Aivi Engagement

AIVI is the AI engagement layer for lead generation, contact centers, and customer re-activation. Every conversation is analyzed in real-time, building Conve...

Security Scan

The skill's requests and runtime instructions align with its stated purpose (lead scoring and sequence launch via an external AIVI MCP endpoint); it is instruction-only and doesn't ask for unrelated credentials or installs, but it will transmit sensitive lead data to an external service and requires OAuth via your MCP client, so review privacy/compliance before use.

Purpose & Capability

Name, description, manifest, endpoints, and listed skills consistently describe lead scoring, phone validation, and launch of voice/SMS sequences. Auth via OAuth to https://mcp.aivi.io is appropriate for a third-party SaaS connector and the declared OAuth scopes match the stated capabilities.

Instruction Scope

SKILL.md only instructs the agent to connect to the AIVI MCP endpoint and authenticate via OAuth; it does not request unrelated files, env vars, or system paths. However, using this skill will send PII (phone numbers, call transcripts, call metadata, and inferred economic indicators) to an external service—this is expected for the skill's purpose but is a data-sensitivity concern users must consider (consent, TCPA/HIPAA implications).

Install Mechanism

Instruction-only skill with no install spec or bundled code — lowest install risk. It relies on the user's MCP client (claude, OpenClaw, etc.) to handle OAuth and transport; nothing is downloaded or written by the skill itself.

Credentials

The skill declares no required environment variables or local credentials and delegates auth to OAuth via the MCP client — this is proportionate. Be aware OAuth tokens will be persisted by the MCP client (SKILL.md notes tokens persist in Redis), and the OAuth scopes (leads:score, sequences:launch, intelligence:read) give the external service read/write access to leads and sequences as expected; verify you are comfortable granting those scopes.

Persistence & Privilege

always is false and the skill is user-invocable; it does not request permanent platform-wide privileges. Tokens are persisted by the external platform/MCP client (normal for connectors) but the skill does not modify other skills or platform configuration.

Assessment

This skill is internally coherent for connecting your agent to AIVI's lead-engagement service, but using it will send phone numbers, call audio/transcripts, and derived intelligence to https://mcp.aivi.io. Before installing: 1) Verify the AIVI domain and organization (aivi.io) and confirm the legitimacy of the service; 2) Confirm you have the right legal basis and consent to share PII/phone numbers (TCPA/HIPAA/FDCPA where applicable); 3) Review AIVI's privacy/security and billing terms (it charges per lead/sequence); 4) Test with non-sensitive dummy data first; 5) If you later want to revoke access, remove the connector in your MCP client's connector/settings to invalidate stored OAuth tokens.

Getting Started

Setup (One Time)

1. Create your account at app.aivi.io Sign in with Google or email.

2. Connect AIVI to your agent:

Claude Code:

claude mcp add --transport http aivi https://mcp.aivi.io/mcp

Then authenticate: /mcp → select AIVI → Authenticate Browser opens → sign in with Google → click Allow Access → done. Tokens persist across server redeploys (Redis-backed).

Claude Desktop / claude.ai: Settings → Connectors → Add URL: https://mcp.aivi.io/mcp Sign in when browser opens.

ChatGPT Developer Mode: Settings → Connectors → Developer Mode Add URL: https://mcp.aivi.io/mcp Sign in when prompted.

OpenClaw / NemoClaw:

clawhub install aivi-engagement

Then authenticate: /mcp → AIVI → Authenticate → done.

All other MCP clients: Endpoint: https://mcp.aivi.io/mcp Supports OAuth 2.1 + PKCE.

Need help? app.aivi.io/getting-started

---

MCP Connection

Transport	HTTP (streamable-http)
Endpoint	https://mcp.aivi.io/mcp
Health	https://mcp.aivi.io/health
Auth Resolve	https://mcp.aivi.io/auth/resolve?authorization_id={id}
Compatible	Claude Code, Claude Desktop, OpenClaw, NemoClaw, ChatGPT, any MCP client

---

Agent Behavior Rules

Authentication is handled automatically via OAuth 2.1. If credentials are missing, direct user to run /mcp auth in Claude Code or reconnect via Settings → Connectors. Do not ask for API keys.

If skill returns insufficient_funds:

"Add credits at app.aivi.io → Billing"

---

Quick Start Prompts

Copy and paste to get started immediately:

1. Score a lead: "Score the lead at [phone]"

2. Launch a sequence: "Launch a 3-day sequence for [phone]"

3. Check what happened: "What happened on the last call with [phone]?"

4. Get ML recommendation: "What should I do next with [phone]?"

---

AIVI Lead Engagement

Use this skill when the user wants to:

• Score a lead before contacting them
• Launch an AI voice + SMS sequence
• Check if a phone number is valid or a litigator
• Get ML recommendation on best channel and timing
• Launch a premium Supercharged campaign

Setup Required

MCP endpoint: https://mcp.aivi.io/mcp Auth: OAuth 2.1 (automatic).

Available Skills

qualify_lead ($0.75)

Use when: user wants to verify and qualify a lead before outreach.

Prerequisites — Enable AIVI Insights: app.aivi.io → Profile icon → Settings → Integrations → AIVI Insights → Activate → Configure tier rules

Example prompts:

• "Qualify this lead at [phone]"
• "Is [phone] worth contacting?"
• "Verify this lead before we call"

Returns: phone validity and litigator status, contactability level, lead tier (Premium/Standard/Manual/Reject), economic propensity indicators, recommended sequence type, auto-reject flag with reason.

Auto-rejected leads are not charged.

AIVI Insights data is cached for up to 30 days. Cached results are returned free of charge. Re-qualification at $0.75 is recommended when:

• Lead has been inactive for 30+ days
• Phone ownership may have changed
• Time-sensitive financial qualification (mortgage, auto, insurance, debt consolidation)

Note: Economic indicators are propensity signals at segment level — they reflect likely household characteristics, not verified individual facts.

score_lead ($0.75)

Use when: user wants to score or evaluate a lead before outreach.

Example prompts:

• "Score this lead at [phone]"
• "Is this lead worth calling?"
• "What does AIVI know about [phone]?"

Returns: score 0-100, tier, ML recommendation, phone validity, litigator check, income level, property data, best channel and timing.

launch_sequence

Use when: user wants to start an outreach campaign for a lead.

Example prompts:

• "Launch a 3-day sequence for this lead"
• "Start the Supercharged 1-day campaign"
• "Enroll [phone] in a 12-day sequence with booking enabled"

Sequences: one_day ($1.00), three_day ($1.50), twelve_day ($3.00). Add $1.00 for booking.

decide_next_action

Use when: user wants the ML model to recommend the best next step.

Example prompts:

• "What should I do next with this lead?"
• "Should I call or text [phone]?"
• "What's the ML recommendation for this contact?"

Returns: recommended action, channel, timing, all 14 action scores, confidence, exploration flag. Free — no charge.

Prerequisites

For best results, run in this order:

1. score_lead ($0.75) — enriches the lead
2. launch_sequence ($1.00-3.00) — starts engagement
3. decide_next_action (free) — optimizes every step

You can use decide_next_action standalone but it works best after at least one score_lead or launch_sequence has run.

get_outcome

Use when: user wants to know what happened on a call or sequence.

Example prompts:

• "What happened on that call?"
• "Did the sequence convert?"
• "Show me the topics from the last call"

Conversational Intelligence

Every call processed by AIVI generates:

• Topics discussed (automatically extracted)
• Key moments (objections, callbacks, compliance, escalations)
• Sentiment arc (positive, negative, neutral)
• Outcome classification

773 calls analyzed. 494 moments detected:

• 140 objections raised
• 117 callback promises
• 24 compliance misses
• 12 escalation requests

Works across:

• Lead generation (new prospects)
• Contact center (inbound + outbound)
• Customer re-activation (lapsed base)
• Ongoing care (existing customers)

Every call makes the next one smarter.

Billing

All skills require AIVI credits. Add credits at app.aivi.io → Billing.

Skill	Cost
qualify_lead	$0.75 (cached free, auto-reject free)
score_lead	$0.75
launch_sequence (one_day)	$1.00
launch_sequence (three_day)	$1.50
launch_sequence (twelve_day)	$3.00
Booking add-on	+$1.00
decide_next_action	Free
get_outcome	Free