By Ashok Panda, Associate Vice President & Global Head – AI and Automation Services, Infosys
The last two years have taught us that business resilience has never been more important than it is today. Resilient organizations are ones that maintain continuous business operations no matter the economic conditions that surround it and safeguard its people and assets.
Forward-thinking businesses have realized that this pandemic-induced uncertainty is merely a sign of times to come and have learned the importance of being prepared for potential volatility. Implemented right, Artificial Intelligence (AI) can offer businesses deep insights that in turn can make businesses more resilient.
Businesses can envision the future better as AI connects the dots
For a business to be agile, it requires harmonization of disparate content from information management solutions. AI can accomplish this and keep a business nimble—a crucial quality in times as uncertain as these. As organizational data grows at a rapid rate, enterprise AI can help glean worthwhile insights to empower people to take well-informed decisions.
AI models have been helping businesses across sectors. In the retail industry, which is deeply impacted by potential supply chain issues, AI’s ability to forecast weekly sales and inventory requirements is transformative. In the food industry, these models can help restaurant managers predict footfalls by factoring in data such as weather forecast. Such insights help businesses make better decisions, plan their operations better, optimize expenses, and ultimately maximize sales.
Today, AI is being used to manage and automate IT infrastructure, glean new insights about customers, identify cyber threats, and even improve the hiring process. But more can be done. To extract maximum value from this technology, companies must make it accessible to employees across different lines of business with varying levels of experience.
Empowering business with technology through democratization of AI
The full potential of technology can be achieved only when it is made available to everyone who has a use for it. Business leaders and users are well versed with the business that they drive but they are not AI experts. On the other hand, data scientists can create cool AI models, but they may not always align with business goals as they are not familiar with nuances of business. By democratizing AI, organizations can bridge the gap between business goals and the AI models built to achieve these goals. Business users and data scientists become partners and work in synergy. Democratization creates dialogues between them to identify the problems they want to solve, build and test hypotheses, assess models, and experiment till they arrive at the right model.
For example, an AI-powered real-time dashboard instead of an Excel sheet in the hands of the sales team would empower them to make well-informed decisions and be proactive in their dealings with access to real-time data. Imagine the number of opportunities lost if this data was available only to select senior employees!
Just as AI use cases need to be accessible to all, its development should be democratized too. Sophisticated businesses are recognizing that the responsibility of creating such solutions need not always rest with the IT team and can be outsourced to citizen data scientists within the organization, using tools such as Low Code No Code software. This doesn’t just reinvent productivity but also drives data-driven culture within the company and helps businesses scale.
What are the best practices to democratize AI within an organization?
To democratize AI, organizations must first scale it and ensure it is embraced by employees across the organization. There are three steps to establishing an enterprise-wide AI agenda in the organization.
- Identify opportunities that promise business viability and technical feasibility for AI adoption
- Standardize AI models, their deployment and lifecycle to ensure ease of adoption
- Reduce or eliminate risks by reviewing AI projects for risks in data, process, and models and building necessary controls to drive trustworthy governance
The above-mentioned framework-based approach helps define the organization roadmap to scale and future-proof AI. Most often, AI is seen as a subject matter of an elite group of scientists. Democratizing AI will need change management to ensure minimum resistance to adoption. Here are some best practices for managing the micro changes.
Accessible to all: AI technology must be available to every employee, irrespective of the role they play, by using a centralized AI platform that could be built to suit different personas such as business heads, data analysts, or systems managers.
Facilitate innovation and collaboration: An environment that encourages employees to think out of the box and create solutions for their business needs without going through complex approval processes can encourage innovation. A gamified design culture that builds collaboration between citizen data scientists and AI experts can pay rich dividends.
Inculcate an AI-First approach: Everyone must be provided with the tools and the motivation to be a problem solver.
Establish an AI academy: Specialized centers of excellence (CoE) and digital enablement platforms can act as catalysts to encourage new ideas and solutions. Reskilling workforce on new-age digital skills by providing them access to latest enablement courses and industry-recognized certifications is a step in this direction.
Create an AI store: A repository of pre-curated array of AI-based platforms, solutions and services can make AI adoption very easy for employees. With self-service options and an experimentation platform, it can become the source for crowdsourcing of models.
For democratization of AI, organizations must measure their readiness with respect to people, processes, and technologies. Each of them could be at different stages of maturity. They must also be wary of the downfall of AI democratization, where technology could be misused or mishandled. Companies must create standards and guidelines needed to ensure that democratization is implemented with the required governance, training, and tra