Few technologies have debuted to as much consumer and media fanfare as Generative AI (GenAI). This rapid expansion has fostered an increasing conviction among enterprises that GenAI holds the promise to greatly enhance, if not replace, the most complex and unstructured paths of value generation. With more than a year of hype around this new technology (after the November 2022 launch of ChatGPT), we have seen an explosion in terms of user-awareness, experimentation, innovation along with a few pioneers spearheading GenAI adoption. What have we learned and what do we anticipate on the GenAI horizon in 2024? 

Reflecting on 2023 – Experimentation before production 

In 2023, GenAI investment surged, exceeding EUR 12 billion in venture capital investments in Q1 of 2023 (almost tripling that of Q1 2022). The game-changing quality of GenAI is evident, as companies that integrate it experience immediate benefits at both micro and macro levels. GenAI assistants applied on a company’s existing knowledge base enhance the contributions of new or inexperienced workers and propel experienced workers to quickly reach new productivity levels. Considering that the GenAI market is foreseen to double every other year for the next decade, organizations are investigating the best route to secure their share. 

Yet, the past year revealed a stark reality — mass activity, interest, and investments mainly delivered proofs of concept (PoCs) and demos, due to disjointed approaches and other market challenges. A Deloitte CEO survey1 in October 2023 shows only 17% operationalized a GenAI use case, while 34% are in limited production, and 49% remain in evaluation and experimentation phases. 

Many companies, despite untapped potential, are in the experimentation phase, providing insights for a potential shift toward GenAI industrialization.  

Overcoming the challenges – Unlock benefits of GenAI in production 

To move from GenAI experimentation to operationalizing these solutions, companies should understand the main challenges. The economic downturn in 2023 introduced investment tradeoffs to many businesses, hindering their ability to acquire foundational technologies and upskill their workforce. In such a complex global environment, company leadership must identify strategies for long-term resilience. 

Organization can use key lessons from the past year to shape their GenAI vision, foster cross-organizational AI awareness, and make smart GenAI investment decisions: 

    - Fast-changing environment: Navigating the ever-changing tech landscape and taking advantage of the latest technology requires strong guardrails for safe and secure solutions. Building enterprise platform capabilities offers flexibility to seamlessly switch between models, whether on-premises or through cloud-based systems. 

    -Lack of transparency in contractual documentation: Ambiguities in Large Language Models (LLM) terms of use could hinder GenAI adoption. Like any software, GenAI providers should clarify how their LLMs are built to ensure alignment on data usage, legal compliance, and security principles. The AI Act is expected to address this issue by mandating AI model assessments to prevent unauthorized activities. 

    - Ethical challenges and user trust: GenAI models face ethical dilemmas, such as hallucinations and accountability issues, bias, and IP rights. Evaluating sources for LLM training ensures data trustworthiness. Consumer demands for transparency push the need for additional legislative and societal structures to address ethical concerns.  

    - Dependence on high quality prompts: User adoption challenges persist, requiring end-user comprehension of GenAI, as the effectiveness of ChatGPT-like systems depends on the quality of user prompts. Educating the organization to effectively prompt GenAI is vital in overcoming this challenge.  

Realizing the future – Our outlook on what is to come 

GenAI is a strategic focus area on C-level agendas across industries, with peak productivity expected between 2025 and 2027, encouraging organizations to prepare and invest now for increased efficiency gains.  

As software vendors integrate additional AI and GenAI solutions into standard offerings (e.g., Microsoft CoPilot integrated into the wider M365 offering), GenAI use will shift into the mainstream and foster broader user literacy. 

In Luxembourg, leaders in GenAI exploration and adoption are primarily Financial Services Institutions (FSI) and EU Institutions. Previously confined to low creativity and context variability, automation now benefits businesses heavily relying on unstructured documents, both paper-based and digital. GenAI’s proficiency in processing and exploring unstructured content hint at broader applications in other sectors in locally and globally in the coming year.  

Looking ahead to 2024, developments in GenAI will not only fortify ongoing developments, bring concrete operational efficiencies and revenue gains for early adopters, but also unveil new groundbreaking use cases. 

Conclusion 

GenAI’s potential is unfolding, requiring widespread AI literacy across all organizations. To unlock its value, empowering teams with confidence and freedom to experiment is crucial, supported by substantial investment. A persistent focus on privacy, security, trust and explainability is vital, ensuring compliance with upcoming AI regulations. Building a strong foundation in digital and AI capabilities is also critical to emphasizing technology infrastructure that gives the flexibility and computing power needed to properly empower AI, data management that feeds AI models, and change management to upskill staff and guide restructuring as needed to adapt to new ways of working.  

Picture: Nicolas Griedlich, Partner, Artificial Intelligence & Data, Deloitte Luxembourg  & Anke Joubert, Senior Manager, Artificial Intelligence & Data, Deloitte Luxembourg