Sep 2, 2026 · 25 min
Trust turns enterprise AI from assistance into automation
EP 64: Lack of ROI Is Actually Lack of Trust — Fixing Enterprise AI for Finance | Binny Gill, Kognitos
Finance teams cannot capture AI’s promised returns if every automated decision still requires human supervision.
- 1Persistent business context lets AI agents retrieve operational knowledge across vendors, accounts, controls, and enterprise systems.
- 2Neurosymbolic design pairs flexible language understanding with deterministic execution, making English-based finance procedures more reliable.
- 3Governance, auditability, and complete data lineage are prerequisites for autonomous AI that organizations can trust in production.
Don't miss
Gill reframes weak AI ROI as a trust failure: organizations cannot realize autonomy while humans remain responsible for monitoring every decision.
The brief
Binny Gill, CEO and co-founder of Cognitos, frames finance automation as a trust problem: AI must understand business context before it can safely act.
Cognitos’s Context Graph connects systems such as SAP, Oracle, and NetSuite, then grows through employee interactions that capture operational knowledge.
The platform treats English as a process language, but ambiguity triggers clarification rather than guesswork; neurosymbolic AI then combines language flexibility with deterministic execution.
Gill’s sharpest challenge to current AI economics is that assistance is not autonomy: ROI remains elusive when humans must monitor every decision.
The proposed answer is governed automation, with human-approved procedures, traceable execution, and complete data lineage to make production deployment accountable.
Featuring
Listen to the full episode and explore every guest, topic, and moment on PodLume.