
Tech Talks Daily 3383: Denodo Technologies on Balancing AI Innovation with Data Integrity
Aug 14, 2025
Charles Southwood, Regional General Manager of Denodo Technologies, brings over 35 years of tech experience, previously as a civil engineer. He discusses the challenge of ensuring data reliability in generative AI to prevent misinformation. Southwood highlights innovative data management strategies like data virtualization, balancing AI innovation with data integrity. He believes 2025 will be a pivotal year for achieving true business agility. The conversation also addresses essential skills for launching a tech career in AI and the integral role of data in enterprise success.
AI Snips
Chapters
Transcript
Episode notes
Trust Is A Precondition For Enterprise AI
- Generative AI must be accurate, trustworthy and reliable before enterprises adopt it widely.
- Regulators like the EU AI Act force organisations to justify AI outputs and their data sources.
Limit Hallucinations With RAG And Oversight
- Use retrieval-augmented generation (RAG) and fine-tuning to reduce hallucinations with domain and real-time data.
- Add human-in-the-loop oversight and contextual guardrails to prevent biased or unethical AI outputs.
Virtualization Delivers Real-Time Data For AI
- Data virtualization unifies real-time data from distributed sources without moving or copying everything.
- That real-time access feeds AI with timely, curated enterprise data and reduces hallucination risk.
