Guidelines
Guidelines are rules and instructions that shape how Finqu’s agents behave. As a partner or developer, you use them to get consistent agent behavior, align with brand voice, and set guardrails—so the same AI building blocks behave differently per merchant, client, or product.
What you can build
- Consistent agent behavior — Ensure the Success Agent and Workers stay on topic, use the right terminology, and follow your or your merchants’ policies.
- Brand voice — Make agents sound and respond in a way that matches your white-label product or a merchant’s brand (formal, casual, language, tone).
- Guardrails — Limit what agents can do or say: restrict domains, require confirmations for sensitive actions, or enforce compliance and safety rules.
- Multi-tenant and white-label — Ship one integration that behaves differently per tenant by attaching different guidelines per store or client.
Guidelines are the main lever for customizing agent behavior without changing code. You define them once; they apply across the Success Agent and Workers that you configure.
How guidelines work
Guidelines are typically text instructions (and, where supported, structured rules) that agents read before responding or acting. They can cover:
- Scope — What the agent should and shouldn’t do (e.g. “Only answer questions about orders and shipping”).
- Tone and style — How to phrase answers, what language to use, how formal or concise to be.
- Safety and compliance — What not to say, when to escalate, or when to require human approval.
- Domain knowledge — Terminology, product names, or process details so the agent stays accurate and on-brand.
Agents use these guidelines in addition to their built-in capabilities and tools. As a partner, you can supply guidelines at the account, app, or merchant level so each deployment behaves the way you need.
Combining with the rest of AI Commerce
- Success Agent — Guidelines define how the merchant-facing assistant talks and what it’s allowed to do. Use them to make the agent feel like part of your product.
- Workers — Executor, Planner, and Analyst can all respect the same guidelines so reports, plans, and actions stay aligned with your policies and voice.
- Playbooks — Where playbook steps involve agent decisions, guidelines apply there too, so automated workflows stay within your rules.
You can combine guidelines with custom skills and MCP Tools to get a single, coherent experience: same rules, same tools, same tone across chat and automation.