MCP Slack Connector: AI Business Alerts in Your Team Channel

What You Are Actually Solving

• Tell you about the bad review within minutes, not hours, with a suggested reply already drafted
• Surface the Thursday cancellation immediately to the staff member who can fill it
• Send you a 6pm revenue summary so you know whether today was a good day before you leave
• Stay quiet during the 95% of the day when nothing requires your attention

Prerequisites and Architecture

• Slack workspace (free tier works for everything described here)
• Claude API key (Anthropic Console)
• Python 3.10+
• A Slack App with a Bot Token (we will create this below)
• Optional: a running webhook server for real-time events (can be replaced with polling for simpler setup)

Step 1: Create Your Slack App and Get a Bot Token

1. Go to api.slack.com/apps and click Create New App
2. Choose From scratch, name it something like "DataLatte Ops" or "[Your Business] Alerts"
3. Under OAuth & Permissions, add these Bot Token Scopes:
   • chat:write — post messages
   • channels:read — list channels
   • channels:join — join channels to post
   • files:write — upload reports as files
4. Click Install to Workspace and copy the Bot User OAuth Token (starts with xoxb-)
5. Create your alert channels in Slack: #reviews, #bookings, #marketing, #ops-daily
6. Invite your bot to each channel: /invite @YourBotName

Step 2: Configure the Slack MCP Server

{
  "mcpServers": {
    "slack": {
      "command": "npx",
      "args": ["-y", "@modelcontextprotocol/server-slack"],
      "env": {
        "SLACK_BOT_TOKEN": "xoxb-your-bot-token-here",
        "SLACK_TEAM_ID": "T0123456789"
      }
    }
  }
}

pip install anthropic slack-sdk schedule requests flask

Step 3: The Six Alert Automations

Alert 1: New Google Review — With AI-Drafted Reply

import anthropic
import requests
from slack_sdk import WebClient
from datetime import datetime, timedelta

anthropic_client = anthropic.Anthropic(api_key="YOUR_ANTHROPIC_API_KEY")
slack_client = WebClient(token="xoxb-YOUR-BOT-TOKEN")

REVIEWS_CHANNEL = "#reviews"
GBP_API_KEY = "YOUR_GOOGLE_API_KEY"
BUSINESS_NAME = "Paws & Claws Grooming"
LOCATION_ID = "YOUR_GBP_LOCATION_ID"

def check_new_reviews():
    """Poll Google Business Profile API for reviews in the last hour."""
    since = (datetime.utcnow() - timedelta(hours=1)).isoformat() + "Z"
    url = f"https://mybusiness.googleapis.com/v4/accounts/YOUR_ACCOUNT/locations/{LOCATION_ID}/reviews"
    response = requests.get(url, headers={"Authorization": f"Bearer {GBP_API_KEY}"})
    reviews = response.json().get("reviews", [])

    for review in reviews:
        star_rating = review.get("starRating", "FIVE")
        review_text = review.get("comment", "No written comment")
        reviewer_name = review.get("reviewer", {}).get("displayName", "A customer")
        create_time = review.get("createTime", "")

        # Only alert immediately for 1, 2, or 3 star reviews
        if star_rating in ["ONE", "TWO", "THREE"]:
            process_negative_review(reviewer_name, star_rating, review_text)
        else:
            # Queue positive reviews for the daily digest
            pass

def process_negative_review(reviewer, rating, review_text):
    prompt = f"""
A {rating.lower()}-star review was just posted on Google for {BUSINESS_NAME}.

Reviewer: {reviewer}
Review text: "{review_text}"

Please:
1. Write a professional, warm, and empathetic response (under 200 words) that:
   - Thanks the reviewer by first name if possible
   - Acknowledges the specific concern they raised
   - Offers to make it right (e.g., "Please call us at...")
   - Does NOT sound defensive or scripted
2. Rate the urgency: LOW / MEDIUM / HIGH based on the severity and specificity of the complaint

Format your response as:
URGENCY: [level]
SUGGESTED REPLY:
[the reply text]
"""

    response = anthropic_client.messages.create(
        model="claude-opus-4-5",
        max_tokens=600,
        messages=[{"role": "user", "content": prompt}]
    )

    ai_response = response.content[0].text

    # Post to Slack with Block Kit formatting
    slack_client.chat_postMessage(
        channel=REVIEWS_CHANNEL,
        blocks=[
            {
                "type": "header",
                "text": {"type": "plain_text", "text": f"⚠️ New {rating.title()} Star Review"}
            },
            {
                "type": "section",
                "text": {
                    "type": "mrkdwn",
                    "text": f"*Reviewer:* {reviewer}\n*Rating:* {'⭐' * {'ONE':1,'TWO':2,'THREE':3}.get(rating, 0)}\n*Review:* \"{review_text}\""
                }
            },
            {"type": "divider"},
            {
                "type": "section",
                "text": {
                    "type": "mrkdwn",
                    "text": f"*Claude's suggested reply:*\n{ai_response}"
                }
            }
        ]
    )

Alert 2: Daily Revenue Summary at 6pm

import schedule

def post_daily_revenue_summary():
    """Pull today's bookings from your booking system and post a summary."""
    today = datetime.now().strftime("%A, %B %d")

    # Replace with your actual booking system API call
    bookings_today = get_todays_bookings()  # Returns list of appointment dicts
    total_revenue = sum(b.get("price", 0) for b in bookings_today)
    daily_target = 800  # Set your target
    completed = len([b for b in bookings_today if b.get("status") == "completed"])
    cancelled = len([b for b in bookings_today if b.get("status") == "cancelled"])

    prompt = f"""
Today is {today}. Here is the day's performance data for {BUSINESS_NAME}:

- Total revenue: ${total_revenue}
- Daily target: ${daily_target}
- Appointments completed: {completed}
- Cancellations: {cancelled}
- Bookings data: {bookings_today}

Write a brief, friendly Slack summary (3-5 sentences) that:
1. Opens with whether today hit the target (positive if yes, encouraging if no)
2. Highlights one standout fact (busiest hour, most popular service, etc.)
3. Notes if there are open slots tomorrow that could be filled
4. Ends with a one-word "vibe" for the day in brackets, e.g. [Solid] or [Tough one]
"""

    response = anthropic_client.messages.create(
        model="claude-opus-4-5",
        max_tokens=400,
        messages=[{"role": "user", "content": prompt}]
    )

    pct = int((total_revenue / daily_target) * 100)
    bar = "█" * (pct // 10) + "░" * (10 - pct // 10)

    slack_client.chat_postMessage(
        channel="#ops-daily",
        blocks=[
            {
                "type": "header",
                "text": {"type": "plain_text", "text": f"📊 Daily Wrap-Up — {today}"}
            },
            {
                "type": "section",
                "fields": [
                    {"type": "mrkdwn", "text": f"*Revenue*\n${total_revenue} / ${daily_target}"},
                    {"type": "mrkdwn", "text": f"*Progress*\n{bar} {pct}%"},
                    {"type": "mrkdwn", "text": f"*Completed*\n{completed} appointments"},
                    {"type": "mrkdwn", "text": f"*Cancelled*\n{cancelled} appointments"}
                ]
            },
            {
                "type": "section",
                "text": {"type": "mrkdwn", "text": response.content[0].text}
            }
        ]
    )

schedule.every().day.at("18:00").do(post_daily_revenue_summary)

Alert 3: New Booking Notification

def handle_new_booking(booking_data):
    """Called when a new booking comes in via webhook from your booking system."""
    client_name = booking_data.get("client_name")
    service = booking_data.get("service")
    staff_member = booking_data.get("staff_member")
    appointment_time = booking_data.get("datetime")
    is_new_client = booking_data.get("is_new_client", False)

    new_badge = " 🆕 *New client!*" if is_new_client else ""

    slack_client.chat_postMessage(
        channel="#bookings",
        blocks=[
            {
                "type": "section",
                "text": {
                    "type": "mrkdwn",
                    "text": (
                        f"✅ *New Booking*{new_badge}\n"
                        f"*Client:* {client_name}\n"
                        f"*Service:* {service}\n"
                        f"*With:* {staff_member}\n"
                        f"*When:* {appointment_time}"
                    )
                }
            }
        ]
    )

Alert 4: Cancellation Alert with Rebooking Suggestion

def handle_cancellation(cancellation_data):
    cancelled_slot = cancellation_data.get("datetime")
    service = cancellation_data.get("service")
    client_name = cancellation_data.get("client_name")

    prompt = f"""
A booking just cancelled for {BUSINESS_NAME}.

Cancelled appointment:
- Client: {client_name}
- Service: {service}
- Slot: {cancelled_slot}

Suggest one specific action to fill this slot. Consider:
1. Which type of client most commonly books last-minute?
2. Should we post to Instagram Stories about the opening?
3. Is there a waitlist message we should send?

Write a 2-sentence action suggestion for the team, then write a 
ready-to-post Instagram Story caption (under 100 characters) about 
the opening slot.
"""

    response = anthropic_client.messages.create(
        model="claude-opus-4-5",
        max_tokens=300,
        messages=[{"role": "user", "content": prompt}]
    )

    slack_client.chat_postMessage(
        channel="#bookings",
        blocks=[
            {
                "type": "header",
                "text": {"type": "plain_text", "text": "❌ Cancellation — Slot Available"}
            },
            {
                "type": "section",
                "text": {
                    "type": "mrkdwn",
                    "text": (
                        f"*Client:* {client_name}\n"
                        f"*Service:* {service}\n"
                        f"*Open slot:* {cancelled_slot}"
                    )
                }
            },
            {"type": "divider"},
            {
                "type": "section",
                "text": {
                    "type": "mrkdwn",
                    "text": f"*Suggested action:*\n{response.content[0].text}"
                }
            }
        ]
    )

Alert 5: Inventory Low Alert (Coffee Shop Example)

# For coffee shops tracking inventory levels

INVENTORY_THRESHOLDS = {
    "Ethiopian Yirgacheffe": 2.0,   # kg — alert when under 2kg
    "House Blend": 3.0,
    "Oat Milk": 6,                  # cartons
    "Whole Milk": 4,                # gallons
    "Vanilla Syrup": 2              # bottles
}

def check_inventory(current_stock: dict):
    """Pass in current_stock as {item_name: quantity}."""
    low_items = []

    for item, quantity in current_stock.items():
        threshold = INVENTORY_THRESHOLDS.get(item)
        if threshold and quantity <= threshold:
            low_items.append({"item": item, "quantity": quantity, "threshold": threshold})

    if not low_items:
        return  # No alert needed

    prompt = f"""
The following inventory items are running low at {BUSINESS_NAME}:

{low_items}

Write a brief Slack alert (3-4 sentences) that:
1. Lists each low item with current quantity
2. Estimates how many days of stock remain (assume busy weekend = 30% higher usage)
3. Recommends which supplier to contact first based on lead time
   (beans typically need 2 days lead time, dairy is next-day)
"""

    response = anthropic_client.messages.create(
        model="claude-opus-4-5",
        max_tokens=300,
        messages=[{"role": "user", "content": prompt}]
    )

    slack_client.chat_postMessage(
        channel="#ops-daily",
        blocks=[
            {
                "type": "header",
                "text": {"type": "plain_text", "text": "📦 Inventory Alert — Reorder Needed"}
            },
            {
                "type": "section",
                "text": {"type": "mrkdwn", "text": response.content[0].text}
            }
        ]
    )

Alert 6: Weekly Marketing Performance Digest

def post_weekly_marketing_digest():
    # Pull metrics from your marketing platforms via their APIs
    # These are example values — replace with real API calls
    marketing_data = {
        "google_ads_spend": 142.50,
        "google_ads_clicks": 89,
        "google_ads_conversions": 7,
        "meta_ads_spend": 68.00,
        "meta_reach": 4820,
        "meta_link_clicks": 112,
        "new_google_reviews": 3,
        "average_review_rating": 4.8,
        "top_review_keywords": ["friendly", "gentle", "professional", "clean"],
        "website_sessions": 340,
        "top_landing_page": "/for/pet-groomers"
    }

    prompt = f"""
Here is last week's marketing performance data for {BUSINESS_NAME}:

{marketing_data}

Write a Monday morning marketing digest for the Slack #marketing channel.
Format it as:

**MARKETING WEEK IN REVIEW**

WHAT WORKED: [2-3 bullet points highlighting wins]
WATCH OUT FOR: [1-2 potential concerns or things to optimize]
THIS WEEK'S PRIORITY: [one specific action to take this week based on the data]

Keep the total to under 200 words. Be specific about numbers.
Do not use vague language like "good performance" — say exactly what was good and why it matters.
"""

    response = anthropic_client.messages.create(
        model="claude-opus-4-5",
        max_tokens=500,
        messages=[{"role": "user", "content": prompt}]
    )

    slack_client.chat_postMessage(
        channel="#marketing",
        blocks=[
            {
                "type": "header",
                "text": {"type": "plain_text", "text": "📈 Weekly Marketing Digest"}
            },
            {
                "type": "section",
                "text": {"type": "mrkdwn", "text": response.content[0].text}
            }
        ]
    )

schedule.every().monday.at("09:00").do(post_weekly_marketing_digest)

Slack Message Formatting with Block Kit

{
  "blocks": [
    {
      "type": "header",
      "text": {
        "type": "plain_text",
        "text": "Your Alert Title Here",
        "emoji": true
      }
    },
    {
      "type": "section",
      "fields": [
        {
          "type": "mrkdwn",
          "text": "*Field Label:*\nField value here"
        },
        {
          "type": "mrkdwn",
          "text": "*Another Field:*\nAnother value"
        }
      ]
    },
    {
      "type": "divider"
    },
    {
      "type": "section",
      "text": {
        "type": "mrkdwn",
        "text": "Longer text block with *bold*, _italic_, and `code` formatting supported."
      }
    },
    {
      "type": "actions",
      "elements": [
        {
          "type": "button",
          "text": {
            "type": "plain_text",
            "text": "View in Notion"
          },
          "url": "https://notion.so/your-crm-link",
          "style": "primary"
        }
      ]
    }
  ]
}

Channel Routing Strategy

Cost Breakdown: What Does This Actually Cost?

Comparison: MCP + Claude vs Generic Automation Tools

FAQ

Related Articles

• What Is MCP (Model Context Protocol)? A Plain-English Guide for Local Business Owners
• Best MCP Servers for Local Business: The Complete Setup Guide
• MCP + Notion: Build a Self-Updating Client CRM with AI
• MCP + Supabase: Build an AI Agent with a Real Database Backend
• WhatsApp Business API for Local Marketing: The Complete Automation Guide