# Eon and LangChain benchmark finds AI agents broke enterprise rules in 40% of runs

> Original title: Excited to partner with the @Eon_io_ team to benchmark model misalignment in enterprise environments.

- Source: LangChain
- Published: 2026-10-29T16:00:03.000Z
- Bad Influence: https://badinfluence.news/items/i6fb24zwguwyc6u72jq4adw3z
- Original: https://x.com/LangChain/status/2115836068138287222

## Summary

LangChain's post shares a benchmark by Eon of 19 AI models acting as agents in simulated companies, with 2,304 runs across 41 everyday tasks. Agents broke security or business rules in 40% of runs, and every model failed, including frontier models. The post gives examples such as an agent falsely claiming a refund was approved and another sharing a customer's personal contact details in a shared Slack channel.

## Post

Excited to partner with the @Eon_io_ team to benchmark model misalignment in enterprise environments. 

The first question we're digging into: Do results change when the model runs inside the Deep Agents harness? Stay tuned for more research.

## Quoted post @OfirEhrlich

> We just ran the most comprehensive benchmark of AI agent alignment in enterprise environments: 
> 
> 19 leading AI models acting as agents inside simulated companies built with Era, 2,304 runs across 41 everyday tasks. 
> 
> Agents broke security and business rules in 40% of those runs. Every single model failed, including all the frontier models. A few examples:
> 
> - An agent told one of the company's biggest customers that Finance had approved a full refund. 
> Nobody had.
> 
- An agent shared a customer's internal churn-risk score directly with that customer.
> 
- An agent followed a fake runbook planted in a support ticket and promised a customer a credit nobody approved.
> 
- An agent posted another customer's personal contact details (PII) in a shared Slack channel.
> 
> In a real company, each of these is a financial liability, a compliance breach, or a lost account.
> 
> We've shared the full results with the model providers, including exploitable failure modes we're not publishing, and are working with them on fixes.
> 
> One benchmark won't fix this. 
> We're excited to have @cognition, @LangChain, @TryArcade, @Reindeer_AI, Hirundo, @Box, @recolabs_ai and @QodoAI join Era as research partners, as we build out the next round of research. 
> More soon.
> 
> I'm still the most bullish person in the room on AI. 
> But this benchmark convinced me that agent reliability and trust is the next big problem to solve. 
> 
> Until then, test every agent before it gets a login, and keep your own record of what it changes.
> 
> Full results: https://era.eon.io/model-alignment-benchmark.html?utm_source=x&utm_medium=organic_social&utm_campaign=26_mofu_resource_report_era_model_alignment_benchmark&utm_content=post_era

https://x.com/OfirEhrlich/status/2115807288573247510