1. Rebound for enterprise AI
Although the overall economic environment related to the aftermath of covid crisis remains uncertain for the time being, we believe that 2021 will see a rebound in enterprise AI.
Organizations’ AI investment will increase by at least 10% compared to 2020, although the global IT spending recovery will be slower.
2. AI for hybrid working
A hybrid working mode is becoming the new normal. Remote working versus working in the office brings many changes to organizations and teams. The challenges for employees to access information and collective knowledge are amplified in the context of hybrid working. Asking a question simply to a colleague in the office or at the coffee machine is no longer the norm. Ensuring employees an easy, secure and effective access to information, knowledge and insights has never been as pressing as it is today.
We predict tools in cognitive search and insights as a growing focus area of AI application in 2021. Customer experience and process automation will continue to be the mainstream AI use case domains.
3. MLOps to scale AI
In the past year, more and more POCs have finally entered production. However, we see that organizations were poorly prepared to manage AI software in production: managing the data and machine learning pipelines, increasing automation, introducing continued training and deployment, etc. Many of them are just starting to see the iceberg of issues. And the situation would be aggravating when they start to have over 20 and 30 models in production at the same time. Implementing MLOps practices will help organizations scale AI and improve their overall AI maturity.
We predict that MLOps is a fast-growing topic for companies as they start to launch such initiatives and acquire related tooling and solutions in 2021.
4. AI ethics beyond guidelines
Principle, charters, and hundreds of discussion papers have been written by various organizations regarding the vision of ethical AI. No one in the industry is exempt from ethics talks and debates. In the meantime, ethics washing has become a real issue for some organizations and vendors, a risk for the development of a trustworthy AI. Now is the time to acknowledge the complexity and certain scientific limits in ethical AI or simply AI. Now is the time for realistic actions. Beyond guidelines and principles, organizations will move towards implementing concrete actions and processes for ethical AI.
5. Emergence of AI factory
As governance and multidisciplinary working continue to be big challenges, large organizations will review their AI governance and set up specific entity like AI Factory that aims to accelerate AI at scale. With the sponsorship from the CXO level, the AI Factory has the mandate to develop and manage AI solutions across the entire organization. In addition to the required expertise, methodologies and processes, the AI Factory is supported by a comprehensive technical platform that provides tools and services enabling the Build and Run of AI software.
