How Open Context Layers Help Enterprises Build, Govern, & Scale Agentic AI - with Prukalpa Sankar of Atlan (25 min)
ai-driven-innovation-economy ai-global-economic-shifts ai-governance-laws ai-in-workforce-disruption ai-investment-trends ai-monetization-strategies
- Release date: 2026-01-08
- Listen on Spotify: Open episode
- Episode description:
Today's guest is Prukalpa Sankar, Co-Founder & CEO at Atlan. Atlan is a metadata platform delivering data and AI governance. Combining enterprise data with AI infrastructure with business context and security, Atlan allows humans and AI agents to find, trust, and govern data across the entire enterprise. Prukalpa joins Emerj Editorial Director Matthew DeMello to unpack why so many enterprise AI pilots still fail in production — and why solving for context rather than just data volume or model complexity is becoming the decisive factor. The conversation also highlights how leading enterprises are closing this context gap and what "context readiness" means for any organisation aiming to operationalise AI in 2026. Want to share your AI adoption story with executive peers? Click emerj.com/expert2 for more information and to be a potential future guest on the 'AI in Business' podcast! This episode is sponsored by Atlan.
Summary
- 🚨 AI Pilot Failure Epidemic: 75-95% of enterprise AI projects fail due to context gaps in data, business meaning, and governance, mirroring past data analytics flops but at greater scale.
- 🔗 Dynamic Context Layer: A living layer integrating data, business, domain, and process contexts with feedback loops surpasses static semantics, enabling precise AI like defining ‘top 10 customers’ dynamically.
- 🤝 Business-Tech Fusion: New operating models demand business as ‘context engineers’ for domain expertise, creating digital twins unlike solo IT projects.
- 💼 Use Case-First Approach: Start with ROI-driving pilots, build reusable context foundations, and scale to avoid multi-year delays while prioritizing context readiness.
- 📈 Proven Enterprise Wins: Workday achieves 5x accuracy, VMO2 onboards thousands, Mastercard goes AI-first—unlocking adoption, usability, and existential edge in innovation races.
Insights
Why do 75-95% of enterprise AI pilots fail to reach production despite massive investments?
Time: 4:36 – 7:24
Category: AI Investment Trends, AI Governance & LawsAnswer: Echoing past data analytics failures where 75% of $250B projects flopped, recent surveys show 75-95% of AI initiatives stall due to a ‘context gap’ in data discoverability, business meaning, and governance, blocking reliable AI reasoning. (Start at 4:36)
What exactly is the ‘context gap’ derailing enterprise AI?
Time: 5:57 – 7:24
Category: AI Governance & LawsAnswer: It spans three layers: finding relevant data (e.g., CIOs with 1,000 projects unsure where data is), infusing business-specific meanings (e.g., ‘TAM’ varying by company), and enforcing governance (e.g., restricting HR chatbots from payroll data). (Start at 5:57)
How can a dynamic ‘context layer’ outperform static semantic layers for AI?
Time: 7:51 – 10:27
Category: AI-Driven Innovation Economy, AI in Workforce DisruptionAnswer: Unlike rigid semantic layers or ontologies, a dynamic context layer weaves data, business meaning, domain, and process contexts with real-time feedback loops from user interactions, enabling AI to clarify ambiguities like ‘top 10 customers’ by revenue vs. NPS. (Start at 7:51)
Why must business teams become ‘context engineers’ alongside data pros?
Time: 10:47 – 11:49
Category: AI in Workforce DisruptionAnswer: AI demands domain expertise held by business users, shifting operating models from IT-led to collaborative, creating a ‘living digital twin’ of the organization impossible without co-ownership. (Start at 10:47)
What existential risks do enterprises face without context readiness?
Time: 12:06 – 14:22
Category: AI & Global Economic ShiftsAnswer: Failing AI projects amid hype exacerbate threats as S&P 500 tenure drops from 75 to 15 years; leaders risk obsolescence while adopters thrive in rapid innovation cycles. (Start at 12:06)
How are pioneers like Workday and Mastercard unlocking AI ROI via context?
Time: 15:02 – 16:35
Category: AI Monetization StrategiesAnswer: Workday boosted AI accuracy 5x by making governed BI consumable; VMO2 onboarded 6,000 employees with 1M+ views; Mastercard builds agentic governance for one-third AI-first products, driving usability and adoption. (Start at 15:02)
What blueprint ensures sustained AI value through 2026?
Time: 16:56 – 18:48
Category: AI-Driven Innovation Economy, AI Governance & LawsAnswer: Prioritize ‘context readiness’ over AI readiness by starting with high-ROI use cases, embedding reusable context engineering with business involvement, avoiding 3-4 year governance pitfalls. (Start at 16:56)
How do you validate and iterate a context layer for production AI?
Time: 19:38 – 21:53
Category: AI-Driven Innovation EconomyAnswer: Bootstrap evals from BI/Slack queries and golden SQL, use human-in-loop for 10x accuracy gains in a month, then always-on enrichment as business context evolves, ditching ‘golden’ static approaches. (Start at 19:38)