What Are Mandates?
Directives that shape how SI communicates — without writing a single line of code
Your brand definition tells SI *who you are*. Your Reusable Knowledge Base gives SI *what to say*. Mandates tell SI *how to behave*.
A mandate is a structured directive that overrides or supplements SI's default behavior for a specific context. You might create a mandate that says "never mention competitor pricing in phone calls," or "always open emails to realtors with a market insight," or "follow up no more than twice per week during the nurturing stage."
When you create a mandate, Zyntro's AI converts your plain-language guidance into a machine-readable JSON schema. That schema gets injected into every relevant interaction — emails SI generates, phone calls Phona conducts, live chat conversations, SMS messages, and reply processing. SI reads the mandate and adjusts its behavior accordingly.
Mandates exist because every business has rules that do not fit neatly into a brand definition. Compliance requirements, industry-specific communication norms, channel-specific tone preferences, competitive positioning rules — these are the kinds of things that vary by audience, product, pipeline stage, or channel. Rather than building custom logic for each user, mandates let you configure SI's behavior yourself, in your own words.
Key Concepts
The category of behavior the mandate controls. There are six types: **Cadence** (timing and frequency), **Communication Flow** (tone and structure), **Hard Facts** (non-negotiable truths), **Prohibitions** (things SI must never do), **Objection Handling** (responses to pushback), and **Comparatives** (competitive positioning rules).
Where the mandate applies. A mandate can be attached to your brand (global), a specific ware (product), an audience segment, a pipeline phase or stage, a contact status, or a communication channel. More specific attachments override more general ones.
The machine-readable JSON output that SI actually consumes. You write guidance in plain language; Zyntro's AI converts it into structured instructions with triggers, rules, and conditions that SI follows during live interactions.
The precedence level that determines which mandate wins when two conflict. From lowest to highest: **foundational** (brand-level), **overlay** (ware/audience), **execution_specific** (stage/channel), **safety_net** (contact status — always wins).
Capabilities
Cadence Control
Define when and how often SI reaches out — max touches per day, channel-specific timing, stop conditions, and escalation logic.
Communication Flow
Shape SI's tone, message structure, opening and closing patterns, CTA style, and interaction behavior per audience or channel.
Hard Facts
Establish non-negotiable truths about your brand, pricing, capabilities, or limitations that SI must always respect and never contradict.
Prohibitions
Create strict guardrails — claims SI must never make, topics it must avoid, actions it must never take. Severity levels from minor to critical.
Objection Handling
Equip SI with structured responses to common pushback — detection signals, core concerns, and approved response strategies.
Competitive Positioning
Define how your brand and products should be compared against competitors — what to emphasize, what language to use, what to avoid.
How It All Fits Together
Mandates sit between your brand definition and SI's real-time decisions. Here is the hierarchy:
Brand Definition sets the foundation — your voice, values, audience segments, and wares. This is who you are. Every piece of AI-generated content starts here.
Mandates add specific behavioral rules on top of the brand definition. They do not replace your brand — they refine how SI applies it in specific contexts. A brand definition says "our tone is professional and warm." A Communication Flow mandate might add "when emailing CFOs, lead with data and ROI metrics before any conversational language."
Reusable Knowledge Base provides the content substance — objections, talking points, FAQs, conversation starters. Mandates control *how* SI uses this material. An Objection Handling mandate might say "when a prospect raises pricing concerns, always acknowledge the concern before presenting value."
SI reads all three layers and generates communication that respects the brand, follows the mandates, and draws from the knowledge base — personalized to each individual contact based on their engagement history.
Mandates are injected as machine-readable JSON into every interaction pipeline:
- Email generation — SI reads active mandates before composing any message
- Phona voice calls — mandates are loaded into the call context so the AI follows them during live conversation
- Live chat — mandate rules are present throughout the chat session
- SMS — mandates govern tone, content, and frequency
- Reply processing — when a contact replies, mandates shape how SI interprets and responds
Mandate Attachment Levels
| Attachment Level | What It Controls | Priority | Example |
|---|---|---|---|
| Brand | Global behavior across all wares, audiences, and channels | Foundational (lowest) | "Always sign off emails with the sender's first name, never the company name" |
| Ware | Behavior specific to one product or service | Overlay | "When discussing the Enterprise plan, emphasize dedicated support and SLA guarantees" |
| Audience Segment | Messaging adapted for a specific persona | Overlay | "When engaging realtors, use real estate terminology and reference MLS data" |
| Pipeline Phase | Strategy for an entire phase (Find, Win, Keep, Grow) | Overlay | "During the Win phase, every touchpoint should include a clear next step" |
| Pipeline Stage | Concrete rules for a specific stage | Execution-specific | "In the Proposal Sent stage, follow up within 48 hours if no response" |
| Communication Channel | Channel-specific formatting and behavior | Execution-specific | "SMS messages must be under 160 characters and never include links" |
| Contact Status | Safety and eligibility boundaries | Safety net (highest) | "DNC contacts: no outbound calls, no SMS, email only with explicit opt-in" |
