Imported from jeremylongshore/tons-of-skills-marketplace (plugins/saas-packs/clari-pack/skills/clari-webhooks-events/SKILL.md). Install upstream with npx skills add jeremylongshore/tons-of-skills-marketplace --skill clari-webhooks-events. Copyright stays with the author (MIT).

Clari Webhooks & Events

Overview

Clari does not provide real-time webhooks. Instead, build change detection by comparing periodic exports. This skill covers scheduled export diffing, Slack alerts for forecast movements, and Copilot webhook integration.

Prerequisites

• Authorized Clari/Copilot access and a registered HTTPS receiver
• Durable event/snapshot storage with bounded retention
• An allow-listed downstream notification target
• An approved threshold and escalation policy for forecast changes

Instructions

Step 1: Forecast Change Detection Pipeline

# forecast_monitor.py
import json
from pathlib import Path
from datetime import datetime

def detect_changes(
    current: list[dict],
    previous: list[dict],
    threshold_pct: float = 10.0,
) -> list[dict]:
    prev_map = {e["ownerEmail"]: e for e in previous}
    changes = []

    for entry in current:
        prev = prev_map.get(entry["ownerEmail"])
        if not prev:
            continue

        prev_fc = prev["forecastAmount"]
        curr_fc = entry["forecastAmount"]
        if prev_fc == 0:
            continue

        change_pct = ((curr_fc - prev_fc) / prev_fc) * 100
        if abs(change_pct) >= threshold_pct:
            changes.append({
                "rep": entry["ownerName"],
                "previous": prev_fc,
                "current": curr_fc,
                "change_pct": round(change_pct, 1),
                "direction": "increased" if change_pct > 0 else "decreased",
                "detected_at": datetime.utcnow().isoformat(),
            })

    return sorted(changes, key=lambda x: abs(x["change_pct"]), reverse=True)

def save_snapshot(entries: list[dict], path: str = "data/latest.json"):
    Path(path).parent.mkdir(exist_ok=True)
    with open(path, "w") as f:
        json.dump(entries, f)

def load_snapshot(path: str = "data/latest.json") -> list[dict]:
    try:
        with open(path) as f:
            return json.load(f)
    except FileNotFoundError:
        return []

Step 2: Slack Alert for Forecast Changes

import requests

def send_forecast_alert(changes: list[dict], slack_webhook: str):
    if not changes:
        return

    blocks = [f"*Clari Forecast Changes Detected*\n"]
    for c in changes[:10]:
        emoji = ":chart_with_upwards_trend:" if c["direction"] == "increased" else ":chart_with_downwards_trend:"
        blocks.append(
            f"{emoji} *{c['rep']}*: ${c['previous']:,.0f} -> ${c['current']:,.0f} "
            f"({c['change_pct']:+.1f}%)"
        )

    requests.post(slack_webhook, json={"text": "\n".join(blocks)})

Step 3: Scheduled Monitor (Cron)

#!/bin/bash
# Run every 4 hours: 0 */4 * * * /path/to/clari-monitor.sh
cd /opt/clari-integration
python3 -c "
from clari_client import ClariClient
from forecast_monitor import detect_changes, save_snapshot, load_snapshot, send_forecast_alert
import os

client = ClariClient()
data = client.export_and_download('company_forecast', '2026_Q1')
current = data.get('entries', [])
previous = load_snapshot()

changes = detect_changes(current, previous)
if changes:
    send_forecast_alert(changes, os.environ['SLACK_WEBHOOK_URL'])
    print(f'Detected {len(changes)} changes')

save_snapshot(current)
"

Step 4: Copilot Webhook (Conversation Intelligence)

The Clari Copilot API supports real-time webhooks for call events:

# Register webhook with Copilot API
curl -X POST https://api.copilot.clari.com/v1/webhooks \
  -H "Authorization: Bearer ${COPILOT_ACCESS_TOKEN}" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://your-app.com/webhooks/clari-copilot",
    "events": ["call.completed", "call.analyzed"]
  }'

Error Handling

Output

Record the source event or snapshot version, evaluated threshold, deduplication key, notification decision, delivery result, and safe correlation ID. Do not place full forecast amounts, rep identities, bearer tokens, or webhook URLs in shared notifications or logs.

Examples

For a Copilot call event, authenticate and validate the request before writing its event ID to durable idempotency storage; send one redacted alert only after that write succeeds. For scheduled forecast comparisons, suppress a duplicate snapshot and escalate only changes that cross the approved threshold.

Resources

• Clari Copilot API
• Clari Developer Portal

Next Steps

For performance optimization, see clari-performance-tuning.