1. Agile goes deeper into the enterprise to reshape business models, culture

IT leaders have adopted agile practices to develop applications, improve machine learning models, automate CI/CD pipelines, and deliver other strategic programs that require teamwork with business stakeholders. The role of the product owner became essential to managing business/IT collaborations. Meanwhile, IT organizations reinvented how they worked as they adopted DevOps culture and practices.

But the move to agile practices and DevOps was just a step in an enterprise journey to establish agile business models and cultures that could support digital transformation. In 2020, more CIOs saw the doors open to partner with business leaders on the agility required to transform business operations. In 2021, expect to see CIOs lead agile deeper into their organizations:

• Multidisciplinary agile teams will include business teammates from marketing, operations, finance, HR, and sales, sprinting and releasing business capabilities.
• Change management programs will move earlier and become transformation programs by engaging customers, early adopters and stakeholders in the development process.
• More organizations will invest in agile portfolio management tools, adopt experimentation cultures, and expand agile to their data science teams.

2. Low-code use cases focus on customer and employee experiences

In 2020, low-code platforms became essential to organizations looking to rapidly develop workflow applications, integrations and automations in response to COVID-19 and remote working. These applications provided personalized assistance to workers on health, safety, and family needs and helped them set up their remote office environments. Automations filled process gaps and helped organizations reduce costs.

Low-code platforms have been around for decades but are now far more strategic to CIOs and business leaders looking to transform their businesses. In 2021, digital leaders will use a mix of low-code, remote process automation, and integration technologies to improve customer and employee experiences in more business areas and at a greater scale. Low-code platforms will not only help more organizations modernize legacy applications and move more workloads to the cloud, but also enable them to do so with fewer software development, cloud architecture and DevOps resources, which continue to be hard to find.

3. Cloud computing service providers enable multi-cloud architectures

Moving to the cloud for many enterprises essentially meant evolving to a hybrid cloud model where IT operations supported applications running in data centers, private clouds and public clouds. While most enterprises are maturing their architectures and services on a primary public cloud, many have come to recognize that they must operate across multiple public clouds. Reasons include the following:

• avoiding vendor lock-in;
• facilitating pricing and service-level negotiation;
• enabling innovation on emerging services;
• complying with data sovereignty regulations; and
• supporting acquisitions.

Supporting multi-cloud architectures is not easy today, but public cloud vendors recognize that enabling multi-cloud integration, management and support is critical to their business relationships with large enterprises. There are services like Google Anthos and Azure Arc that help IT manage system resources and Kubernetes clusters across multiple clouds. The public cloud service providers also recognize the role of edge computing in workloads involving human safety and other factors requiring low latency and analytics capabilities close to the data source; infrastructure such as Azure Stack Edge and AWS Outposts enable these edge computing deployments.

Vendor support of multi-clouds will address some of the challenges of digital transformation. The growing choice of deployment options and multi-cloud management tools is good news for enterprises looking to move and grow strategic workloads on the right clouds for the job.

4. Real-time data processing and event-driven architectures go mainstream

Real-time data processing used to be a difficult technical objective reserved for industries like financial services and ad tech, where businesses with split-second reductions in data latency create significant business advantages. Many companies lived with unreliable data updates scheduled to run nightly, weekly or monthly and with manual processes to fix stoppages, data quality issues or to execute complex queries.

CIOs now have several architecture options to support real-time data processing. These include cloud architectures that scale on demand, open source data platforms like Apache Kafka and Apache Spark, real-time data streaming platforms, platforms supporting event-driven architectures, and RPA software that can automate many aspects of data collection.

5. Simplified MLOps brings machine learning from POC to production

While big technology, social media and technically advanced enterprises have machine learning models running in production, other businesses are lagging. In Algorithmia's "2020 state of enterprise machine learning" report, only 45% of respondents have deployed models into production. Even organizations that successfully deployed machine learning (ML) models struggled, with over 40% taking more than thirty days to deploy a model. An IDC survey published in June found that over 28% of artificial intelligence/ML initiatives fail.

But don't mistake the enterprise's current struggles as an indication that ML is fading. As more IT and data science organizations better understand MLOps -- the collaboration of data scientists and IT pros to automate ML algorithms -- and as more platforms offer ML lifecycle management capabilities, ML will take hold.

The reasons are clear: MLOps offers advanced organizations the opportunity to reduce deployment time and monitor ML models for model drift in production. For organizations early in their AI journeys, the platforms provide a management and operational framework, so that data scientists can focus on developing and testing their models. The net is that in 2021, we should see more companies demonstrating successful ML models in production.

Businesses prioritize practical solutions that deliver near-term value

Digital transformation trends run the gamut, in keeping with the term's broad mandate. This list of 2021 trends is different. It focuses on technologies and practices that help more organizations deliver near-term digital transformation benefits via competitive data and technology capabilities.

Digital transformation failures arise when goals are not connected to improved business performance. Organizations in 2021 need improved customer experiences, operational efficiencies and advanced capabilities -- minus the extensive technical expertise and myriad implementation risks that have typically gone hand in hand with implementing the emerging technologies that enable these goals.

In other words, platforms in low-code, MLOps, multi-cloud management and data streaming will be key for businesses across many industries to go from agile technology processes to capabilities that drive business agility.