As any US business executive will tell you Artificial Intelligence (AI) is not just a passing trend; it’s a transformative capability worth developing internally as it will differentiate each business, and hit the bottom line, in a way we haven’t seen with tech yet. I guarantee you every single fortune 500 meeting has a portion of their calls on “This what we’re doing with AI, here is our progress to date.”
However, managing AI expectations within your organization, especially with your exuberant forward-thinking CEO and Board of Directors, requires a more strategic, well thought out and logical approach. The key is to educate and create realistic AI use cases while enabling your CEO to clearly understand and articulate how AI will be leveraged now and in the future at company XYZ.
AI Best Practices
- Clear set of AI goals and metrics to ensure it’s a real test
- Accept that pushing the boundaries of AI, requires failures & learning
- AI may take time to tweak, improve and rollout, realistic timelines are critical
- Start small, test, test, dogfood it, get feedback from your AI champions, test again
- If the test use case doesn’t increase efficiency, or sales (or path to either) kill it
- Have a clear check list for when AI is ready for rollout as well as when to give up
- Complete an AI Risk Analysis and AI Risk Assessment
- Benchmark your current AI Risk to any future AI rollouts
- Internal Legal team-review all software contracts for AI data use,
- AI Lawyers prepare internal AI use policy for human resources,
- Review use of AI in document preparation (see ChatGPT), ensure all client deliverables that have any use of AI are still legal and / or noted in client deliverables
With these best practices in mind. The following ten steps will help you stick with a strategy and keep everyone on the same page of AI realism vs. AI sizzle and flash puffery.
10 Steps to manage AI expectations
- Advocate for a Dedicated AI Team and Budget
AI development requires a dedicated team and budget. While it doesn’t have to be excessive, starting with a team of around five data scientists or analysts is a good approach for most companies.
- Educate on Reality vs. Hype
It’s crucial to educate your CEO and executive staff on the realities of AI versus the hype. If necessary, bring in external consultants and spend several hours with your CEO to demystify AI, including some of the underlying math. This helps ensure AI doesn’t remain an esoteric concept.
- Focus on Business Outcomes
Lead with the business outcomes you aim to achieve with AI. Remember, AI and Machine Learning (ML) are not panaceas; they won’t transform your business into something it’s fundamentally not. Business leaders should think creatively about how AI can support their goals.
- Prepare for Iteration and Failures
Developing AI capabilities requires trying numerous projects to build the necessary expertise. Set the expectation with your CEO that rapid iteration and a high volume of projects will be the focus for the first six months. Be prepared for several failed initiatives along the way.
- Emphasize Data Availability
AI and ML are heavily dependent on data availability. If your company lacks sufficient data, building effective models will be challenging. Set the expectation that a cultural shift towards data-driven decision-making is necessary, and the CEO and executive staff should lead by example.
- Revisit and Utilize Existing Data
Examine the data your company already has but hasn’t utilized effectively. For instance, many companies have machine sensors generating large amounts of data that go unused. Revisit these data stores and explore new possibilities.
- Ensure Data Quality and Infrastructure
AI projects are doomed to fail if the underlying data is poor. Continuously improve your data infrastructure to ensure all data sources needed for business outcomes are in good shape. This doesn’t mean embarking on a massive “data lake” project, but rather ensuring data quality incrementally.
- Clarify AI vs. Analytics Needs
Often, when people say they want AI, they actually mean they want better analytics—more reliable and quicker answers to their questions. Test this assumption and consider hiring strong analysts if this is the case.
- Pair Business Leaders with Analysts
Pair key business leaders with analysts who can answer data-related questions and educate them on interpreting data. Surprisingly, many leaders are unfamiliar with terms like p-value and correlation coefficient. Making leaders moderately data-savvy can significantly impact organizational culture.
- Understand AI, ML, and Deep Learning
- Artificial Intelligence (AI): When machines perform tasks typically considered the realm of humans.
- Machine Learning (ML): When a machine improves its performance on a task with experience.
- Deep Learning: A subset of ML that uses neural networks.
One of the other best practices your AI champions should do is stay on top of new use cases that may not be necessarily in your industry. It is impossible for most companies to test all the new AI coming out. But you can watch, learn and apply what others are doing and hopefully do better with fewer errors.
Examples of Successful AI Implementation
- Healthcare: IBM Watson Health
- Case Study: IBM Watson Health has collaborated with Memorial Sloan Kettering Cancer Center to improve healthcare outcomes. Watson processes vast amounts of medical literature and patient data, aiding oncologists in diagnosing and recommending treatment options.
- Impact: Reduced diagnosis time from weeks to hours and increased accuracy in identifying cancer types1.
- Finance: JPMorgan Chase
- Case Study: JPMorgan Chase’s Contract Intelligence (COiN) platform uses AI to review legal documents and extract essential data points, enhancing risk management and fraud detection.
- Impact: Reduced document review time from 360,000 hours annually to seconds and improved fraud detection accuracy1.
- Manufacturing: Siemens
- Case Study: Siemens integrates AI with its production lines for predictive maintenance and process optimization.
- Impact: Reduced unplanned downtime by up to 50% and increased production efficiency by 20%1.
- Retail: Amazon
- Case Study: Amazon uses AI to personalize customer experiences and optimize inventory management.
- Impact: Enhanced customer satisfaction and streamlined inventory processes1.
- Transportation and Logistics: UPS
- Case Study: UPS uses AI to optimize delivery routes and improve logistics efficiency.
- Impact: Reduced fuel consumption by 10 million gallons annually and improved delivery times2.
- Agriculture: John Deere
- Case Study: John Deere employs AI in its precision agriculture technology to optimize crop yields.
- Impact: Increased crop yields by 15% and reduced resource usage2.
- Entertainment: Netflix
- Case Study: Netflix uses AI to recommend content to users based on their viewing history.
- Impact: Increased user engagement and retention rates2.
By following these steps and learning from successful AI implementations, you can effectively manage AI expectations with your CEO and pave the way for successful AI integration in your organization.
Tags: #CEO #BOD #AI #AIadoption #AIRisk #AIStrategy #Boardofdirectors #AItesting #AIusecase
