AI implementation dilemma: Buy, build, or partner for success?

The GenAI strategy landscape: Organizations are grappling with the decision to buy, build, or partner as they develop generative AI applications and services, with a mixed approach emerging as the most common strategy.

• According to a KPMG survey, 79% of organizations are either buying/leasing technologies (50%) or employing a combination of building, buying, and partnering (29%).
• Only 12% of organizations are opting to build GenAI solutions entirely in-house, citing cost savings, customization needs, and intellectual property protection as key reasons.
• Todd Lohr, principal and U.S. technology consulting leader at KPMG, emphasizes the importance of finding the right balance between building and buying solutions to gain a competitive edge.

Critical considerations for GenAI implementation: IT departments face several key decisions when developing and deploying GenAI solutions, which can significantly impact the success of their AI strategy.

• The scope of use cases is a primary consideration, including the architecture, processes, and tools required to achieve desired outcomes.
• Potential use cases range from digital assistants retrieving organizational information to AI copilots for code generation and tools for navigating digital twins.
• Choosing the right language model is crucial, with preferences for pre-trained models that offer fine-tuning capabilities and can be run in-house or at the edge for better control over performance, security, and costs.
• Meta’s Llama open-source LLM is cited as a solid choice, already in use by enterprises like Goldman Sachs, AT&T, and Accenture for various applications.

Enhancing AI capabilities: Organizations are looking beyond pre-trained models to create more tailored and effective GenAI solutions.

• Retrieval augmented generation (RAG) is highlighted as a popular technique for generating content that incorporates organization-specific data.
• RAG is particularly useful for summarizing information, retrieving relevant documents, and analyzing data for various business functions.
• This approach helps break down knowledge silos that have long been a challenge for enterprises.

The importance of partnerships: As organizations navigate the complex landscape of GenAI implementation, strategic partnerships are emerging as a crucial factor for success.

• Many organizations lack the internal resources to execute all aspects of their GenAI strategy independently.
• Partnering with experienced providers can offer guidance on bringing AI to data, selecting appropriate infrastructure, and leveraging the growing AI ecosystem.
• Professional services from partners can be instrumental in successfully implementing chosen use cases.

Drivers and expectations: The push for GenAI investment is largely driven by C-suite expectations of tangible business benefits.

• Revenue growth is cited as the top driver for GenAI investment according to executives surveyed by KPMG.
• Organizations are under pressure to develop GenAI applications and services that can provide a competitive edge in the market.

Strategic approach to GenAI: A cohesive strategy that balances building, buying, and partnering is emerging as the most effective approach for organizations implementing GenAI solutions.

• The journey begins with identifying the right use cases that align with organizational goals and can deliver a competitive advantage.
• Careful consideration of data preparation, model selection, and infrastructure choices is critical for success.
• The complexity of GenAI technologies and processes underscores the value of choosing trusted partners to guide organizations through the implementation process.

Looking ahead: As organizations continue to explore and implement GenAI solutions, the landscape is likely to evolve rapidly.

• The success of early adopters and the lessons learned from their experiences will shape future strategies in the field.
• Continuous evaluation and adjustment of GenAI strategies will be necessary as the technology and its applications mature.
• Organizations that can effectively balance innovation with practical implementation are poised to reap the most significant benefits from GenAI technologies.