Building a Data-Driven Culture: Tackling Unstructured Data

The rise of generative AI has renewed interest in data, but simply adopting AI won’t turn a company into a data-driven organisation. A recent survey shows that only 37% of businesses describe themselves as data-driven—down from previous years. Another 92% cite cultural and change-management challenges as primary barriers to becoming data-driven. At the same time, 94% of AI and data leaders say increased interest in AI has prompted greater focus on data management. To build a truly data-driven culture in 2025, companies must address both culture and the technical challenge of unstructured data.

### Why culture matters more than technology

A data-driven culture goes beyond installing analytics tools. It requires employees at all levels to make decisions based on data rather than instinct. Leaders must model data-informed behaviour, encourage experimentation and reward decisions made through evidence. Without these cultural shifts, even the most advanced AI models will struggle to deliver value.

Resistance often stems from fear of change or concern that data will expose mistakes. To overcome this, provide training so employees feel comfortable interpreting data, and communicate how data supports rather than replaces their expertise. Celebrate small wins where data-driven decisions lead to better outcomes, and use those examples to build momentum.

My post

### The unstructured data challenge

One reason organisations struggle to become data-driven is the sheer volume of unstructured data. Text, audio, video, emails and PDFs make up an estimated 97% of a large company’s data. Traditional databases and analytics tools are designed for structured data like numbers in rows and columns. Emerging techniques such as retrieval-augmented generation (RAG) allow AI models to search and summarise unstructured sources, but implementing these systems remains manual and resource-intensive.

### Strategies for success

1. Centralise data management – Establish a single source of truth for data, and implement governance policies to ensure quality and security. This includes defining data standards and appointing stewards.

2. Invest in unstructured data tools – Use technologies like RAG, natural language processing and document indexing to make unstructured data searchable and usable. Start with high-impact domains such as customer support records or research reports.

3. Promote data literacy – Offer training programmes to teach employees how to interpret dashboards, ask the right questions and critique AI outputs. When people understand the data, they’re more likely to trust and use it.

4. Encourage cross-functional collaboration – Break down silos by creating teams that include data scientists, domain experts and business leaders. Collaboration fosters shared ownership of data projects and increases adoption.

### Conclusion

Generative AI has generated excitement about what’s possible with data, but achieving a data-driven culture requires intentional change. With only a third of organisations seeing themselves as data-driven, there is plenty of room for improvement. By addressing cultural resistance, investing in tools that tame unstructured data and promoting data literacy, companies can turn AI interest into a lasting data-driven transformation.