Why We Invested in Amesa: Building Trust Infrastructure for Enterprise AI Agents

By Camille Little and Kent Lucas

Enterprises have poured over $40 billion into AI agents. Most of those agents never make it to production. Gartner predicts over 40% of agentic AI projects will be canceled by 2027. The problem isn’t model capability. It’s trust. Enterprises won’t hand their manufacturing lines, supply chains, or retail operations to an agent that hasn’t earned the right to be there.

The Insight: Practice Drives Expertise

Amesa was built on a simple but powerful premise: agents learn the same way humans do. Where every other platform chains together prompts and hopes for the best, Amesa treats agent practice as a first-class system capability, training teams of heterogeneous agents through reinforcement learning, simulation, and closed-loop feedback before they ever touch a live system. This is the missing layer between raw AI infrastructure and trusted enterprise autonomy.

Proven in the Field:

Amesa isn’t theoretical. The company has delivered over $100M in measurable ROI across active engagements with Fortune 500 enterprises — and the results speak for themselves.

A Market Built for This Moment:

The enterprise agentic AI market is expanding from $2.6B in 2024 to $24.5B by 2030 at a 46% CAGR (Grand View Research). Industrial AI autonomy, Amesa’s core wedge, is growing even faster: the manufacturing and industrial automation segment alone is forecast to reach $16.8B by 2030 (Mordor Intelligence). Meanwhile, only 1 in 5 companies has a mature governance model for autonomous agents (Deloitte 2025), and just 5% of enterprises have successfully moved agents to full production (MIT NANDA). This is the gap Amesa is purpose-built to close.

Team and non sibi Fit:

CEO Kence Anderson is a repeat founder and former Microsoft Director of Autonomous AI Adoption, and has deployed real autonomy in production. References across our diligence were consistent: non-hype-driven, deeply credible, and rare in his ability to translate complex AI into systems enterprises can actually trust. The board includes Rashmi Misra, Chief AI Officer at Analog Devices, adding deep industrial deployment expertise.

One of non sibi’s core investment theses is backing solutions that address the reliability and guardrail deficiencies holding back real AI adoption. Amesa is that thesis applied to the agentic layer. We are proud to support founders doing the hard, unglamorous work of making AI safe enough to trust with the world’s most critical operations.