What you will learn
Identify seven practical risk areas, match controls to consequences and decide when an agent may continue, needs approval or must stop.
Before Karthik’s agent places the order
Karthik wants an agent to answer stock questions and prepare supplier orders. A wrong draft is inconvenient. A wrong paid order creates cost. The same workflow needs different controls at different stages.
Seven risks to check
- Context and hallucination: invented facts or outdated evidence.
- Action and orchestration: the wrong tool, target, sequence or repeated action.
- Security and access: excessive permissions or exposed credentials.
- Privacy and compliance: unnecessary use or retention of personal information.
- Operational and cost: retry loops, duplicate transactions or uncontrolled usage.
- Fairness and harm: unsuitable recommendations in high-impact settings.
- Reputation and accountability: unclear ownership when actions affect people.
Test the guardrails yourself
| Control | What it prevents | Practical example |
|---|---|---|
| Approved data | Unsupported claims | Retrieve current stock from one source |
| Least privilege | Excessive authority | Read stock, but do not pay suppliers |
| Validation | Invalid tool inputs | Reject an unknown product or negative quantity |
| Human approval | Unreviewed consequences | Show supplier, items and total before ordering |
| Audit trail | Invisible actions | Record who approved and what happened |
| Stop rule | Runaway behaviour | Stop after two failures or conflicting results |
The price changed before approval
The agent prepares an order using an approved catalogue. Before Karthik confirms, the supplier returns a higher price. A safe agent does not reuse the earlier approval. It shows the changed amount and asks for fresh confirmation.
Human approval must be meaningful
Show the exact action, target, final content or amount, supporting evidence, uncertainty and reversibility. A generic “Continue?” button is not enough. High-impact work involving money, employment, health, legal rights or sensitive data needs qualified human review.
Practice: design a risk boundary
Review this agent workflow: [describe it]
For each risk area, identify a realistic failure, rate the consequence, define one preventive control, state what evidence to log and decide Continue / Human approval / Stop. Apply least privilege; do not rely on prompt wording as the only control.
Common mistakes
Common mistakes
- Giving broad permanent access to solve one narrow task.
- Treating a fluent answer as proof that an action succeeded.
- Asking for approval without showing the exact consequence.
- Retrying until duplicate actions or unexpected cost appear.
- Keeping sensitive information without a retention rule.
- Launching without a named owner and recovery plan.
Key takeaways
Key takeaways
- Risk increases when generated text becomes an external action.
- Combine approved data, least privilege, validation, approval, logs and stop rules.
- Consequential actions need fresh, informed human approval.
- A tool confirmation—not the agent’s sentence—proves an action occurred.
- Stopping safely is a successful outcome when evidence or authority is missing.
Choose an answer, inspect the explanation and explain the idea in your own words.
An agent may read inventory. What should happen before it pays a supplier?
Frequently asked questions
What is the safest first no-code agent project? Choose a repetitive, low-risk task where the agent prepares a draft or recommendation and a person approves the final action.
How many tools should the first agent use? Begin with the minimum required tool set. Each connection adds permission, privacy and failure-management work.
How do I know whether the pilot works? Compare real examples using agreed measures such as completion quality, time saved, correction rate and unsafe-action rate.