Process Management

The Secret to Successful AI-Driven Process Redesign

/ Hans van 't Riet / Leave a comment

Building on my previous article, How to Marry Process Management and AI, I take this issue a step further by leveraging insights from the Harvard Business Review (Jan–Feb 2025) article by H. James Wilson and Paul R. Daugherty. These authors, also known for Human + Machine—a must-read for understanding the future of work (a full book review is available on www.bestofdigitaltransformation.com)—explore how AI is reshaping process redesign.

Their article focuses on AI and the Evolution of Kaizen. Initially, I found the parallel between AI and kaizen (continuous improvement) intriguing, but the more I reflected on it, the clearer it became: AI enables humans to make continuous, incremental improvements to processes.

Key Themes:

  • The Toyota Production System, built on kaizen, has long enabled incremental process improvements.
  • Kaizen 2.0, powered by AI, allows employees to leverage data-driven insights to optimize workflows.
  • The article explores how companies use AI to redesign processes, empower employees, and drive business transformation.

In this newsletter, I borrow great examples from the article and add my own insights on leveraging AI for process redesign.

Empowering Employees Throughout the Enterprise

Examples:

  • Mercedes-Benz’s MO360 Data Platform connects plants globally, enabling real-time AI-powered insights for shop-floor workers.
  • Mahindra & Mahindra’s production workers use AI virtual assistants for step-by-step robot repair guidance, reducing downtime and improving morale.
  • Companies like Mercedes-Benz invest in AI training programs (e.g., Turn2Learn), equipping employees with skills in prompt engineering and natural language processing.

Insights:

  • Empowering employees with AI starts with trustworthy and well-managed data, as data quality is critical for AI effectiveness.
  • AI-driven tools eliminate reliance on predesigned reports, allowing employees to interact with data in their own language and gain real-time insights.
  • The ability to ask AI the right questions is a crucial skill, and training employees in prompt engineering is essential.

Redesigning Scientific Processes

Examples:

  • Gen AI is revolutionizing pharmaceutical R&D, reducing waste, accelerating drug discovery, and enhancing quality control.
  • Merck employs AI-generated synthetic image data, reducing false rejects in drug manufacturing by 50%.
  • Absci’s AI-driven zero-shot learning creates new antibodies in silico, cutting drug development from six years to 18 months.

Insights:

  • AI accelerates not only operational processes but also scientific research, leveraging vast, fast access to data.
  • AI rapidly simulates multiple potential solutions, significantly accelerating the research cycle.
  • A remarkable example: Microsoft recently helped identify a lithium alternative for batteries, reducing lithium consumption by 70%—an achievement made possible by AI screening 32 million materials in a single week, a process that would normally take years.

Augmenting Creative Processes

Examples:

  • Colgate-Palmolive, Nestlé, and Campbell’s use AI to validate product ideas and conduct market research.
  • Coca-Cola integrates GPT-4 and DALL-E, allowing digital artists to generate AI-assisted branding materials.
  • NASA’s AI-driven CAD process reduces design cycles from weeks to hours, producing lighter, stronger components for space missions.

Insights:

  • AI enhances creativity in product development, marketing, and design.
  • AI can generate multiple design options, allowing humans to curate and refine the best ones.
  • AI-generated content is transforming marketing—I personally use DALL-E to create visuals for my newsletter instead of manually searching for images.

Animating Physical Operations

Examples:

  • Sereact’s PickGPT enables warehouse robots to follow natural language commands, making robotics more accessible to non-technical employees.
  • Digital twins—virtual models of real-world systems—are used in preclinical drug testing, factory optimization, and hospital operations.
  • Atlas Meditech’s AI-driven virtual brain models allow surgeons to practice on patient-specific digital twins before real-life procedures.

Insights:

  • AI integrates with sensors, enabling robots to collaborate seamlessly with human workers.
  • AI optimizes human-robot collaboration, ensuring each group focuses on their strengths.
  • Digital twins provide simulated environments for process planning and workforce training.

Autonomous Agents

Examples:

  • AI agents are evolving to autonomously make decisions and take action.
  • DoNotPay’s AI agent automatically identifies unnecessary subscriptions and negotiates lower bills.
  • Walmart, Marriott, and Nestlé use AI for inventory, booking, and supply chain optimization.
  • AI agents display human-like reasoning in three ways:
    1. Goal-oriented behavior – Acting independently to achieve objectives.
    2. Logical reasoning & planning – Breaking tasks into structured steps.
    3. Long-term memory & reflection – Learning from past interactions to enhance decision-making.

Insights:

  • AI agents are becoming more powerful, handling complex process optimizations independently.
  • Salesforce’s Agentforce AI resolves customer service issues autonomously—without being pre-scripted.
  • AI-based agents will transform Robotic Process Automation (RPA):
    • RPA handles repetitive tasks, with structured data.
    • AI agents tackle complex tasks involving both structured and unstructured data.

Ecosystems of Autonomous Agents

Examples:

  • Complex tasks often require multiple AI agents working in unison, rather than a single AI performing isolated tasks.
  • Mortgage underwriting: AI agents analyze documents, check compliance, and generate loan recommendations in parallel.
  • Google & Stanford’s AI simulation demonstrated that autonomous agents can develop human-like decision-making and learning.

Insights:

  • End-to-end process automation is still a challenge, given the many variables and process variations.
  • Instead of full-process automation, companies should integrate AI into specific tasks that enhance overall workflows.
  • AI agents can collaborate, forming an ecosystem that continuously learns and improves over time.

Conclusion: AI-Driven Process Redesign Remains Human-Centered

Key Takeaways:

  • AI does not replace humans—instead, it augments employees, enabling continuous improvement at scale.
  • AI allows employees to focus on strategic decisions, while AI agents optimize repetitive and analytical tasks.
  • Successful AI adoption depends on leadership-driven empowerment, ensuring AI tools enhance human creativity rather than replace it.
  • The future of kaizen is AI-augmented, human-led, and continuously evolving, as AI and human expertise merge to drive business transformation.