Summary
Panelists described why enterprise AI needs unified, contextual data from fragmented marketing, healthcare, and customer systems before agents can be useful. They emphasized privacy foundations, per-patient encryption, brand-safety simulations, parallel moderation, and forward-deployed teams as practical ways to scale real-time AI without sacrificing compliance or trust.
Key Topics
5 key topics from Ioana Hreninciuc, Ardy Arianpour, David Bromberg and Ron Miller at HumanX Amsterdam 2026. Thicker branches were mentioned more often in the talk.
Lessons Learned in Unifying Fragmented Data & Scaling Safely
Click a topic to open it
Key Takeaways
- Build privacy, security, and encryption foundations before deploying real-time AI.
- Map customer-facing systems into a shared ontology before expecting intelligent agents.
- Use simulations and brand-guideline agents to validate scaled personalization.
- Run safety checks in parallel with inference to avoid slowing legitimate requests.
- Start pilots with narrow ROI-heavy workflows, then expand through change management.
“if you don't start with that foundation, it's really hard to go back and build it.”
Share
Up Next
S
WatchNext in agenda
The Next Frontier in Enterprise AI: Video's Compounding Value
SSoyoung Lee
video aiai agents
SJD
WatchAlso on enterprise ai
Deploying at Scale: The True Cost of AI
Susan Jones, James Waters, Dasha Afanasieva
ai adoptionenterprise ai
KT
WatchAlso on ai safety
Building at the AI Frontier
Kareem Ayoub, Tom Mackenzie
agentic aiai safety