AI, or not AI: What is the question?

Since the emergence of ChatGPT two years ago, AI has rapidly gained popularity. It’s everywhere—news outlets, articles, magazines, books, and podcasts are buzzing with discussions about its benefits. “This is the future,” they proclaim. Governments across the globe have begun establishing regulations and guidelines to define acceptable AI usage. Meanwhile, entrepreneurs and scientists—key players in AI's development—have raised alarms about its potential negative impacts on socio-economic systems. Some articles even suggest that AI could be an overhyped bubble, potentially bigger than the dot-com era. The main criticism? The return on investment (ROI) from AI might not justify its extensive adoption or its perceived revolutionary potential.

As an average person, you may find yourself wondering: Could AI really benefit my business, or is it just another overhyped trend? One of the first things you’ll notice is that AI isn’t as new as it seems. Its foundational breakthroughs date back to the 1950s, though they largely remained confined to academic labs and niche publications. That’s a whole separate topic—one deserving of its own “history of AI” article, along with an exploration of why highly intelligent robots won’t be replacing public servants like in I, Robot anytime soon (at least not in the next 30+ years).

Then, you might ask yourself: How did ChatGPT, only two years old, spark such a surge of “experts” eager to sell their AI expertise? Many of them seem to be newcomers or, at best, specialists with limited experience. Come back in ten years, you might think, when you have more time-tested development and sales results under your belt, Mr. Expert.

Eventually, it dawns on you that the only way to truly understand AI’s potential is to start your own journey—researching, learning, and treating AI as an experimental tool with outcomes that remain uncertain for now.

Questions that this article will try to explore:

1. What is AI?

2. Why do they blend AI with automation? How is AI different from automation?

3. Why do I need to focus on defining my processes first before automating and implementing AI?

4. What is a standardized process and why is it important?

5. What are Value-Added (VA) and Non-Value-Added (NVA) activities, and why are they significant?

6. How can I build my own AI, or should I buy a ready-made solution from a software firm?

7. When should I expect to see the first outcomes?

8. What is the cost of implementing AI?

9. If I successfully implement AI, can I lay off my employees and save money by doing so?

   
   

1.    What is an AI?

Artificial Intelligence (AI) refers to computer systems designed to perform tasks that typically require human intelligence. These tasks include:

1. Learning: Acquiring knowledge from data (e.g., machine learning).

2. Reasoning: Making decisions or solving problems based on rules or patterns.

3. Perception: Understanding and interpreting sensory inputs like images, sounds, or text.

4. Language Understanding: Communicating effectively in natural human languages.

AI systems can range from narrow AI, which specializes in specific tasks (e.g., recommendation systems, facial recognition), to the theoretical concept of general AI, which would possess human-like cognitive abilities across a wide range of activities.

2.    Why they blend AI with automations?

How is AI different from automations? 

AI and automation are closely related but distinct concepts. They are often blended because they complement each other in enhancing efficiency, flexibility, and decision-making.

Automation:

• Definition: Automation involves the use of machines or software to perform predefined, repetitive tasks with minimal human intervention.

• Characteristics:

  • Rule-based: Follows a fixed set of instructions or workflows.

  • Predictable: Operates in structured environments where processes are consistent.

  • Examples:

    • Robotic Process Automation (RPA): Automating data entry, invoice processing.

    • Assembly line robots: Repeating the same steps in manufacturing.

Artificial Intelligence (AI):

• Definition: AI involves systems that simulate human intelligence, enabling them to learn, reason, and adapt.

• Characteristics:

  • Adaptive: Learns from data and can improve over time.

  • Decision-making: Handles uncertainty and makes complex decisions based on patterns.

  • Examples:

    • Chatbots with natural language understanding.

    • Predictive maintenance in machines.

How They Work Together:

• Enhanced Automation with AI: AI is integrated into automation to make it smarter, flexible, and capable of handling dynamic, unstructured environments. For example:

  • Dynamic Decision-Making: AI-enhanced RPA can decide how to route a customer service ticket based on sentiment analysis.

  • Learning & Adaptation: AI systems in automation can analyze performance data to optimize workflows.

• Scaling AI with Automation: Automation handles routine tasks, freeing AI to focus on more complex challenges or high-value activities. For instance:

  • AI detects anomalies in data (intelligent decision-making), while automation handles the response (e.g., alerting a technician).

      

Key Differences:

Blending the two creates systems that not only do tasks efficiently (automation) but also think about how to improve them (AI). This synergy is foundational to technologies like self-driving cars, where automation controls the vehicle and AI processes complex inputs to make decisions.

3. Why I need to focus on defining my processes first before automating and AI?

Defining your processes is a critical step before implementing automation and AI. It ensures that you are addressing the right problems, optimizing efficiency, and avoiding unnecessary complexity. Here’s why focusing on process definition first is essential:

1. Understanding the Current State

• Identify Bottlenecks: Clearly mapped processes help pinpoint inefficiencies, redundancies, and areas ripe for improvement.

• Standardization: Ensures consistency in how tasks are performed across your organization, which is crucial for automation and AI to function effectively.

2. Clarity for Technology Application

• Customization: AI and automation tools work best when tailored to specific workflows. If processes are unclear, it’s harder to configure tools to meet your needs.

• Avoiding Missteps: Automating or applying AI to poorly designed processes can amplify inefficiencies rather than resolve them.

3. Aligning Objectives

• Business Goals: Defined processes ensure that automation and AI initiatives align with organizational objectives, whether it's reducing costs, improving customer experience, or increasing speed.

• Scalability: Clear processes provide a roadmap for scaling automation or AI solutions as the organization grows.

4. Facilitating Data Collection

• Input Data: AI requires high-quality, relevant data. Defined processes clarify where and how data is generated, ensuring AI models are trained effectively.

• Feedback Loops: Understanding workflows enables proper design of feedback mechanisms for continuous improvement.

5. Cost Efficiency

• Avoiding Overengineering: Without clear processes, you may invest in overly complex solutions for tasks that don’t require them.

• Focused Investment: You can prioritize areas where automation and AI will provide the greatest return on investment (ROI).

6. Change Management

• Employee Buy-In: When processes are well-defined, it’s easier to communicate changes to employees and get their support.

• Transition Ease: A clear understanding of workflows makes it smoother to integrate new technologies and train teams.

4. What is standardized process and why?

A standardized process is a structured and documented way of performing a task or series of tasks in a consistent, repeatable manner. It ensures that everyone involved follows the same steps, uses the same methods, and adheres to the same rules or guidelines to achieve predictable and reliable results. Put it simply, it’s the process everyone believes is the right thing to do their work.

A standardized process should be:

1.      Documented: Clearly written procedures or flowcharts outline each step in the process.

2.      Consistent: Everyone performs the process in the same way, regardless of who is involved.

3.      Repeatable: The process can be executed multiple times with the same results.

4.      Measured: Key performance indicators (KPIs) or metrics are often used to track and evaluate performance.

5.      Optimized: Processes are designed to minimize waste, inefficiencies, or errors.

Benefits of Standardization in Automation and AI - the WHYs

• Foundation for Automation: Automation thrives on well-defined, repeatable processes.

• Data Consistency: AI requires clean and consistent data, which standardized processes help ensure.

• Reduced Errors: Standardization eliminates variation, which can lead to fewer mistakes.

A standardized process is the backbone of operational excellence, enabling efficiency, quality, and scalability. It’s particularly critical when integrating technologies like automation and AI to ensure smooth implementation and maximum benefit.

5. What is VA and NVA and why?

VA (Value Added) and NVA (Non-Value Adding) are terms used in Lean and Six Sigma process improvement, to categorize tasks based on their contribution to customer value.

NVA are activities that don’t directly contribute to delivering the desired outcome or service to the customer.

VA are activities that directly contribute to delivering the desired outcome or service to the customer.

VA and NVA are important for identifying waste, improving customer satisfaction, resource optimization, and are foundation for automation and AI.

Focusing automation or AI on NVA or wasteful tasks can amplify inefficiencies if the process isn’t optimized first. Automating before streamlining you simply will exponentially multiply errors delivering it faster to your customers – big mistake, don’t do it.

How can I build my own AI or I should buy it ready from a software firm?

Deciding whether to build your own AI or purchase a pre-built solution depends on your specific needs, resources, timeline, and long-term goals.

Build your own AI if your business model is structured around AI and you have the tech expertise, development resources, and budget. Pros of doing that are customization, control and ownership, you’ll have your own experimentation platform and a valuable asset. Cons are high-initial costs, time consuming, and resource-intensive for maintenance and updates.

Buying pre-built is easier and faster, do it if you aim for quick deployment, lower initial costs, proven solutions and support included. Cons are that customization will be limited and dependence on vendors - you’ll be dependent on vendors limiting flexibility and increasing costs. The biggest concern buying AI could be data ownership where vendors could control or access your data.

Hybrid approach is to customize pre-build solution or to partner with a software firm to co-develop a solution balancing speed, budget and customization.

6. When should I expect first outcomes?

Expect first signs of prototype success within 1-3 months. Visible business impact and operational results around 6-12 months. Full transformation and integration impact may take 12+ months, depending on complexity and scale.

7. What is the cost of doing it?

The cost of an AI project can vary widely depending on its scope, complexity, and scale. The total cost of an AI project will depend on your goals, the technology stack you choose, and the resources you already have in place. Demand is high for these types of projects and you should expect higher than normal pricing. To give you a perspective, startup initial AI projects may go up to $100k, mid-size businesses could expect up to $500k for AI applications and large enterprise-wide AI initiatives to exceed $1M budgets. Don’t forget maintenance, hidden and opportunity costs. 

8. I’ve implemented successfully an AI, can I now permanently lay off my employees and save money doing so?

While successfully implementing AI can lead to cost savings and increased efficiency, laying off employees as a direct result should be approached cautiously. AI implementation doesn't automatically justify workforce reduction, and in many cases, employees remain critical for optimizing and managing AI systems.

Before even thinking about layoffs, you should consider upskilling, reskilling, role redeployment, hybrid roles, and you should perform cost-benefit analysis (short term, long term, risks). Layoffs after automatons and AI implementation will reduce remaining staff morale and productivity, will lead to negative public perception or backlash form customers and stakeholders.

Layoffs should be a last resort, not the default approach, after implementing it. Consider automatons and AI as a tool to enhance productivity and innovation. We can help you prepare and decide what to do.

Ultimately, the key question is one you need to discover for yourself. As a starting point, consider shifting your focus away from the constant search for the next technological "magic wand" and instead revisit the roots of your challenges. Often, the pursuit of simplicity reveals that the core issues lie in a lack of standardized processes and fundamental organizational management. Perhaps it’s time to concentrate more on optimizing your processes, empowering your people, and cultivating the right environment—rather than relying solely on tools.

Whether you choose to dive into the rabbit hole of AI, we’re here to help you become AI Ready. Remember, with or without AI, the best approach is to start at the roots: standardize, streamline, and optimize your processes and procedures. We’re here to support your process improvement efforts, including SOP creation, process mapping, and DMAIC implementation.

Ready? Let's talk.