Future 100 – 2025 Winner: Energent.ai

An autonomous AI agent that verifies AI-generated analysis against source documents, catching hallucinations before delivery.

Energent.ai

You're reading Entrepreneur Middle East, an international franchise of Entrepreneur Media.

  • FOUNDED IN: 2025
  • LOCATION: ABU DHABI
  • SECTOR: FINTECH
Future 100 - 2025 Winner: Energent.ai
Energent.ai Founders Rachel Hu and Kimi Kong

Enterprises now deploy AI across finance, operations, and research faster than they can check its work — and one hallucinated figure in a board deck or client report carries real cost. Energent.ai closes that gap with an autonomous AI quality-control agent that reviews AI-generated work the way a second analyst would: it re-derives every number from the original files, recounts rows, recomputes totals, and traces each claim back to its source across 150+ formats, from spreadsheets and scanned PDFs to CAD drawings and handwriting.

Teams receive a shareable, audit-ready verdict instead of another answer to double-check. Crucially, verification is independent of the model that produced the work. The platform ranks #1 on Hugging Face’s data-agent leaderboard at 94.4% accuracy —30% ahead of Google’s agent— and serves Fortune 500 teams in finance, logistics, telecom, and retail. Hub71 and Y Combinator-backed, founded by AI researchers from AWS and Google DeepMind.

  • FOUNDED IN: 2025
  • LOCATION: ABU DHABI
  • SECTOR: FINTECH
Future 100 - 2025 Winner: Energent.ai
Energent.ai Founders Rachel Hu and Kimi Kong

Enterprises now deploy AI across finance, operations, and research faster than they can check its work — and one hallucinated figure in a board deck or client report carries real cost. Energent.ai closes that gap with an autonomous AI quality-control agent that reviews AI-generated work the way a second analyst would: it re-derives every number from the original files, recounts rows, recomputes totals, and traces each claim back to its source across 150+ formats, from spreadsheets and scanned PDFs to CAD drawings and handwriting.

Teams receive a shareable, audit-ready verdict instead of another answer to double-check. Crucially, verification is independent of the model that produced the work. The platform ranks #1 on Hugging Face’s data-agent leaderboard at 94.4% accuracy —30% ahead of Google’s agent— and serves Fortune 500 teams in finance, logistics, telecom, and retail. Hub71 and Y Combinator-backed, founded by AI researchers from AWS and Google DeepMind.

Related Content