The AI Revolution Isn’t About Doing Things *Better* – It’s About Doing Entirely Different Things
The hype around artificial intelligence is reaching fever pitch, but a crucial message emerged from the recent Fortune Brainstorm Tech dinner at CES: AI isn’t just a faster, smarter version of what we already do. It’s a fundamental shift in capability, demanding a complete rethink of how organizations operate. Simply automating existing processes risks “industrializing waste,” as Disney’s Chief Information and Data Officer, Susan Doniz, put it.
Beyond Automation: The Trap of Incremental Improvement
For years, businesses have approached new technologies by seeking incremental improvements to existing workflows. With AI, that approach is a dead end. Deloitte CTO Bill Briggs warns against falling into the “trap” of simply trying to do things “a little bit differently and a little bit better.” Agentic AI – AI that can act autonomously – requires a more radical approach. It’s not about optimizing the horse-drawn carriage; it’s about inventing the automobile.
Consider the manufacturing sector. Traditionally, quality control involved manual inspection or statistically sampled checks. Now, AI-powered vision systems can inspect 100% of products in real-time, identifying defects invisible to the human eye. This isn’t just faster inspection; it’s a fundamentally different approach to quality assurance, leading to reduced waste, improved customer satisfaction, and potentially, entirely new product designs based on the data collected. A recent report by McKinsey estimates that AI-powered quality control could reduce manufacturing defects by up to 90%.
Designing for Failure: The Importance of Kill Switches and Auditability
Embracing this new capability requires a healthy dose of realism. AI systems, particularly agentic ones, will inevitably make mistakes. Hari Bala, CTO of Health Information Systems at Solventum, stresses the need to “design for failure.” This means building in safeguards like “kill switches” to halt runaway processes and robust audit trails to understand *why* an AI made a particular decision.
This is particularly critical in regulated industries like healthcare and finance. Imagine an AI-powered loan application system that unfairly denies credit to a protected group. Without auditability, identifying and correcting this bias would be nearly impossible. The EU AI Act, set to be fully enforced in 2026, will mandate precisely this level of transparency and accountability for high-risk AI applications.
Avoiding Tech Sprawl: Simplification is Key
The rush to adopt AI can easily lead to the same “tech sprawl” that plagued previous technological revolutions. Lauri Palmieri of Salesforce points to the pitfalls of service-oriented architecture and microservices – promising technologies that often resulted in complex, unmanageable systems. The key, she argues, is strong leadership that says, “Just go,” but also a commitment to simplification.
Doniz echoes this sentiment, stating that an “AI-first mindset is firstly about simplification.” Automating existing inefficiencies simply amplifies them. Instead, organizations should use AI to fundamentally redesign processes, eliminating unnecessary steps and streamlining workflows. A prime example is robotic process automation (RPA) combined with AI. RPA can automate repetitive tasks, but when coupled with AI’s ability to learn and adapt, it can optimize entire processes, not just individual steps.
Data: The Fuel for the AI Engine
All this potential hinges on one critical element: data. As Doniz emphasizes, “The data is the fuel and the foundation.” Organizations need well-orchestrated, integrated, secure, and accessible data to power their AI initiatives. This isn’t just about having *more* data; it’s about having *better* data.
Poor data quality can lead to biased AI models and inaccurate predictions. A recent study by Gartner found that poor data quality costs organizations an average of $12.9 million per year. Investing in data governance, data cleansing, and data integration is therefore essential for realizing the full potential of AI.
Urgency and Imperfection: The “11:30” Principle
Finally, Briggs offers a pragmatic perspective: don’t let perfection be the enemy of the good. Borrowing a quote from Lorne Michaels of Saturday Night Live – “We don’t go on because we’re ready; we go on because it’s 11:30” – he urges organizations to embrace a sense of urgency. Waiting for the perfect AI solution will only leave you behind. Start small, iterate quickly, and learn from your mistakes.
This approach aligns with the principles of Agile development and Lean Startup methodologies, which emphasize rapid prototyping and continuous improvement. The key is to start experimenting with AI now, even if it means accepting a degree of imperfection.
Frequently Asked Questions (FAQ)
Q: Is AI going to replace all our jobs?
A: While AI will automate some tasks, it’s more likely to augment human capabilities and create new job roles focused on AI development, maintenance, and ethical oversight.
Q: What’s the first step my company should take with AI?
A: Focus on identifying a specific business problem that AI can solve, and ensure you have access to the necessary data.
Q: How much should we invest in AI?
A: Investment should be proportional to the potential return and aligned with your overall business strategy. Start with pilot projects to demonstrate value before making large-scale investments.
Q: What are the biggest risks of implementing AI?
A: Risks include data bias, lack of transparency, security vulnerabilities, and the potential for unintended consequences. Proactive risk management is crucial.
Did you know? The global AI market is projected to reach $1.84 trillion by 2030, according to a report by Grand View Research.
Pro Tip: Don’t underestimate the importance of employee training. Equipping your workforce with the skills to work alongside AI is essential for success.
Ready to explore how AI can transform your business? Contact us today for a free consultation. Share your thoughts and experiences with AI in the comments below!
Worth a look