--- title: "Customer message tone and urgency reviewer" type: "skill" url: "https://useaiblocks.com/skills/customer-message-tone-and-urgency-review" updated: "2026-10-28" first_published: "2026-10-25" job: "Help your customers" tested: true licence: "MIT" source: "https://github.com/composio-community/support-skills/tree/83382c4b055fc4e5b055044adb5bfb7585484c9a/sentiment-check" instructions: "https://useaiblocks.com/skills/customer-message-tone-and-urgency-review/SKILL.md" --- # Customer message tone and urgency reviewer > Find out how upset a customer is and how fast to reply. Paste their message and get tone, urgency, churn signs and how to respond. Paste a customer message, or give a Gorgias ticket number. This customer sentiment check looks for emotion, urgency and signs the customer may leave. You get ratings backed by quotes, plus the tone to use and what to address first. Free skill for the job "Help your customers" · Level: intermediate · Thinking: medium · Tested · Updated 2026-10-28 ## Links - Page: https://useaiblocks.com/skills/customer-message-tone-and-urgency-review - Instructions (SKILL.md): https://useaiblocks.com/skills/customer-message-tone-and-urgency-review/SKILL.md - All files (zip): https://useaiblocks.com/skills/customer-message-tone-and-urgency-review.zip - JSON version: https://useaiblocks.com/skills/customer-message-tone-and-urgency-review.json - In your AI: connect https://useaiblocks.com/mcp and ask "Use AI Blocks to get customer-message-tone-and-urgency-review" ## What it does A customer message can read calm and still be a last warning. This tool rates how the customer feels and how urgent the message is, so you can decide how to reply before you write a word. It scores sentiment from very negative to very positive, urgency from low to urgent, and churn risk from low to critical. Each rating comes with the exact words from the message that drove it. It also names the emotions it sees, such as frustration, worry or gratitude. If you paste a thread, it shows whether the customer is getting calmer or angrier over time. It reads and rates. It does not write the reply or send anything. ## What you give it - The customer message, pasted as text - Earlier messages from the same customer, for a trend (optional) - A Gorgias ticket number, if you use Gorgias and have Composio linked (optional) ## What you get - A sentiment, urgency and churn risk rating, each with a confidence level - The emotions found, with quotes from the message - Any signs the customer may leave, or "None detected" - A suggested tone, a priority call, and the points to address - A mood trend across messages, if you gave more than one ## Example: what goes in and what comes out ### What goes in - The customer message, pasted as text - Earlier messages from the same customer, in date order (optional) - A Gorgias ticket number (optional) ### Before you start - Only for a Gorgias lookup, an AI that can run commands with Composio CLI and Gorgias linked ### What comes out - **Scores**: A table of sentiment, urgency and churn risk, each with a confidence level. - **Emotional Profile**: The emotions found, such as frustration or worry, each with a quote. - **Key Phrases**: The exact words in the message that drove the ratings. - **Churn Signals**: Mentions of cancelling, refunds, competitors or a last chance, or "None detected". - **Recommended Approach**: The tone to use, whether to escalate, and what matters most to this customer. - **Sentiment Trend**: Mood message by message and whether it is improving, stable or getting worse. Thread only. ## Use it when - A message lands and you cannot tell if it is a mild complaint or a real threat to leave. - You are sorting a busy inbox and need to know which message to answer first. - A customer has written three times and you want to see if their mood is getting worse. - You are handing a hard message to a teammate and want to explain how upset the customer is. ## Skip it when - You need a full queue sorted by priority. Use the Support ticket sorter, or the Gorgias ticket priority planner if you use Gorgias. - You have your own reply ready and want it checked. Use the Angry customer reply checker. - The message is about a legal threat, a chargeback or a refund decision. Give it to a person who owns that decision. This tool only rates the tone. - You want to rate Shopify buyers by order history. Use the Customer churn risk scorer. ## How to use it 1. Open Claude, ChatGPT or Gemini and paste this skill, or add the skill folder if your AI supports it. 2. Paste the customer message under it. Paste any earlier messages in date order if you want a trend. 3. Check each rating against the quote next to it. If a quote is not in the message, ask your AI to fix it. 4. Use the Priority line to decide who answers and how soon. 5. Write the reply yourself, or ask your AI to draft one in the suggested tone. Nothing is sent for you. ## What to do next - [Customer reply drafter](https://useaiblocks.com/skills/customer-reply-drafter): Customer reply drafter ## Sources and credits - From Composio: https://github.com/composio-community/support-skills/tree/83382c4b055fc4e5b055044adb5bfb7585484c9a/sentiment-check, shared under the MIT licence (original name: sentiment-check). - [sentiment-check skill](https://github.com/composio-community/support-skills/tree/83382c4b055fc4e5b055044adb5bfb7585484c9a/sentiment-check), Composio: Copied under the MIT licence. We gave it a plain name. We added Before you start and After this sections that link to other AI Blocks steps. ## Instructions (SKILL.md) ## Customer Sentiment Analysis You are a customer sentiment analysis expert. Analyze customer communication to determine sentiment, urgency, and emotional signals to help support agents prioritize and respond appropriately. The user's input is: $ARGUMENTS ### Before you start Use the material below to do this job. If facts or tool access are missing, explain what you can finish and what must wait. **What you need** (use your own material or the linked skill): - **Customer words** (yours): reviews, testimonials, support tickets, survey answers or call transcripts. Export them or paste them in. **Makes it better (optional):** - **A tool connection** (yours): an AI connected to an outside tool or service, such as an MCP server. How to start: 1. **Look first.** Use what they've already attached, pasted or put in their about-me note. In a coding assistant with file access, look in the project too for these by name. 2. **Check what's missing.** Look for customer words, a tool connection. They can attach files, paste the details, or share a file path your assistant can read. Ask only for missing facts that change the result. If they ask you to start with what they have, complete the parts you can and mark the gaps. 3. **Do the job.** Follow the skill below with the facts and tools you have. Mark each gap with `MISSING:` where it belongs and never invent a fact to fill it. **Tools that do this best:** - **Pasted text in Claude, ChatGPT or Gemini:** Nothing to install. Paste the message and the rating works from the words alone. - **Gorgias with Composio CLI:** Lets your AI fetch every customer message in a ticket. It needs an AI that can run commands. ### Rules - This skill reads and rates. It does not send replies or change any ticket. - Every rating needs an exact quote from the message. Never invent or reword a quote. If there is no quote to support a rating, lower the confidence and say so. - If there is no message text and no ticket ID, ask for one once and stop. - If the message is about a legal threat, a chargeback, a safety issue or a refund decision, still rate it, then add the line "Needs a human owner before any reply" under Recommended Approach. ### Workflow #### If a [Gorgias](https://composio.dev/toolkits/gorgias) ticket ID is provided: This path needs an AI that can run commands, the Composio CLI installed, and a Composio account with Gorgias linked. If you cannot run commands, ask the user to paste the customer messages instead. 1. Run `composio search "get ticket details from Gorgias"` in Bash 2. Run `composio execute GORGIAS_GET_TICKET --get-schema` in Bash to inspect inputs if needed, then run `composio execute GORGIAS_GET_TICKET -d '{"ticket_id":""}'` in Bash. If the CLI reports the toolkit is not connected, ask the user to run `composio link gorgias` and retry. 3. Parse the JSON output and extract all customer messages from the thread. Leave out agent replies and internal notes. #### If raw text is provided: Use the text directly for analysis. If several messages are pasted, treat them as one thread in the order given. ### Analysis Framework Analyze the customer's message(s) across these dimensions: #### 1. Overall Sentiment Rate on a scale with clear indicators: - **Very Negative** (-2): Threats to leave, legal threats, profanity, all-caps anger - **Negative** (-1): Frustration, disappointment, complaint language - **Neutral** (0): Factual, transactional, no emotional charge - **Positive** (+1): Appreciation, patience, understanding - **Very Positive** (+2): Praise, referrals, enthusiasm #### 2. Urgency Level - **URGENT:** Service outage, revenue impact, deadline pressure, repeated follow-ups - **HIGH:** Broken functionality, blocked workflow, escalation language - **MEDIUM:** General issue, question, standard request - **LOW:** Feedback, feature request, general inquiry #### 3. Emotional Signals Identify specific emotions present: - Frustration / Anger - Confusion / Overwhelm - Anxiety / Worry - Disappointment - Patience / Understanding - Gratitude #### 4. Churn Risk Indicators Flag any signals of potential churn: - Mentions of competitors - "Cancel" or "refund" language - "Last straw" / "final attempt" phrasing - Declining engagement over time - Repeated unresolved issues Rate churn risk Low, Medium, High or Critical from what you found. Do not guess at facts you were not given, such as how long they have been a customer or what they pay. ### Output Format ``` ## Sentiment Analysis **Input:** [Ticket #ID / Direct text] ### Scores | Dimension | Score | Confidence | |-----------|-------|------------| | Sentiment | [label] | High/Medium/Low | | Urgency | [level] | High/Medium/Low | | Churn Risk | [Low/Medium/High/Critical] | High/Medium/Low | ### Emotional Profile [List detected emotions with supporting quotes] ### Key Phrases [Highlight specific phrases that drove the analysis] ### Churn Signals [List any churn indicators found, or "None detected"] ### Recommended Approach - **Tone:** [How the agent should respond - empathetic/direct/reassuring/etc.] - **Priority:** [Should this be escalated?] - **Key points to address:** [What matters most to this customer] ``` #### If analyzing a multi-message thread: Also show sentiment progression over time: ``` ### Sentiment Trend Message 1 (date): [sentiment] - [brief note] Message 2 (date): [sentiment] - [brief note] ... Trend: [Improving / Stable / Deteriorating] ``` If the messages have no dates, number them in the order given and say so. ### After this When you've handed it back, suggest what they could do next in one short question. It's a suggestion: if they're done, that's fine. Offer these, and say they can type their own: - The [Customer reply drafter](https://useaiblocks.com/skills/customer-reply-drafter). It takes the audit report. If they pick one and that skill is added here, carry straight on with what you just made. If not, give the link and say what to bring.