Reflections 15.10.2024
Navigating the AI revolution – Choices to be made
In today’s data-driven world, success is not about adopting the latest technology for its own sake, but about harnessing data to improve business outcomes. Whether it’s optimising operations, forecasting demand, or the reliability of fault reports, the key to thriving in this landscape lies in making strategic choices that align data initiatives with real business goals. This requires clear objectives, smart investments, and a deep understanding of how data can enhance decision-making, increase efficiency, and drive new revenue streams.
The market and technology landscape – The global AI race is on
The global AI landscape is dominated by key players, particularly the United States and China, whose investments and advancements far exceed those of Europe. America holds an impressive 80.3% share of the worldwide platform economy, with a total value of $11 321 billion, driven by tech giants like Microsoft, Apple, Meta, Amazon, and Alphabet. Asia-Pacific accounts for a total platform value of $2 225 billion, with key platforms emerging from China, Korea, India, and Japan (Tencent and Alibaba). Africa, though still in the early stages of its digital development, its platform value represents a 1.7% share of the global platform economy.
Europe, however, struggles with a 2.2% global platform share, facing internal challenges such as fragmented markets and regulatory hurdles. To compete with powerhouses, Europe needs greater collaboration at the EU level to pool resources and boost AI through cutting-edge platform solutions. (Source: Ecodynamics, 2023).
Building a successful machine learning project – harness the data
When embarking on a machine learning (ML) project, companies must recognize that AI is not a one-size-fits-all solution. The process involves many stages, including data collection, ensuring data quality, cleaning, building reliable data flows and pipelines, structured and unstructured data storage, feature engineering, analytics, model development, training, validation, deployment, and continuous optimisation. In fact, the majority of time in AI projects is often spent on ensuring high-quality data – as poor data can undermine even the most advanced models. Each step must be approached with thoughtful execution and a focus on aligning with business objectives, ensuring that all actions are validated by the needs of the business. This requires a cross-functional team that spans software engineers, data engineers, data scientists, and business domain experts.
Choices to be made in the AI journey
According to several recent studies, “AI first” or “data first” approaches often result in 80% of projects failing. This is typically because the business target was defined too vaguely or there wasn’t the right data for the business case. The companies that have succeeded in utilizing AI and gaining real benefits from their data have tapped into all of the following:
- Invest in time and money: Organizations need to allocate resources to build a strong AI foundation, from data infrastructure to AI talent.
- Start with business benefits: AI initiatives should always be linked to clear business goals, ensuring that AI delivers tangible value.
- Build a cross-functional team: Collaboration between technical teams, business units, and leadership is critical for AI success.
- Focus on governance and regulation: The upcoming EU AI Act and EU Data Act will have a significant impact on AI projects, and businesses must ensure they comply with new regulatory requirements.
Preparing for the EU Data Act and EU AI Act
The regulatory environment around data and AI is changing rapidly, with the EU Data Act and EU AI Act setting the stage for new rules on data usage, sharing, and AI system development. These regulations aim to protect user rights while trying not to hinder innovation. For organisations, this means making sure their AI systems are compliant, particularly those that involve high-risk AI use cases.
The EU Data Act, for example, requires companies to share raw data with users securely and free of charge, with exceptions for data that has been heavily processed or derived. The transition period for this regulation is already underway, meaning businesses must begin preparing now. Likewise, the EU AI Act will impose traceability obligations on some companies, meaning processes to trace the behaviour of models will be implemented in many organisations.
The way forward
To fully leverage data, companies must make strategic decisions today – understanding their data, choosing the right technologies, and building capable teams. Success lies in aligning data with business goals, continuously improving, and integrating data into operations. Whether through predictive maintenance, fault analysis, increasing the automation in customer service, real-time energy consumption visualisation, demand forecasting, or optimising any resource, businesses that invest in data infrastructure and strategy now will lead in efficiency, productivity, and new revenue streams tomorrow.
Hanna Hagström was the keynote speaker at Aidon Energy Vision seminar 11t September 2024 in Helsinki.