Artificial intelligence (AI) is the simulation of the human intelligence process by machines. It includes learning, reasoning & self-correction. In the last few years, machine learning has evolved rapidly. Several enterprises are building AI assisted systems using various cognitive components like Natural Language Processing (NLP) and machine learning models. This article discusses the current market outlook, the key challenge and the ways to overcome the challenge in AI adoption across enterprises.
Current Market Outlook
According to a new update to the International Data Corporation (IDC) Worldwide Artificial Intelligence Systems Spending Guide, spending on AI systems will reach $97.9 billion in 2023, more than two and one half times the $37.5 billion that will be spent in 2019. The compound annual growth rate (CAGR) for the 2018-2023 forecast period is estimated at 28.4%.
According to the IDC spending guide, on a geographic basis, the United States will deliver more than 50% of AI spending, which will largely be in the retail & banking industries. Western Europe will be second largest geographic region where most spends will be in banking & manufacturing. China will be the third largest region by AI spending in retail, state/local government & professional services being the major recipients. The strongest spending growth over the five-year period forecast will be Japan (45.3% CAGR) and China (44.9 % CAGR).1
1https://www.idc.com/getdoc.jsp?containerId=prUS45481219
Even though as per the forecast, AI spend is expected to reach $97.9 billion by 2023, the numbers showing the actual usage of AI reflects a different story. Ben Lorica, O'Reilly Media's Chief Data Scientist, interviewed 1,300 respondents to pin down how enterprises are planning, deploying and struggling with AI.
Out of the 1300 respondents, only 27% of them had a mature practice. 54% of them are in the evaluation phase & 19% are not using AI in their organization. From the same sample data some of the bottlenecks for AI implementation are depicted in the below diagram :
The Challenge
Leaders across various industries think of AI as a plug and play technology with immediate returns. They invest millions in data infrastructure, AI tools, data scientists etc. to get the projects up and running. Some of the pilot projects have delivered small wins but there have been no tangible big wins as was expected. Companies find it difficult to move from small pilot projects to enterprise wide programs. Hence, it is important for the leaders to align company’s mission, culture, structure to support AI adoption.
In order to achieve this, leaders/companies will need to make a change in the processes as suggested below:
1. From a siloed approach to enterprise wise collaboration
AI will have a huge impact when use cases/modes are developed by various cross-functional teams as when these teams (business & operational) work together with data scientists, it ensures the entire process is prioritized. When the development team involves end users/business users in the design process of application, the chances of adoption of that application increases.
2. From experience/judgement based decision to data-driven decision
Leaders should use algorithms’ recommendations to arrive at a better decision rather than relying on one’s judgement and intuition. For this approach, the stakeholder at different hierarchal levels should be allowed to trust the algorithms’ suggestion. If employees have to consult a higher-up before taking any action, than this will inhibit the adoption of AI.
3. From rigid and risk averse to agile, innovation & adaptable mindset
Organization needs to change the mindset that AI will solve all the business problem in the first iteration. Adoption of AI is a continuous process where “test and learn” mentality will help in discovering news sources of optimization made from the past mistakes. Getting feedbacks from past mistakes/success and incorporating them into the next iteration will allow companies to fix minor issues before they become an expensive problem.
4. From multiple tools to a centralized platform
Companies have invested in independent AI tools to solve siloed problems. In this approach, there is a problem in optimization as tools are used in siloed business units and not leveraged to their full capacity. Companies should identify the opportunity to consolidate independent tools to a central platform. This central platform should act as an automation market place to build all the automation use cases and provide a single place for the governance of the digital workforce.
For easy adoption & scalability of AI in an organization, it is important to have a governance layer consisting of business and IT operations, SMEs, analytics leaders and data scientists.
There should be a central team headed by an executive like a CAO – Chief Analytics Officer, who can help align the strategy. This team will have to be responsible for partnering with the data and AI service provider, AI standard, policies, innovation and the performance management team.
The innovation team for independent business functions must be headed by a business unit head and supported by an AI product owner. This team should be responsible for AI solution/use case identification, solution deployment, adoption and performance tracking.
There should be an execution team with members from both central and innovation teams with business analysts, data architects, data engineers, data scientists and product owners. The core responsibility of this team will be to do organization capability assessment, shortlisting AI service provider, solution feasibility, data strategy, building algorithms, setting direction for AI projects.
Conclusion
Enterprises consider AI among the top three investment areas for the next four to five years. For seamless AI adoption, CXOs must build AI strategies that align with their business strategies. Necessary change management should be done to drive enterprise-wide adoption. To scale up AI Adoption,companies should invest in benefits realization office/Center of excellence, which would provide a platform to co-innovate, co-create automation solutions with SMEs from different lines of business.3
References:
1https://www.idc.com/getdoc.jsp?containerId=prUS45481219
3https://hbr.org/2019/07/building-the-ai-powered-organization
Rahul Tyagi
Rahul Tyagi is a Pre Sales Consultant for HOLMES - Wipro’s Artificial Intelligence & Automation Platform. He has experience working for driving GTM and automation advisory for BFSI, Retail & Hi-Tech clients. He is an MBA from Goa Institute of Management.