Strategic Clarity: The Foundation for AI Success
The roundtable began with a critical examination of an issue. Common pitfall: Lack of a defined AI strategy. Despite widespread interest in AI, many organizations lack a clear vision for AI applications, often torn between the desire to use a large number of generic models and the need for unique solutions tailored to specific business problems. This strategic vacuum often results in fragmented efforts, unrealized potential, and failure to translate excitement into tangible business value.
Establishing clear governance emerged as a top concern. The discussion emphasized that the definition of ownership, policies, and operating models for AI implementation often precedes actual technology deployment. Several participants mentioned the creation of a dedicated AI council, typically comprised of IT, compliance, and automation experts, tasked with laying the foundation for a coherent strategy and responsible use of AI. These councils are important for addressing complex issues such as data ownership and ethical AI deployment.
“Looking back, many organizations failed to define a clear AI strategy and tried to use it aimlessly or in large quantities, or tried to solve everything with a single tool. The fundamental question, ‘What do we want to achieve with AI?’ was often overlooked.” – Eszter Lukács, GBS Advisory and Finance Transformation Director, Deloitte Hungary.
Navigating your practice: approaches and tangible benefits.
Exploration of dialogue Diverse implementation approaches, from top-down to CEO-level authority Found in financial institutions and technology companies, Toward a more organic bottom-up approach We focus on employee empowerment. For example, some companies are redefining internal operations with AI solutions, while others are leveraging internal “AI accelerators” and hackathons to drive adoption and generate innovative ideas from scratch. These initiatives aim to demystify AI, reduce employee anxiety, and foster a culture of experimentation.
Success stories shared included optimizing pricing and promotions, and systems to detect fraudulent product returns. Other examples highlight automated processing of incoming emails and documents, AI-driven simulations for process optimization, and intelligent inventory management. These cases are AI’s potential to increase operational efficiency and improve customer experience; For priority projects, we typically deliver a measurable return on investment within a payback period of 18 to 24 months.
but, Participants also acknowledged pitfalls. The key insight is the dangers of framing AI efforts solely around cost reduction; You may be missing opportunities to generate revenue, strengthen capacity, or address existing gaps. What was emphasized in the discussion was The importance of distinguishing between large-scale language models (LLMs)is better at understanding language, but can “hallucinate” information and is more specialized. Small Language Model (SLM) Or, use traditional machine learning for applications that require high accuracy, such as financial forecasting or contract analysis.
“AI offers immense capabilities, but we must understand its probabilistic nature and the need for human expertise in validation and content curation. Misaligned expectations and poor user understanding can lead to significant setbacks, highlighting the critical role of seasoned subject matter experts.” – added Zoltán Páll, Technology & Transformation AI Manager, Deloitte Hungary.
Talent development and AI-enabled ecosystem
A recurring theme was the urgent need for talent development and upskilling. Hungary boasts a strong base in analytics, data science, and engineering talent, but the rapid pace of AI evolution is outpacing current education and in-house training efforts. Organizations are increasing investment in internal programsWe focus not only on technical skills such as Python, but also on important soft skills such as resilience, problem-solving, and proactiveness, which are critical to navigating an AI-driven environment.
Access to R&D grants and government support was discussed as a factor that is often underutilized. In Hungary, AI-related developments, particularly initiatives that involve experimentation, uncertainty, or new combinations of existing technologies, are often eligible for R&D tax incentives and cash grants, where organizations can potentially recover 20-50% of eligible costs. Participants noted that these mechanisms can significantly improve the business case and enable projects where securing internal funding may be difficult.
Human factors such as potential resistance and the need for “inclusive AI” were also critically discussed. Misplaced trust in AI, or conversely, a complete refusal to engage with it, can ruin even the best-laid projects. The roundtable emphasized: AI is a tool that expands human capabilities, It is not completely replaced. The focus should shift from “AI will take our jobs” to “our colleagues will excel using AI.” Validation of AI output and continuous improvement of the system requires strong involvement of SMEs (Subject Matter Experts).
Additionally, a solid data strategy and robust data governance are fundamental. Modern AI, especially LLM, can handle unstructured data, but standardized, high-quality data greatly increases accuracy and reliability. Integrating an AI strategy with existing data management frameworks, often under separate leadership, presents an ongoing challenge that requires increased cross-functional collaboration to ensure that data is not just “AI-enabled” but truly fits enterprise-wide objectives.
Looking to the future: adaptability and continuous learning
The consensus among GBS leaders was that working towards optimal AI integration is dynamic and requires continuous learning and adaptability. The speed of AI development requires an infrastructure that can evolve to accommodate future advances, rather than just meeting current needs. Organizations must foster a mindset that embraces change, encourages experimentation, and sees AI as a powerful enabler of strategic growth and innovation. This far exceeds any immediate cost savings potential. This collaborative effort, combining strategic vision with agile execution and human-centered development, will define success in the evolving AI landscape.
