales activities and customer onboarding, in which the AI/ML model can highlight the topical map of anomalies and inflection points. Leading advisory firms like Gartner and Forrester have been highlighting data fabrics as the AI trend for the past two years, but the primary reason it has not taken off is the lack of understanding about data fabrics by organizational leaders. When leaders partner with the right data advisors, however, they will be able to transform their companies by learning from the past and evolving over time.

BFSI, Manufacturing and Telecommunication companies can benefit from data fabrics if designed and implemented in a clearly defined and structured way. A data fabric must include the following architectural components:

Component 1: Ingestion, which includes all data sources (internal or external, paid or public, domestic or international, private or open source)

Component 2: Processing, including analytics data and knowledge graphs

Component 3: Insights, which compiles various advanced algorithms and machine learning models

Component 4: Interfaces, including all the APIs, SDKs and connectivity tools

Component 5: Data layers, including the consumption and transport layers for all the organizational data

Component 6: Platform, which is the primary hosting platform and environment used in the hosting environment

5. Graph Analytics

Graph analytics are analytics tools used to determine strength and direction of relationships between objects in a graph. The focus of graph analytics is on the pairwise relationship between two objects at a time and the structural characteristics of the graph as a whole. Graph analytics saves time in data organization and consumes less effort while merging more data sources or points. The graphs are visually appealing and easier to understand than other data analytics tools and models.

According to a Markets and Markets survey, the graph analytics market is poised to grow to $2.5 billion by 2024 from its current market size of little less than a billion dollars. Social media companies deploy graph algorithms using various techniques like centrality analysis, connectivity analysis and path analysis. It showcases the network of relationships between its users and identifies key behavioral trends.

A now classic example of graph analytics is the identification of networks of relationships conducted by the International Consortium of Investigative Journalists (ICIJ) in the Panama Papers. This research sheds light on how authoritarian leaders and politicians use complex sets of shell companies to obscure their wealth from the public.

Insurance companies, governments, Oil & Gas and publication houses can benefit from graph analytics to find indirect relations and express trends in large and complex data in a cohesive way.

ISG helps companies take advantage of emerging technologies in AI and ML to solve real business problems. If you want to brainstorm ideas about how these AI/ML trends might benefit your organization, please reach out to us at ISG. We love to hear from our readers and engage into symbiotic professional partnerships.

About the author

Sush co-leads ISG consulting Analytics solution and has substantial experience in helping Fortune 1000 organizations adopt real-time insights driven connected enterprise business models.