There’s little doubt that Artificial Intelligence (AI) is a game changer. Over the coming decade, and perhaps far more quickly, AI holds the potential to disrupt virtually every aspect of how we do business.
And change can be intimidating. For many organizations, inertia is a powerful force. “What we’re doing is working just fine” employees may say. Or, “We’ve been following the same procedures for years; why should we change now?” Or, “We’re too busy with our day jobs to figure out how to apply some newfangled technology.”
So how can managers and executives cultivate organizations where AI is not only welcomed, but embraced? How can you foster a culture of innovation and experimentation?
We recommend a 10-step plan that will help you systematically plan, execute and evaluate AI initiatives:
1. Establish an AI steering committee - Every initiative requires planning and oversight. The AI steering committee should be a cross-functional team comprising senior leaders, domain experts, data scientists, and IT professionals. As we note elsewhere, it’s important to recognize that AI will benefit different functions in different ways. As such, you will need broad representation on the steering committee.
2. Define strategic objectives - Strategic objectives for major AI investments generally align with a company’s key competitive priorities. For instance, a bank may focus AI resources on fraud detection, risk assessment, and customer services. An automotive manufacturer may invest heavily in research for self-driving vehicles. And a consumer packaged-goods company may use it for supply chain management and demand forecasting. Walmart, which has more than 100,000 suppliers, went so far as to let an AI-driven chatbot conduct purchasing negotiations, “allowing its procurement team to focus more on strategic relations, exceptions and continuous improvement,” according to Harvard Business Review.
3. Identify and prioritize potential AI use cases - Resources for implementing AI solutions, particularly complex ones, are obviously limited. The steering committee should therefore identify and prioritize AI use cases across different business functions and departments, based on strategic alignment, feasibility, and potential impact on business outcomes.
4. Assess organizational readiness - Organizational readiness is critical. Like any technology, AI cannot be foisted on an organization that is not receptive to it. If the organization is ill-prepared or resistant to AI (perhaps due to concerns over job loss), investments are vulnerable to failure.
There are five groups to identify in assessing AI readiness: stakeholders, sponsors, advocates and allies, naysayers, and partners. You will want to pair them with appropriately qualified individuals to find gaps in acumen, technical ability, and change management.
5. Determine the potential impact, ROI and costs/investments required - As in any business investment, this step entails quantifying the expected benefits and comparing them with the estimated costs, including development and maintenance. Both direct and indirect costs and benefits should be considered.
6. Assess the risks, challenges, and ethical considerations - Risks, challenges and ethical considerations include data privacy breaches, regulatory violations, disruptions caused by organizational change, and unfair treatment due to biases in algorithms or training data. (Please see this AMA blog for a more detailed discussion of risks.)
7. Build the business case - To build the business case, you can identify pain points and ideal solutions, and then determine how AI can get your organization closer to that solution. You should outline the benefits: Will the solution reduce costs, improve efficiency, drive revenue growth, or improve customer satisfaction? Ultimately, the business case aligns the solution with the organization’s overarching goals.
8. Identify and engage stakeholders - You will need to engage stakeholders—including business leaders, end users and subject-matter experts—to solicit feedback and gather insights. Note that this is not an activity to be left entirely to the IT or data-science team. Engaging nontechnical team members in functions such as marketing, sales and customer service can improve problem-solving, identify issues and improve usability and adoption.
9. Train employees - Training and capacity-building programs are essential. They should enable employees to acquire the AI literacy, technical skills, and domain knowledge required to leverage AI technologies in their roles, so they can contribute to the conversation about AI possibilities.
10. Communicate about AI policies, opportunities and wins - Socializing the organization’s AI framework, initiatives and benefits is critical to building support, enthusiasm and momentum behind the technology. As these 10 steps demonstrate, the management and implementation of AI involves large swaths of a typical organization, including executives, managers, and staff. It’s not a technology to be relegated to the IT department.
