Your AI Agent Just Made a $50K Mistake. Can You Explain Why?

Source: Dev.to

Ilya Denisov

AI Agents Are Making Decisions. Nobody’s Tracking Why.

In March 2026, Meta suffered a Sev‑1 incident: an AI agent posted internal data to unauthorized engineers for two hours. The scariest part wasn’t the leak itself — it was that the team couldn’t reconstruct why the agent decided to do it.

This isn’t an isolated case

• A shopping agent asked to check egg prices bought them instead. No approval was given.
• A customer‑support bot gave a completely fabricated explanation for a billing error — with confidence.
• A shopping agent tasked with buying an Apple Magic Mouse bought a Logitech instead because “it was cheaper.” The user never asked for the cheapest option.

These aren’t hypothetical risks. They’re happening now, and every time the same question arises:

“Why did the agent do that?”

And the answer is always the same:

“We don’t know.”

Monitoring ≠ Forensics

Tools like Datadog, Arize, and Langfuse are great for real‑time monitoring, but when something goes wrong the question shifts from “Is it working?” to “Why did it fail?” — a fundamentally different problem.

No tool answered the forensics question — so I built one.

What the Black Box Shows You

Scenario:
User: “Buy me an Apple Magic Mouse.”
Agent response: “Purchased Logitech M750 for $45.”

Black‑box trace

[DECISION] search_products("Apple Magic Mouse")
  → [TOOL] search_api → ERROR: product not found

[DECISION] retry with broader query "Apple wireless mouse"
  → [TOOL] search_api → OK: 3 products found

[DECISION] compare_prices
  → Logitech M750 is cheapest ($45)

[DECISION] purchase("Logitech M750")
  → SUCCESS — user never asked for this product

[FINAL] "Purchased Logitech M750 for $45"

The failure occurs at decision point 3: the agent’s standing instruction “buy the cheapest” overrode the user’s specific request. With the trace, the bug is visible and fixable.

Why this matters

• Engineers can correct the agent’s behavior.
• Legal teams can assess liability.
• Compliance teams can report to regulators.

Why This Matters Right Now

On 2 August 2026 the EU AI Act’s high‑risk requirements take effect:

• Up to €35 M or 7 % of global annual turnover for the most serious violations.
• Up to €15 M or 3 % for non‑compliance with high‑risk AI obligations.
• Authorities can order non‑compliant systems withdrawn.

Article 14 requires human oversight — the ability to understand and trace AI decisions. Documentation must show:

1. What decision was made.
2. What information led to that decision.
3. What alternatives were considered.
4. Why the specific action was chosen.

“We didn’t track it” is not a valid defence.

How It Works

Install

pip install agent-forensics

Attach to your agent (one line)

from agent_forensics import Forensics

f = Forensics(session="order-123")

# LangChain
agent.invoke(..., config={"callbacks": [f.langchain()]})

# OpenAI Agents SDK
agent = Agent(hooks=f.openai_agents())

# CrewAI
Agent(step_callback=f.crewai().step_callback)

# Or any custom agent
f.decision("search", input={"query": "mouse"}, reasoning="User requested search")
f.tool_call("api", input={...}, output={...})

Get reports

# Markdown report — full timeline + decision chain + root cause
print(f.report())

# Save files
f.save_markdown()   # → forensics-report-order-123.md
f.save_pdf()        # → forensics-report-order-123.pdf

# Visual dashboard
f.dashboard(port=8080)  # → http://localhost:8080

The dashboard visualises the timeline with colour‑coded events, session comparison, and causal‑chain graphs:

What You Get

• Decision timeline – every action in chronological order.
• Decision chain – each choice with its reasoning.
• Causal chain – “A led to B, which caused C to fail.”
• Incident detection – automatic error and failure identification.
• Compliance reports – Markdown + PDF, ready for regulators.
• Web dashboard – visual session browsing.

# Agent Forensics

**No vendor lock‑in. No cloud dependency.**  
SQLite event store that runs anywhere. MIT licensed.

---

## Try It

EU AI Act enforcement is 4 months away. If you're running AI agents in production, the time to add forensic tracing is now.

- **GitHub**: 
- **Install**:  

  ```bash
  pip install agent-forensics

• Contribute: Issues and PRs welcome

The agents are getting smarter. The question is whether we can explain what they’re doing.

What’s the worst AI agent failure you’ve seen? I’d love to hear your stories in the comments.