Our latest Executive Impact Series article is a collaboration with Anush Naghshineh.
Introduction: The Mistake of Applying AI to Broken Processes
Companies are making a costly error with their AI investments. They’re using AI to automate bad processes before fixing them first. It’s like turbocharging a car with square wheels – you’ll have plenty of power, but you won’t get far until the car tears itself apart.
Recent research indicates that while 78% of organizations utilize AI in at least one business function, fewer than 2% consider themselves mature in its deployment. The gap between adoption and success often comes down to one fundamental mistake: applying AI to fundamentally broken workflows.
Consider this: if your current process involves redundant approvals, unclear handoffs, and data silos, AI cannot resolve these issues. It’ll just help you create new confusion. Companies like Lumen discovered this when they used AI to transform sales processes, cutting preparation time from four hours to 15 minutes. However, they only realized the benefits after redesigning their sales teams’ work processes.
The companies seeing real results from AI aren’t just plugging it into existing systems. They’re answering the tough questions first:
- What if we could start from scratch?
- What would this process look like if we designed it to meet our current needs?
- Which steps add measurable value, and which ones exist because of the status quo?
Enhancement can deliver quick wins, but redesign creates lasting competitive advantages. You need to take a hard look at which business functions will benefit the most from AI automation and then determine if those processes require redesign.
Which processes need redesign vs. enhancement
The decision between redesigning and enhancing isn’t about the technology but the process itself. Digital enhancement introduces technology without fundamentally changing how work gets done, while transformation changes how business is conducted.
Here’s how to tell the difference. If your process has clear, logical steps and the primary issue is speed or accuracy, enhancement may be sufficient. However, if people regularly work outside the official process, if there are multiple approval layers that don’t add value, or if different departments have their own versions of the same workflow, a redesign is necessary.
Consider insurance claims processing. Many insurers initially utilized AI to expedite existing approval processes. But the smart ones stepped back first. They realized the real problem wasn’t slow approvals but redundant reviews, inconsistent documentation, and data scattered across multiple systems. Companies that redesigned first, then added AI, saw revenue increases of 10-15% from streamlined processes.
Manufacturing offers another clear example. HARTING redesigned its configuration process before implementing AI, reducing setup time from 15-20 minutes to just one minute. They redesigned the workflow by reimagining what the ideal configuration could look like.
The pattern is consistent across industries. As McKinsey research shows, AI enables better, faster decisions, but only when companies first address fundamental process issues. Companies that apply AI to broken processes often end up with costly solutions that don’t deliver returns.
Enhancements are beneficial for processes that are well-designed but can benefit from improvements in speed, scale, or precision. Redesign is essential when the process itself is the bottleneck.
The Business Case for Fundamental Workflow Changes
While enhancements can yield quick wins, true competitive advantage often comes from rethinking workflows from the ground up. Fundamental redesign enables organizations to:
- Align operations with shifting market demands.
- Eliminate redundant or unnecessary steps entirely, not just automate them.
- Capture richer, cleaner data streams for more powerful AI models.
- Build customer experiences that differentiate the business.
Leaders should view redesign as an investment, not a cost, delivering greater long-term ROI than piecemeal optimizations.
Change Management Strategies for AI-Driven Process Transformation
Successful redesigns hinge on how well people adapt to new ways of working. Effective strategies include:
- Clear Vision & Communication: Articulate why change is necessary and how it benefits employees and customers.
- Stakeholder Involvement: Engage process owners, frontline staff, and customers early to identify hidden challenges.
- Pilot Programs: Test redesigned processes in controlled environments before full rollout.
- Training & Support: Equip teams with the skills and confidence to use new AI tools and workflows.
- Continuous Feedback: Establish channels to refine processes based on user experience and evolving needs.
Real-World Examples of Successful Redesigns
- Insurance Claims Processing: A major insurer eliminated redundant approvals and standardized documentation before introducing AI-based fraud detection, cutting processing time. A case study of mid-sized insurer Dialzara (~500K customers) explored AI for automated data entry, smart review workflows, and fraud detection. The result: claims processed 40% faster, 25% fewer errors, 30% reduction in labor costs, and a 20% boost in customer satisfaction.
- Manufacturing Quality Control: A global electronics firm redesigned its inspection workflows to collect better defect data, then layered AI visual inspection on top, reducing defect rates. The Advantech case study describes “PowerArena AI” in electronics assembly workflows: cameras capture each workstation’s output and image data, enabling real-time alerts and traceability, improving cycle balance and defect tracing.
- Retail Customer Service: A retailer reimagined returns processes, integrating omnichannel self-service options, then used AI to predict return patterns, improving customer satisfaction and reducing costs. The SupplyChainBrain forum discusses how AI analyzes customer-submitted photos or videos to verify item conditions, enabling instant refunds and faster turnaround.
Conclusion: Decision Framework for Process Transformation
To make the redesign versus enhancement decision, start with three simple questions that cut through the complexity and get to what matters most for your business.
- Do people regularly bypass this process to get work done? If your team has created unofficial workarounds, shortcuts, or “shadow processes,” that’s a clear signal that the workflow is broken. Enhancement won’t fix fundamental design flaws.
- Consider your data situation: Is your process generating clean, useful information or just moving data around? Processes that capture richer, cleaner data streams enable more powerful AI models. If your current workflow creates data silos or requires manual cleanup before the information can be used, redesign it first.
- Explore customer impact: Does this process create meaningful value for customers, or does it exist primarily for internal convenience? Companies that see transformative AI returns focus on processes that directly improve the customer experience or create competitive differentiation.
If you answered ‘Yes’ to the first question or ‘No’ to the others, evaluate a redesign. If your process is fundamentally sound but could benefit from improvements in speed or accuracy, enhancements might be enough.
Remember that enhancement and transformation aren’t mutually exclusive – enhancement initiatives can deliver quick wins while contributing to larger transformation over time. Start with pilot programs, involve frontline staff in the design process, and focus on one process at a time. The companies winning with AI aren’t treating it as a new technology adoption. Instead, they’re using it as an opportunity to build better ways of working. That’s the difference between automation and transformation.