Was machen die Unternehmen richtig, die den größten Nutzen aus der Künstlichen Intelligenz ziehen?

When investing in Artificial Intelligence (AI), one must not only think about the possibilities of in - house development, purchase, integration, or cooperation but also about how to build organizational capabilities and sensibly use these four strategies. At the same time, one should establish a decision - making framework to ensure that each strategy achieves the maximum strategic value. Companies that master this multi - dimensional approach can not only optimize their AI investments but also create a sustainable competitive advantage and reasonably justify every investment decision.

Why do some companies invest millions in AI while their competitors can achieve better results with far less money?

This question precisely reflects the strategic dilemma of some companies: What is the best way to invest in AI? When should a company build its own AI capabilities, and when should it buy external solutions? The answer is not simply a binary decision.

Today, many companies have moved beyond this simple binary decision between in - house development and purchase and instead chosen more refined strategies. According to data from the International Data Corporation (IDC), only 13% of IT leaders plan to build AI models from scratch, while 53% want to first use pre - trained models and then optimize them with corporate data. This shift towards strategic implementation - especially the increasing importance of strategic partnerships - shows that success in the field of AI does not depend on the amount of investment but on whether one invests wisely between in - house development, purchase, integration, and cooperation.

As a consultant for AI transformation, I have personally witnessed how companies make these decisions and at the same time adapt their employee teams to get used to the new technology. With the increasing adoption of AI by companies of all sizes, the urgency of these decisions is becoming more and more obvious. According to the "2024 Labor Market Situation Report" by TriNet, 88% of small and medium - sized enterprises and 71% of employees already use AI in the workplace. The companies that achieve the greatest benefits have developed systematic methods that go beyond mere cost considerations.

Strategic Decision - Making Framework

The most successful companies evaluate each AI capability based on a systematic framework. The first question should not be "In - house development or purchase?" but "Can this capability create unique value for our customers in a way that is difficult for competitors to replicate?"

This evaluation of strategic value requires considering three key dimensions: Potential for competitive differentiation, organizational readiness, and long - term strategic alignment. Companies that perform well in this evaluation generally have better results than those that mainly make decisions based on upfront costs or technical preferences.

When to Choose In - house Development

Companies choose in - house development when a capability represents a central competitive advantage, the data and industry knowledge available to them form a unique entry barrier, the long - term cost advantages justify a high upfront investment, or the protection of intellectual property is crucial for their business model.

In - house development requires comprehensive planning and systematic implementation. First, a detailed capability planning must be carried out - all AI capabilities from customer - centric applications to operating systems must be defined. For each capability that needs to be developed individually, a comprehensive feasibility study must be conducted, taking into account technical requirements, personnel needs, and infrastructure needs.

A special interdisciplinary team should be established, integrating existing internal personnel and recruiting strategically. These teams should include not only technicians but also industry experts with a solid business background to ensure that the AI solutions can handle practical operational challenges. A development period of 12 to 24 months should be planned and released iteratively to continuously receive feedback and improve.

In addition, a strong development infrastructure must be built, including scalable computing resources, comprehensive data pipelines, and Machine Learning Operations (MLOps) capabilities that support the entire machine - learning lifecycle. This investment in infrastructure usually accounts for 30% to 40% of the total project cost but is crucial for long - term success.

Finally, the success indicators must be defined, which should not be limited to technical performance but also include business impact indicators such as development speed, system reliability, user acceptance, and measurable competitive advantages. A regular review cycle should be established to evaluate progress based on these indicators and adjust the strategy if necessary.

Risk management is particularly important for the in - house development strategy. Contingency plans must be created for problems in employee retention, technological development, and changing business requirements. At the same time, it must be considered how the individually developed system can be integrated with future technological acquisitions to ensure that the architecture can adapt to organizational requirements.

JPMorgan Chase is a prime example of this approach. In 2024, the bank invested $1.7 billion in technology, a large part of which went into self - developed AI systems. Their individually developed AI platform for fraud detection can analyze the transaction patterns of certain customer groups and provide a customized risk assessment that no standard solution can achieve. This investment has reduced the rejection rate in account verification by 15% to 20% and at the same time significantly reduced the number of false alarms, impressively demonstrating how strategic in - house development can create measurable competitive advantages.

"Follow Your Passion"

Companies choose to purchase external solutions when market entry speed is crucial, the providers have excellent expertise, or the internal development costs exceed the long - term value creation. The purchase strategy is particularly effective for standard functions because in these cases, the competitive advantage comes more from excellent implementation than from the differentiation of the underlying technology.

Successful procurement requires a complex provider evaluation process that considers not only the current capabilities but also the alignment with the future roadmap and the integration flexibility. Comprehensive evaluation criteria should be established, including technical performance, security compliance, scalability potential, and provider stability.

A comprehensive provider evaluation should be carried out, including requesting references from similar companies, piloting important functions, and conducting a detailed analysis of the total cost of ownership (i.e., license fees, implementation costs, training costs, and ongoing support costs). Particular attention should be paid to the integration requirements to ensure that the acquired solutions can work seamlessly with existing systems and data streams.

Contracts should be negotiated to provide flexibility for changing requirements and to avoid dependence on a single provider. They should include clauses on data portability, access to application programming interfaces (APIs), and performance guarantees. One should also consider a multi - provider strategy to avoid over - dependence on a single provider and create a competitive situation for the company.

Remember, develop a comprehensive change management process for the acquired solutions. Even off - the - shelf software requires significant adaptations from the company, including user training, process changes, and cultural adaptation. An implementation schedule of 6 to 12 months should be planned, including comprehensive testing, user training, and phased roll - out.

Finally, the performance of the providers should be continuously monitored through the agreed - upon service - level agreements and regular business reviews. One should keep an eye on alternative providers and be ready to switch if the provider's performance declines or the strategic alignment is lost.

Salesforce's acquisition strategy is a typical example. They have acquired specialized AI companies such as Einstein Analytics and integrated these capabilities into their core platform. Instead of developing every AI function internally, they have accelerated the development of their AI capabilities through strategic acquisitions of established technologies and teams while concentrating internal development on core innovations in the field of Customer Relationship Management (CRM) to differentiate the platform.

When to Choose Integration

The hybrid strategy - i.e., developing some components in - house and purchasing others - is most suitable when some components need to be customized while others can be standardized, or when a company wants to control the core algorithms and at the same time use external infrastructure. With the search for a balance between speed, cost, and competitive differentiation, the integration strategy is becoming increasingly popular.

Successful integration requires precise architecture planning to ensure the seamless integration of internal and external components. Modular systems with clearly defined interfaces should be designed so that different components can be developed, updated, or replaced independently.

In addition, strong application programming interfaces (APIs) and data exchange protocols must be developed to ensure smooth communication between internal systems and external solutions. Particular attention should be paid to data security and regulatory compliance, especially when integrating cloud - based external services with internal systems that contain sensitive information.

A clear governance architecture should be established to define the ownership rights and responsibilities for the different system components. An interdisciplinary team should be established to monitor the integration, review the performance, and drive the strategic development of the integrated solution forward.

Given the continuous development of internal and external components, continuous optimization must be planned. The integrated solution requires constant attention to ensure that the update of one component does not disrupt the other components and that the entire system remains consistent and powerful.

Capital One has effectively implemented this approach. They have built their own Machine - Learning platform for the core function of credit decision - making while purchasing a pre - built AI solution for the automation of customer service. This hybrid strategy has significantly improved the processing efficiency and customer satisfaction, showing how strategic integration can maximize the return on AI investments.

When to Choose Cooperation

Strategic partnerships represent a fourth way, which is different from traditional provider relationships and offers a comprehensive solution that combines technology, expertise, and continuous service delivery. This method is most suitable when certain capabilities are important but not differentiating, specialized providers have excellent expertise and technology, or a company needs a flexible service model to adapt to changing requirements.

Strategic partnerships require a careful evaluation of partners based on multiple criteria, including technical capabilities, industry knowledge, service quality, and cultural alignment. One should look for partners who offer end - to - end solutions, not just software licenses, including implementation support, continuous optimization, and strategic advice.

A detailed partnership agreement should be negotiated, which in addition to traditional service - level agreements also includes commitments on strategic alignment, innovation cooperation, and mutual performance motivation. This relationship should be more like an extension of the internal team than an external provider relationship.

An integration strategy should be developed to ensure that the partners' solutions can work seamlessly with internal systems, taking into account security and compliance requirements. This usually requires setting up special communication channels, sharing performance dashboards, and conducting regular strategic reviews.

Finally, a governance structure should be established to ensure that the partnership adapts to organizational requirements. Regular strategic reviews should evaluate not only the operational performance but also the strategic alignment, innovation cooperation, and long - term value creation.

Domino's Pizza has entered into a partnership with Microsoft Azure to develop an AI - based platform for optimizing orders and deliveries. This strategic partnership is an excellent example. Domino's did not develop these capabilities internally and did not just choose to buy software licenses. Instead, they collaborated with Microsoft to develop an AI solution that optimizes delivery routes, predicts customer preferences, and automates order processing. This cooperation has enabled Domino's to use Microsoft's advanced AI capabilities while bringing in its own expertise in pizza delivery logistics. In this way, Domino's has increased the accuracy of its AI prediction of order readiness from 75% to 95% by using a load - time model that takes into account workforce variables and order complexity. Microsoft has improved its AI services for other retail customers through the insights from the practical application, while Domino's did not have to build its corporate AI capabilities from scratch with large internal investments.

Strategic Priorities

Companies that get the most out of AI have moved beyond the simple debate between in - house development and purchase. They have a decision - making framework...