AI as your strategic ally: customer-centric portfolio decisions, powered by Artificial Intelligence

Munich, December 2023

Portfolio decisions have become increasingly complex as the pace of transformation in the automotive industry accelerates. Munich, December 2023

P

ortfolio decisions have become increasingly complex as the pace of transformation in the automotive industry accelerates. Driven by intensifying competition to meet customer demands ever more accurately, we expect this trend to continue for the foreseeable future.

A well-crafted portfolio strategy has never been more essential. Selecting the right products, services, and features that meet customer desires and needs is the key challenge in portfolio decision making, alongside managing risk, allocating resources, and maintaining agility. Could AI be the new standard for how portfolios can be managed more effectively? Think of AI as your most versatile team member, equipped for a variety of tasks. Much like a Swiss Army knife in the field of portfolio strategy. The only question is: how do you best use it?

Our approach offers a practical roadmap for using AI effectively. Not all pieces of the puzzle allowing usage of AI’s full potential are yet in place. However, the picture of how these pieces will fall into place is becoming increasingly clear. By integrating AI into portfolio strategy, a variety of benefits, such as time efficiency, deeper analytical insights, and easier scenario analyses, can be expected. The urgency to integrate AI into strategic planning cannot be emphasized enough. By doing so, not only present work and decision making can be improved, but also the groundwork for significantly greater benefits in the future is laid. Now is a crucial time to consider AI integration; those who act promptly are likely to be better positioned in an increasingly competitive market.

This article is the first in a series of articles exploring the potential of AI for portfolio strategy. As this technology is evolving fast, we look at AI with a dual lens: assessing what is already possible, and subsequently, identifying the next steps for leading automotive players towards (even more) customer-centric portfolios. In this context, this first article will focus primarily on the opportunities for automotive OEMs.

Artificial Intelligence (AI) is a broad field of computer science dedicated to developing systems capable of performing tasks that normally require human intelligence; for example, solving problems, recognizing patterns, and understanding natural language.

Large Language Models (LLMs), such as the latest GPT models, are advanced AI models trained on vast datasets and designed to understand, process, and generate human-like text based on the context provided. LLMs have revolutionized the way we interact with AI systems and bear great potential for applications like customer segmentation and – when done right – can even support strategy work.

 

Unleashing AI’s potential for customer-centric portfolio decision making:

The transformative power of AI for customer and portfolio strategy lays in the ability to inform human decision making. If complemented by a robust framework and accurate data, AI has the potential to enhance customer understanding and increase the speed of analysis in portfolio decision-making. This can lead to greater cost-efficiency compared to traditional data science methods such as advanced analytics or big data. By harnessing vast amounts of structured and unstructured data from different sources, such as customer demographics, purchasing behaviour, vehicle usage, and online interactions, AI can unlock nuanced insights about customer preferences and behaviour enabling a dynamic, real-world understanding of customers. From the discerned patterns, AI can then offer a glimpse into the future, fundamentally informing human decision-making. This granular understanding empowers automotive companies to refine their portfolio strategy with timely insights and tailor their products and services in a way that resonates with customer preferences fostering competitiveness as well as strategic, sustainable growth.

We believe that harnessing the potential of AI for portfolio strategy offers distinct key benefits for OEMs:

Authors
Malte Broxtermann

Partner

Timo Krall

Project Manager

Philipp Brandner

Consultant

Claudius Feldmann

Thesis Student

Benefit 1

1. Dynamic customer understanding & sensing tomorrow’s landscape:

The utilization of AI as a tool for dynamic understanding of customers and market prediction can be a strategic advantage for OEMs. AI can provide timely updates on customer behaviors and needs as required, coupled with detailed segment and persona analysis. This allows automotive companies to gain comprehensive insights into their target audience's preferences and anticipate shifting demands. Through pattern identification AI can anticipate emerging market changes before they fully emerge. By this proactive detection of trends, risks, and opportunities as they emerge, OEMs are better equipped to make more effective decisions in an industry that is more dynamic than ever before.

Benefit 2
2. Aligning strategy with real-world insights:

In today’s rapidly evolving market, aligning strategy with real-world insights is crucial, and AI can play a pivotal role in this process. By comparing strategic concepts with a wealth of internal and external data, AI performs well in scenario recognition. It detects inconsistencies effectively and allows the refinement of strategies to ensure they align more closely with actual market conditions. With actionable insights, these strategies are not only robust on paper but also practical in real-world conditions.
Benefit 3

3. Empowering decision-making through human-AI synergy:
Consider AI as the perfect partner for complex analysis. When humans and AI collaborate, decision-making is supercharged. While humans bring empathy, ethical considerations, and intricate practical understanding to the table, AI ensures precision, speed, and vast data handling. Together, this partnership takes decision-making to the next level, combining the depth of human intuition with the efficiency and analytical capabilities of AI.

After exploring the key benefits and understanding the significant impact of AI, we recognize that the automotive industry is on the threshold of a new era. This leads us to these key questions:

“How do we transition from recognizing these advantages to implementing them in tangible, actionable strategies for portfolio management? How can OEMs leverage AI to not only understand but also anticipate the dynamic demands of the market?”

Operationalizing AI insights: A strategic approach to portfolio management

A clear roadmap is essential to harness the AI benefits and close the gap between AI’s potential and its practical application in portfolio management. This is where we, at Berylls, step in. As agents of operationalization, our goal is to transform abstract concepts into tangible, actionable strategies.

Our approach follows five clear steps for portfolio strategy development:

A) It begins with summarizing existing relevant aspects of corporate, functional, and business unit strategies into key statements, forming a comprehensive understanding of the strategic landscape. Where feasible, key statements are also quantified to enable more meaningful data-based analysis.

B) This is followed by a detailed analysis of the portfolio environment, which is crucial for understanding the stakeholder landscape and market trends. This analysis especially highlights the significant role of understanding customers and competitors to ensure a thorough evaluation of latest leading portfolio strategies and best practices. There is no one-size-fits-all solution, and we pay close attention to what makes a portfolio work (or fail) depending on each specific company and its target customers.

C) We then conduct a comprehensive review of the current internal portfolio in question, analyzing its composition, performance, and enablers.

D) These initial steps lay the foundation for the subsequent phases, where we identify and evaluate potential improvements to the portfolio, focusing on portfolio gaps, key levers, and scenarios for enhancement.

E) Finally, the analysis concludes with defining or updating the portfolio strategy, the setting of specific objectives, and the development of a detailed implementation roadmap.

This approach is optimized for AI integration with the first three steps (A-C) providing the relevant input for AI-supported decision making and the last two steps (D-E) involving AI to analyze all available information effectively.

Figure 1: Berylls framework for portfolio strategy

Source: Berylls Strategy Advisors

The integration of AI into portfolio strategy development represents a significant improvement. Already in the initial stages where qualitative and quantitative data is made accessible for AI, it can play a key role in improving the speed, cost efficiency and quality of strategic decision making. In later steps, AI supports in identifying data patterns, pinpointing missing data points, suggesting portfolio improvements, and preparing meetings by decision makers effectively. Moreover, this integration of AI enables a more dynamic, agile, and efficient iteration of the entire process that closely aligns with evolving market needs.

Figure 2: Berylls framework for AI-enabled portfolio strategy

Source: Berylls Strategy Advisors

An example ChatGPT case illustrating the framework can be found by clicking on this box

Output generated by current LLMs should never be used without diligent review and confidential data must not be submitted to ChatGPT. The example ChatGPT output above was created based on dummy data for a hypothetical company. This output was left completely unchanged by Berylls to help provide a realistic example of the type of quality that is feasible with the latest generation of LLMs. Replicating similar results for real applications requires access to the latest generation of LLMs that guarantees full data security. A real-world project covers drastically more inputs, analytical steps, and output than is included in the simplified example case.

An example ChatGPT case illustrating the framework can be found by clicking on this box

Output generated by current LLMs should never be used without diligent review and confidential data must not be submitted to ChatGPT. The example ChatGPT output above was created based on dummy data for a hypothetical company. This output was left completely unchanged by Berylls to help provide a realistic example of the type of quality that is feasible with the latest generation of LLMs. Replicating similar results for real applications requires access to the latest generation of LLMs that guarantees full data security. A real-world project covers drastically more inputs, analytical steps, and output than is included in the simplified example case.

Using AI to take smarter portfolio decisions by putting the customer front-and-center:

AI can demonstrate significant advantages over traditional analyzing methods. AI’s speed and adaptability enable quicker decisions, eliminate human biases, reduce manual research, and efficiently process large datasets for precise market alignment and feedback analysis. Correctly providing relevant input and gradually refining the prompts provided to AI does require expertise and AI currently cannot (and in our assessment: should not) replace human decision making in matters of strategic importance. Latest AI models, nevertheless, can be applied to accelerate analysis and help human experts take more informed and well-prepared decisions. Remarkably, an analysis as outlined in the example case linked above can be replicated in minutes, assuming an AI model comparable to the latest version used by ChatGPT has access to all essential data in a way that ensures data security. While the case linked above is hypothetical, AI indeed can create significant value once it has access to an organization’s real-world data and strategies. Berylls invites you to start working towards unlocking the full potential of AI to achieve a deep, data-driven understanding of customer behaviors, preferences, and expectations.

Engage several powerful AI flywheels and lay the foundation for success by embracing the following enablers:
1. Redefine data for AI

Quality data is the backbone of meaningful AI insights. Ensure your data is well-structured, high-quality, meaningful, and thus prepared to enable accurate results.

2. Forge AI-ready strategies

Condense your core business strategies into written key statements to make them AI-readable. This allows AI to analyze data and generate insights with your strategy in mind.

3. Human-AI decision synergy

Use AI to support your decision-making process, ensuring faster responses to dynamic market demands. Combining the strengths of human intuition with AI’s analytical capabilities creates a powerful duo, maximizing both precision and efficiency.

4. Tailor your AI approach

Different needs require different solutions. Whether it's leveraging AI for specific tasks or a more comprehensive AI-driven strategy, aligning the technology with individual organizational goals and capabilities is crucial.

Engage several powerful AI flywheels and lay the foundation for success by embracing the following enablers:

1. Redefine Data for AI

Quality data is the backbone of meaningful AI insights. Ensure your data is well-structured, high-quality, meaningful, and thus prepared to enable accurate results.

2. Forge AI-ready strategies

Condense your core business strategies into written key statements to make them AI-readable. This allows AI to analyze data and generate insights with your strategy in mind.

3. Human-AI decision synergy

Use AI to support your decision-making process, ensuring faster responses to dynamic market demands. Combining the strengths of human intuition with AI’s analytical capabilities creates a powerful duo, maximizing both precision and efficiency.

4. Tailor your AI approach

Different needs require different solutions. Whether it's leveraging AI for specific tasks or a more comprehensive AI-driven strategy, aligning the technology with individual organizational goals and capabilities is crucial.

We expect a broad spectrum of relevant use cases for applying AI to portfolio strategy. The role of AI can range from fundamental applications in stakeholder analysis to more complex scenarios in business case preparation and continuous improvement. Each use case demonstrates a specific aspect of AI’s contribution to portfolio strategy and has unique requirements for successful implementation. Different use cases can be rolled out sequentially, gradually creating more value with AI over time.

Four types of portfolio strategy use cases for AI

Low Complexity
Low Output

High Complexity
High Output

Use case I: stakeholder analysis

LLMs enable detailed profiling and analysis of customers, competitors, and other key stakeholders. Profiles can also cover the portfolios and best practices by key competitors. LLMs can provide frequent, cost-effective updates based on AI-assisted research, competitor monitoring, and customer data analysis. This speed and breadth of analysis helps improving the understanding of market expectations, behaviors, and dynamics.

Main requirements:

- Introduce LLM integration: Provide access to a Large Language Model (LLM) like ChatGPT for analyzing anonymized stakeholder data.

- Establish customer data platform: Develop a platform to manage and analyze customer data across relevant data sources, ensuring it is structured and AI-ready.

- Ensure data updates: Implement a system for continuous updating of stakeholder data to keep the analysis relevant.

Use case II: Identification of decision option

AI aids in broadening decision options for portfolio adjustments. It helps in generating new ideas for portfolio changes by using pattern recognition in the available data to identify potential opportunities and optimize strategic choices. Specifically, AI can help identify untapped potential by highlighting discrepancies of the current portfolio with available customer insights, market trends and best practices by comparable competitors.

Main requirements:

- Ensure available strategies are AI-readable: Turn the relevant aspects of current strategies into AI-readable (quantified) key statements.

- Integrate results of stakeholder analysis: Grant AI access to all relevant analysis of the impact and interests of various key stakeholders, especially by customers and competitors. This should include a thorough review of competitor portfolio best practices and strategies with clear assessments of the company-specifics that make each competitor succeed (or fail).

- Describe portfolio for AI Analysis: Submit a comprehensive documentation of the current product and service portfolio, the portfolio performance, and any relevant information about company-specific portfolio enablers. The information about enablers should include the main resources allocated across the portfolio as well as key competitive strengths (or weaknesses). Examples for relevant competitive attributes would be information about the sales and manufacturing model used, describing (e.g.) the resulting supported production speed, variety, flexibility, and cost-efficiency

Use case III: Assessment of decision options

Latest generation LLMs can assist in qualitative comparison of decision alternatives, for example by suggesting (dis)advantages for each available decision. For quantitative decisions, they can also support in preparing business cases. At the current technological level of LLMs, business cases are mainly manually crafted, with LLMs offering suggestions that require in-depth expert review for accuracy. Over time and as LLMs evolve, their suggestions become more reliable, allowing them to draft business cases based on company data. This evolution will gradually shift the human role towards systematically refining these AI-generated suggestions, leading to more complex and accurate business cases that account for scenarios involving competitor responses, thus enhancing strategic decision-making

Main requirements:

- Integrate results of stakeholder and decision option analysis: Inform the AI about the final set of decision alternatives and the related scenario assumptions. This will be the starting point for the business case. In addition, provide input to the AI with any relevant findings from the stakeholder analysis such as identified customer preferences. This information, for example, can help the AI to suggest meaningful parameters and assumptions for the business case.

- Provide access to operational and financial data: Both internal and market-related operational and financial data must be accessible for a comprehensive AI analysis. Where available, validated forecast data will improve the accuracy of the analysis.

- Involve industry experts: From start to finish, AI prompts should be engineered by industry experts to ensure relevant and accurate AI outputs.

Use case IV: Continuous improvement

AI supports the ongoing refinement of a portfolio strategy. It quickly identifies opportunities for action based on current data and ensures that the portfolio remains aligned with market trends and customer preferences, leading to continuous improvement and strategic agility.

Main requirements:

- Establish regular data updates: Ensure AI analyses are based on most current information.

- Implement feedback mechanisms: Integrate customer feedback and performance data into ongoing AI analysis.

- Develop a dynamic strategy adaption process: Allow regular strategy refinements based on AI insights.

Leveraging AI for your company… but different

The potential of AI to drastically change the way strategy work is done grows rapidly. Berylls stands ready to help you transform successfully and in a way that is uniquely centered around your customers, company, situation, and aspirations. Stepping into the AI realm can be complex; however, having a dedicated partner ensures that you harness its full potential efficiently and strategically. Berylls provides a tailored way to make AI work best for your needs, from initial assessment to final implementation. Backed by a team of experts and collaboration partners with custom software solutions at the forefront of AI advancements. Embrace the AI-driven future of automotive customer strategy and gain the AI advantage with us. The automotive industry faces a time of accelerating change, and those who can sense and adapt to these changes first as AI matures will have the upper hand.

Malte Broxtermann

Malte Broxtermann (1986) joined the Berylls team in 2014. After extensive experience as emergency medical technician, he has been working in consulting since 2012. He helps customer to leverage digital strategies & products across the entire automotive value chain. He is an expert in deploying machine learning-powered applications. As Partner at Berylls’ own unit for digital solutions, Berylls Digital Ventures, he focuses on scaling start-ups as part of our venturing practice.
Studied economics and international business at Maastricht University (Netherlands) and Queen’s University (Canada).