Blog

AI technologies are rapidly maturing as a viable means of enabling and supporting essential business functions. However, creating business value from artificial intelligence requires a thoughtful approach that balances people, processes, and technology.

AI comes in many forms: machine learning, deep learning, predictive analytics, natural language processing, computer vision, and automation. Organizations must first start with a solid foundation and a realistic perspective to determine the competitive advantages that an AI implementation can bring to their business strategy and planning.

“Artificial intelligence encompasses many things, and there are many exaggerations and sometimes exaggerations about how intelligent it really is,” said John Carey, executive director of management consultancy AArete.

An early implementation of AI is not necessarily a perfect science and may need to be experimental first – starting with a hypothesis, then testing, and finally measuring the results. Early ideas are likely to be flawed, so an exploratory approach to gradually adopting AI is likely to produce better results than a big bang setting. To avoid mistakes, these 10 steps can help ensure a successful AI implementation in your company.

1. Establish data flow

Hands-on conversations about AI require a basic understanding of how data drives the whole process. “Data flow is a real and challenging barrier – more than just tools or technology combined,” says Penny Wand, technology director at IT consultancy West Monroe. In a 2020 report by Forrester Research, Forrester Research found that 90% of data and analytics decision makers surveyed view increasing use of data insights as a business priority, but 91% said leveraging those insights is challenging Company represents. Forrester went on to report that the gap between realizing the meaning of insights and actually applying them is mainly due to a lack of advanced analytical skills necessary to produce business results. “Understanding this maturation process and driving sustainable change requires understanding and support from leaders,” noted Wand.

2. Define your primary business drivers for AI

“To successfully implement AI, it is critical to learn what others are doing inside and outside of your industry to generate interest and stimulate action,” said Wand. When developing an AI implementation, identify key use cases and assess their value and feasibility. Also, consider your influencers and who should be championing the project, identify external data sources, figure out how you could externally monetize your own data, and create a backlog to ensure that the project’s momentum is maintained.

3. Identify areas of opportunity

Focusing on business areas with high variability and significant payoff, advised Suketu Gandhi, partner at Kearney, a consulting firm for digital transformation. Teams of business stakeholders with technology and data expertise should use metrics to measure the impact of an AI implementation on the company and its employees.

4. Assess your internal skills

Once use cases have been identified and prioritized, business teams need to figure out how those applications align with your company’s existing technology and human resources. Education and training can help fill the technical skills gap in-house, while corporate partners can facilitate on-the-job training. In the meantime, external expertise could help accelerate promising AI applications.

5. Identify suitable candidates

It’s important to limit a broad range of options to practical AI deployment – for example, invoice matching, IoT-based facial recognition, predictive maintenance of legacy systems, or customer buying habits. “Be adventurous,” said Carey, “and include as many people as possible [in the process] as you can.”

6. Piloting an AI project

In Gandhi’s view, turning an AI implementation candidate into an actual project requires a team of AI, data, and business process experts to collect data, develop algorithms, provide scientifically controlled releases, and measure impact and risk .

7. Create a basic understanding

The successes and failures of early AI projects can help improve understanding across the organization. “Make sure you keep people informed to build trust and tie your business and process experts with your data scientists,” said Wand. Also, realize that the road to AI begins with understanding the data and good old-fashioned rearview mirror reporting to lay a foundation for understanding. Once a baseline has been established, it is easier to see how the actual AI implementation confirms or disproves the original hypothesis.

8. Scale incrementally

The overall process of creating momentum for an AI deployment starts with small wins, argued Carey. Incremental successes can help build trust across the organization and inspire more stakeholders to conduct similar AI implementation experiments on a stronger, more established basis. “Adjust algorithms and business processes for a scaled publication,” suggested Gandhi. “Embed [them] in normal business and technical operations. “

9. Bring general AI skills to maturity

As AI projects scale, business teams need to improve the overall life cycle of AI development, testing, and deployment. To ensure sustainable success, Wand offers three core practices for developing overall project skills:

• Build a modern data platform that streamlines the collection, storage, and structuring of data for reporting and analytical insights based on the value of the data source and desired business performance metrics.
• Develop an organizational design that sets business priorities and supports the agile development of data governance and modern data platforms to drive business goals and decision making.
• Create and build the overall management, ownership, processes, and technology required to manage critical data items with a focus on customers, suppliers, and members.

10. Continuously improve AI models and processes

Once the overall system is in place, business teams need to identify opportunities to continuously improve AI models and processes. AI models can deteriorate over time or in response to rapid changes due to disruptions like the COVID-19 pandemic. Teams must also monitor feedback and resistance to AI deployment from employees, customers, and partners.

Coexistence with machines

Problems arise at every step of the AI ​​implementation process. “The tougher challenges are the human ones, which has always been the case with technology,” said Wand.

A steering committee should be set up to share the results and represent the key functional areas of the company, she added. The introduction of organizational change management techniques to promote data literacy and trust between those involved can make a major contribution to overcoming “human” challenges.

“The AI ​​skill can only mature as quickly as your entire data management maturity,” advised Wand, “so create and execute a roadmap to move these skills in parallel.”