Agents, Trust, and Real ROI with SAS

Over the last two weeks, I sat down with two leaders from SAS to cut through noise and focus on impact. Marinela Profi unpacked agentic AI for real enterprise use. Manisha Khanna walked through the SAS–IDC Data and AI Pulse and why trust drives outcomes. Together, these conversations map a path from pilots to production that teams can use today.

Part 1: Marinela Profi on Agentic AI You Can Ship

Marinela keeps it simple. Agents are not just bigger models. They are systems that plan, act, and learn inside guardrails. We talked about autonomy levels, decisioning, and orchestration with a human in the loop. Her message is direct. LLMs are not agents. You need both deterministic and probabilistic methods, tied to policy. We covered where teams get stuck. Hype over readiness. Data gaps. No orchestration layer. She shared concrete use cases on Viya across banking, insurance, and manufacturing that prove value when governance is baked in from day one.

Watch the interview: Marinela Profi on Agentic AI

Part 2: Manisha Khanna on Trust That Shows Up in the P&L

With Manisha, we went deeper on trust. The SAS–IDC findings are clear. Leaders who win do the basics well. Data lineage. Access control. Model monitoring. Clear ownership. Order matters. Fix data quality and governance first. Then productize. Then scale. Viya guardrails make “safe by default” real with policies, repeatable workflows, and measurable outcomes. Agentic readiness is not a tool choice. It is about reliable data, governed actions, and audit-ready feedback loops.

Watch the interview:
Manisha Khanna on Trust and ROI

Why this matters now

Enterprises chase bigger models and miss the foundation. Agents without governance create risk. Governance without action creates shelfware. The win is the combination. A clean data layer. Clear policies. Monitored models. Orchestrated agents that know what they can and cannot do. That is how you move from demo to daily use.

My take

If you are a CIO or CDO, set the order and hold the line. Start with data quality and access control. Define policies and outcomes. Stand up an orchestration layer that blends deterministic rules with LLM reasoning. Measure with business KPIs, not only model scores. Pick one cross functional use case. Close the loop with monitoring and ownership. Then scale.

What to do next

1. Share both interviews with your data, AI, and risk leads.

2. Map today’s gaps across lineage, access, monitoring, and ownership.

3. Pick a use case and pilot on Viya with policies turned on from day one.

4. Track time to value and show the win.
   

Reply and tell me where you are in this journey. I will feature a few real stories in my next issue.

🔍 Stay Ahead in AI & Data! Join 137K+ Data & AI professionals who stay updated with the latest trends, insights, and innovations.

📢 Want to sponsor or support this newsletter? Reach out and let's collaborate! 🚀

Best,

Ravit Jain

Founder & Host of The Ravit Show