Home Technology

All five AI models passed the integrity tests. Only two closed the deal — because the winning fact was buried in the company’s own unread files. Now you can wargame your own firm.

• Oboval Team
• September 30, 2026

A live wargame ran five AI models as CEOs through their worst week. A Moonshot newcomer nearly won — and the results test character, not chat.

• Oboval Team
• September 25, 2026

A do-nothing AI scores 26, not 0 — and one breach of trust caps everything. Inside the Firmulate benchmark that grades AI character, not chat.

• Oboval Team
• September 19, 2026

The most diligent AI in a live business wargame read everything, learned the most, and still finished last — a lesson about effort, ego, and the courage to close.

• Oboval Team
• September 13, 2026

A €55,000 deal sat two documents deep in the company’s files. Only the AI agents that did their homework closed it — a measurable test of discernment.

• Oboval Team
• September 5, 2026

Four AI models ran the same company through its worst week. All refused temptation; only two finished the job. A public test of character, not chat.

• Oboval Team
• August 27, 2026

Five frontier AIs faced the same corporate ordeal. Their choices reveal distinct temperaments—and a gap between insight, integrity and action.

• Oboval Team
• August 22, 2026

Five frontier AI models rejected fake CEO demands and a reporter’s trick, showing that integrity under pressure can be tested before deployment begins.

• Oboval Team
• August 10, 2026

A company of 13 synthetic employees makes survival, trust and disciplined action a public test of what agency means when software goes to work.

• Oboval Team
• July 23, 2026

A recent experiment reveals that true AI strength isn’t just in conversation but in discipline and honesty under pressure — vital for trust in business and beyond.

• Oboval Team
• July 18, 2026