As the pandemic has accelerated the transition to digital, the general consensus is that many businesses that have remained primarily physical have taken the step in this direction. Even if they do go back, a sizable portion of their transactions would continue to happen electronically. Indeed, innovation through digital and data will enable leaders to differentiate and stand out.
Since data is at the heart of digital, managing it and the associated infrastructure, as well as having the right strategy and planning, will be critical to business success. That’s why we expect a lot of innovation in the areas of infrastructure and data management architectures. Here are the big five trends we see having the biggest impact in 2023 as it relates to data and data analytics.
Faced with the specter of recession, companies will seek to optimize their infrastructure costs
Whether or not France is in a recession in 2023, companies are actively reducing their costs, as well as their IT infrastructure, which has always been an easy fix for their leaders. Processing and storage costs continue to decrease due to the use of the cloud. However, they can still incur heavy bills for companies given their huge investments in data analytics infrastructure. Partly due to the wide choice of storage, compute and application solutions, companies often adopt a complete replacement strategy to modernize their infrastructure in this area. Not only is this approach expensive, but it can often disrupt IT operations. In 2023, more and more companies will focus on modern and non-disruptive solutions for updating their IT infrastructures. This is whether their data resides entirely in a single cloud, across multiple clouds, or in a hybrid environment maintaining on-premises facilities.
With multicloud it becomes necessary to control cloud costs
For many businesses, data is spread across multiple clouds and geographies. This may be due to different preferences in choosing a cloud service provider (CSP) or as a result of mergers and acquisitions between entities dependent on different CSPs. As data migration to the cloud intensifies and some CSPs gain traction in some regions over others, the adoption of a multicloud architecture is accelerating in multinational organizations. There is currently no simple solution to manage and integrate data and services across these various CSPs. The persistence of this problem always leads to the creation of data silos and a fragmentation of data management, leading to complications in their access and governance.
Also, contrary to popular belief, cloud costs are becoming more and more hardware-based due to the sheer volume of data and associated egress costs, to name a few reasons. For many businesses, cloud investments aren’t delivering the expected economic and business benefits. They use FinOps methods to audit cloud costs and usages, identify cost-to-value, and determine how to optimize cloud management across today’s hybrid and multicloud environments. In the coming year, FinOps is expected to grow and play a pivotal role in helping enterprises better manage their hybrid cloud and multicloud spending.
Accelerate adoption of data fabrics and data meshes
Over the past two decades, data management has gone through cycles of centralization and decentralization: databases, data warehouses, cloud data stores, data lakes, etc. While each approach has its proponents and opponents, recent years have shown that data is more distributed than centralized in most organizations. While there are many options for implementing an enterprise data architecture, 2022 has seen the accelerated adoption of two – data fabrics and data meshes – designed to improve the management and access of distributed data. The two are different in nature: data fabric is a composable set of data management technologies, and data mesh is a process orientation that allows distributed teams to manage enterprise data as they see fit. Both are essential for companies that want to better manage their data. Easy access to data as well as its governance and security are important for every data actor, from data scientist to business leader. These are, in fact, essential for the production of dashboards and reports, advanced analytics, Machine Learning (ML) or artificial intelligence (AI).
Both data fabric and data mesh can play a vital role in accessing, integrating, managing and disseminating data across the enterprise, when implemented with the right infrastructure. Therefore, in 2023, a decisive acceleration in the adoption of both architectures in medium and large enterprises is foreseeable.
Ethical AI becomes critical as more and more decisions rely on AI
Companies are increasingly turning to AI for data-driven decision-making, whether it’s moderating social media, connecting healthcare professionals with patients, or issuing credit from consumer banks. However, when AI biases the decision, there is currently no way to eliminate the algorithm’s inherent bias. This is why legislation in the works, such as the EU’s proposed “artificial intelligence” directive, is starting to regulate the use of AI in commercial enterprises. These new regulations classify AI applications according to the risk they pose (unacceptable, high, medium or low) and ban or regulate their use accordingly.
In 2023, companies will need to be able to comply with these regulations, especially in terms of privacy protection and data governance, transparency of algorithms, fairness and non-discrimination, traceability and auditability. To this end, they need to implement their own frameworks for ethical AI, for example in the form of guidelines for trustworthy AI, peer reviews or even dedicated ethics committees. As more companies implement artificial intelligence, ethical AI is set to gain unprecedented prominence next year.
Improved data quality and preparation, metadata management and analytics
While often intended to power advanced analytical tools and AI and ML techniques, proper data management is itself essential to business success. Data is often referred to as the new black gold because data analytics is constantly driving innovation. As companies increase their usage, it is crucial for them to not lose sight of their governance and quality, as well as metadata management. However, as their volume, variety and speed continuously increase, these various aspects become too complex to be managed on a large scale. Witness the time data scientists and data engineers have to spend researching and preparing data before they can even start using it. That is why several industry players have recently proposed augmented data management that allows companies, through the application of AI, to automate a large number of tasks in this field.
According to some of the leading analysts, every layer of a data fabric – ingestion, processing, orchestration, governance, etc. management – should incorporate AI or ML, to automate each stage of the data management process. In 2023, augmented data management will gain strong traction in the market, helping professionals focus on data analysis without being hampered by routine administrative tasks.
While these are five powerful trends, there are other areas of analysis that will determine both the survival and success of digital businesses in 2023 and beyond. The last three years have certainly taught us that digital is really not a second-best solution when face-to-face meetings are impossible, but a solution for the future.