AIviSORIAT

A Long but Rewarding Journey: Organizational Learning in the Age of AI

In the rapidly evolving landscape of artificial intelligence (AI), business leaders face a critical question: How can we harness the transformative power of AI to drive innovation, efficiency, and sustainable growth? While technical expertise is valuable, the strategic integration of AI across the organization is paramount.

So, do all business leaders and practitioners need to be machine learning engineers or data scientists? I think no. Do they need to understand how AI works and what it can/not do for their business model - and is this possible without overstretching the organization and people? I absolutely think yes. 

Reflecting on … Many conversations I had recently, e.g. at Cannes Lions and the Puls Automobil Kongress, I have observed that companies are advancing at different speeds and depths of organizational learning. Some embrace AI and ensure everyone understands and uses the technology to the degree required for their role. Some prefer a “wait and see” approach. Many are in-between.

Well, but … It’s a lot to take in. New AI models are being built and improved at high speed. And new ways to use them, e.g. agents and agentic workflows, are being constantly innovated. With organizations already stretched these days, how to make room for learning about and adapting AI?

My thoughts … It's necessary and possible to embrace AI. Necessary, because AI is a powerful toolkit that makes businesses and people more efficient, productive, and frees the human mind from boring and repetitive tasks to focus on innovation and human interactions. This will inevitably lead to better business performance.

It's possible by continuously investing time in learning. It takes about 10,000 hours to be a world-class expert (e.g. olympic level sports), but only about 20 hours to learn the basics of a new skill (e.g. asking for directions to and ordering food at a restaurant in a new language). 

Putting this principle into practice: If you invest 2.5 - 5 hours every week into learning about AI, you will have reached a basic level of understanding within 1-2 months. Continuous learning will further develop your skills and knowledge over time.

By embracing a strategic approach to AI adoption and fostering a culture of continuous learning, business leaders can position their organizations for long-term success in an increasingly AI-driven world.

Ultimately … It's a choice for both leaders and practitioners. And for organizations, to create room and resources for everyone to go on this learning journey. I think this is one of the best choices you can make in the age of AI.
 

"The What":

Don't Start with AI: The Key to Preparing Your Business for the AI Revolution

In the current climate of AI excitement, it's easy to get swept away by the buzz and start looking for ways to implement AI into your business without a clear strategy. However, the most effective approach doesn't begin with AI itself. Instead, it starts with understanding the core of your business.

Begin with Your Business Fundamentals

Consider the example of Amazon. Jeff Bezos has consistently emphasized that Amazon's core focus has always been and will always be about offering a vast selection, competitive prices, and rapid delivery. This fundamental truth guides their every decision, including their AI strategy.

The first step to preparing your business for AI is to identify your core principles. What is it that your business truly excels at? What is the central value proposition that you offer your customers? Once you have a firm grasp of this core, you can begin to explore how AI can empower and accelerate it.

AI as an Enabler, Not a Solution

Once you've identified your core business focus, the next question is whether AI can enhance its performance or efficiency. Can AI help you provide even more choices, faster delivery, or better prices? The key is to see AI as an enabler, not a solution in itself.

To determine if AI can help, start by examining if your business challenge can be converted into an AI challenge. What type of data (structured or unstructured) do you have available? What problem do you want the AI to solve?

Choosing the Right AI Approach

Next, explore the available AI options. Are there pre-trained AI models that suit your needs? Or foundation models that you can easily adapt to your purpose? Or do you need to develop an AI model from scratch? McKinsey refers to this decision-making process as the "taker-shaper-maker" logic. There is more than this framework to guide your decision, but this is a good thought-starter.

Test, Measure, Learn, and Enjoy the Journey

The AI implementation process is iterative. Test your AI solution, measure its performance, learn from the results, and continually improve. Remember to have fun along the way. The process of exploring and integrating AI into your business can be a rewarding journey of discovery and innovation.

In conclusion, the most effective way to prepare your business for AI isn't to start with AI. Instead, begin by understanding your core business principles. Use AI as a tool to empower and accelerate what you already do best. With a clear strategy and a focus on your core values, you can leverage AI to achieve greater success.

"The How":

Choosing the Right AI Projects: The "Fab Four" Framework

In today's AI-driven landscape, organizations face a vast array of possibilities. With so many options, how do you select the right AI projects that will genuinely propel your organization forward?

In my experience working with numerous companies and teams, the "Fab Four" framework has proven to be a valuable guide. Remember, this isn't a rigid roadmap, but rather a set of principles to tailor to your specific circumstances.

1. Significant Business Impact

The first question is straightforward: Does the project have a significant impact on your revenue, margin, cost base, or product/service performance? Steer clear of "shiny objects" and "first evers" unless you're engaged in foundational AI research. Prioritize projects that deliver tangible business benefits.

2. Concrete Implementation

Once you've chosen an impactful business objective, the next step is to make it concrete. Is the description of what the AI should do specific enough to enable the selection and training of the right model? What type of data is needed? What kind of model is suitable? What are the AI's optimization goals? For instance, a vague requirement like "an AI avatar that sells my products to everyone" is challenging to translate into concrete design requirements. In contrast, "a chatbot that informs customers about all my product features, based on the existing database of product manuals" is much easier to design and build.

3. AI Availability

After step 1 and 2 you have selected an impactful project, and specified the AI design requirements. Now it's time to assess the type of AI required. Are there pre-trained AI models ("takers") available? Or foundation models ("shapers") that you can fine-tune for your specific needs? Do you need to build an AI model from scratch ("makers")? Takers and shapers typically involve no-code or low-code solutions, while makers necessitate machine learning engineering and data science expertise.

4. Production Readiness

Finally, you should ask yourself the question: Can the solution be deployed into large-scale daily operations? Do you have the necessary data and AI infrastructure? For example, you might possess two years of historical CRM and sales data to train an AI model to predict future sales based on CRM signals four weeks out. However, if your retail network only reports this data in a batch four weeks after each quarter's end, the model won't be able to predict future sales in real-time operations.

In conclusion, the ideal AI project portfolio focuses on business impact, has a concrete design, leverages "taker" or "shaper" types of AI whenever possible, and is ready for production. By adhering to these principles, you can ensure that your AI initiatives contribute meaningfully to your organization's success. And that your teams have a feeling of accomplishment and fun along the way.

Learnings, Insights, and Observations on AI:

September 2024: Fueled by AI conversations with Munich's digital leaders!

I’m grateful for the invitation to speak at PWC’s digital leader’s lunch in Munich. A big thank you to Marcus Worbs from PWC and Sascha Bilen from Careerteam for organizing this insightful gathering.

It was energizing to connect with such an engaged group of leaders from diverse industries - Consulting, Automotive, Consumer Goods, Insurance, Media, Startups, Executive Recruiting, and more. The conversation was buzzing, all centered around AI and how to accelerate its adoption and usage across organizations to gain a competitive edge.

Some key questions that sparked lively debate:

• Can you afford to wait with AI adoption? My take: No. The technology advances daily, and it takes time for an organization to fully embrace AI and enable every function and member.
• Do business leaders need to understand AI? If yes, how deep? My view: The minimum is understanding what business problem can be solved by what type of AI, and how to choose the right tech and talent to achieve it.
• Should AI change be driven from a separate “AI unit” or from within the corporate functions? I believe: The closer to the heart of the business, the better. So, preferably within the functions themselves.
• Do you need to win hearts or minds first for AI? My answer: Both.

I left feeling delighted and inspired by this exchange and the positive mindset of all these leaders towards AI. It's clear that Munich's business community is ready to harness the power of AI to drive innovation and growth.

October 2024: AI for Automotive: Wow! The “How” is Now.

On October 24, 2024, I attended the McKinsey + Google "AI for Automotive" event in Munich. What a difference 8 months makes! Back in February, the focus was on potential. This time? Results. Real-world applications, tangible outcomes, and a clear shift from "Wow, that's cool" to "Wow, that's working!"

The energy was palpable, with fantastic insights from McKinsey, BMW, Bosch, and Beam.ai. Here are my key takeaways:

• Foundation Models as Building Blocks: Forget reinventing the wheel. Powerful, multimodal foundation models are readily available, making it easier than ever to tailor AI solutions for specific automotive use cases.
• Data is King (and Queen): High-quality data is essential. And when working with customer data, ethical considerations and consent are paramount.
• Agents are the Real Deal: Goal-based agents are driving serious efficiency gains (up to 80%!) and cost reductions. The trick? Break down complex problems into smaller tasks, with dedicated agents for each.
• AI as a Force Multiplier: AI isn't about replacing humans; it's about empowering them. The most successful implementations integrate AI with existing processes, technologies, and – most importantly – people.
• Hybrid Operating Models for the Win: A balanced approach, with a central Center of Excellence (CoE) and AI expertise embedded across business units, seems to be the sweet spot for organizational AI adoption.
• Democratization of AI: Make AI tools accessible to everyone! Let people experiment, explore, and overcome their fear of the technology.
• User-Friendly Interfaces are Key: A conversational interface or LLM can make even the most complex AI models (like a digital twin of a factory) accessible to a wider audience.

The speed of progress is mind-blowing. I can't wait to see what the next 6-12 months bring for AI in the automotive industry!

November 2024: The Spark Ignites: Key Takeaways from Germany's Digital Award

Thank you to McKinsey and Handelsblatt for inviting me to attend The Spark, The German Digital award, in Berlin this week! This year's theme, "AI for business", truly sparked my imagination. The evening was buzzing with inspiring presentations and insightful conversations. A huge congratulations to all the nominees and, of course, the winners – you can check them out on the spark website!

Here's what I took away from this electrifying event:

1. A Vibrant AI Community: The DACH region boasts around 1,000 AI startups, according to McKinsey, and the energy in the room could be felt. It's clear that there's a thriving community of innovators pushing the boundaries of AI in the German-speaking region.

2. From Platforms to Specialization: We can build on the foundation of big AI platforms. The focus now is on specialized applications, tailoring AI solutions to businesses of all shapes and sizes.

3. Real-World Impact: The top nominees and winners showcased the incredible potential of AI across diverse sectors. From Spread accelerating mechatronics engineering to Aedifion optimizing building management for CO2 reduction and Aignostics revolutionizing biopharma research, these companies are making a real difference.

4. Finding the Sweet Spot: The key to success lies in striking the right balance between differentiation and scale. This can be achieved by developing AI solutions for entire categories (e.g., beauty/skincare/health, mechatronics) or critical business functions (e.g., purchasing, legal).

5. Navigating the Regulatory Landscape: The biggest concern? AI regulation. While some regulation is necessary, excessive or overly vague rules can stifle innovation and discourage investment.

6. The Road Ahead: There's widespread optimism about the future of AI. Advancements in domain-specific foundation models and spatial agency (robotics/”physical AI”) will unlock even more powerful and economically viable business models. Despite significant investment, we still need to invest more in AI to fully realize its potential.

I left The Spark feeling incredibly inspired by this community of innovators, investors, and business leaders. Their positive, can-do attitude was contagious. It's an exciting time to be working in AI, and I’m looking at the future with optimism!