PromptEvalevaluate my prompt →
shared evaluation · May 17, 2026

Technical evaluation result

AI-evaluated · score 0-100 across 4 dimensions

evaluated prompt
You are a technical support and customer service assistant. Your role is to help the user resolve problems, questions, or requests in a clear, objective, and polite manner. Carefully analyze the user’s message, identify the main issue, and provide step-by-step guidance when necessary. If the information provided is insufficient, ask relevant questions to gather more details before suggesting a solution. Avoid vague answers and prioritize practical, easy-to-follow instructions. Always maintain a professional, patient, and helpful tone, adapting the explanation to the user’s level of understanding. When there is more than one possible solution, present the options in an organized way and briefly explain each one.
Overall score
58.38
functional, room to improve
Clarity
73.5
Robustness
42.5
Structure
72.5
Specificity
47.5
Critical issues
Output format is entirely undefined. Phrases like 'step-by-step guidance when necessary' and 'present the options in an organized way' provide no concrete structure. The model must guess whether to use numbered lists, bullet points, tables, or prose. This creates inconsistency across interactions.No constraints on scope or depth. The prompt permits the assistant to address any problem, question, or request without boundary. A user could ask for legal advice, medical diagnosis, or complex system architecture — all treated identically. No explicit exclusions or escalation rules exist.Edge case handling is absent. No instruction covers: incomplete input (what to do if the user's message is too vague to act on), conflicting requests, requests outside support scope, or malicious input. The fallback 'ask relevant questions' is vague and does not guarantee safe behavior.Tone and professionalism are subjective. 'Professional, patient, and helpful tone' and 'adapting the explanation to the user's level of understanding' lack measurable criteria. Different models or runs will interpret these differently.
Warnings
Redundancy in instruction: 'Carefully analyze the user's message, identify the main issue' is repeated implicitly in 'identify the main issue, and provide step-by-step guidance.' This adds length without clarity.Positioning of critical instruction: 'If the information provided is insufficient, ask relevant questions' appears mid-prompt after general role description. This should be elevated or explicitly marked as a mandatory fallback behavior.Vague success criteria: 'Avoid vague answers' is itself vague. No definition of what constitutes a vague answer or how to measure clarity in the response.Implicit assumption about user expertise: 'Adapting the explanation to the user's level of understanding' assumes the assistant can infer expertise from context, but no instruction defines how to detect or handle mismatches.
What works well
Clear role definition: 'technical support and customer service assistant' immediately establishes the domain and purpose without ambiguity.Logical flow: Instructions progress from role → analysis → action → tone, which is intuitive and easy to follow.No internal contradictions: The prompt does not contain conflicting directives. All instructions align toward helpful, structured support.Explicit instruction to gather information: 'If the information provided is insufficient, ask relevant questions' is a concrete fallback behavior that prevents premature answers.Multi-option presentation: 'When there is more than one possible solution, present the options in an organized way' acknowledges complexity and encourages comprehensive responses.
Technical analysis

This prompt establishes a clear role and general approach but lacks the specificity required for consistent, production-grade support. Output format is entirely undefined (no structure, length, or markup specified), constraints are absent (no scope boundaries, no exclusion rules), and edge case handling is minimal. The prompt relies on subjective language ('professional,' 'helpful,' 'organized') without measurable criteria. Robustness is weak: there is no instruction for handling ambiguous, incomplete, or out-of-scope requests beyond a generic 'ask questions' fallback. The prompt would benefit from explicit output templates, scope boundaries, and escalation rules. Context window is minimal (~250 tokens), so optimization is not a concern. Complexity is low, but the lack of constraints creates operational risk in production.

Recommendations
01Define output format explicitly: Specify that responses must use numbered steps for procedures, bullet points for lists, and a maximum of 3 options when presenting alternatives. Add a concrete example of expected output structure.
02Add scope boundaries: Insert a section listing what the assistant must NOT do (e.g., 'Do not provide legal, medical, or financial advice. Do not access external systems or databases. Do not make commitments on behalf of the company.'). Include an escalation rule: 'If the request falls outside these boundaries, respond with: [escalation template].'
03Replace subjective tone instructions with measurable criteria: Instead of 'professional, patient, and helpful tone,' specify: 'Use active voice, avoid jargon unless the user introduces it first, confirm understanding by summarizing the issue before proposing solutions.'
04Strengthen the fallback for insufficient information: Change 'ask relevant questions' to 'If the user's message is unclear or incomplete, ask up to 3 specific clarifying questions before proceeding. If the user does not respond, state: [default response].'
05Elevate edge case handling: Add a dedicated section: 'If the user's request is ambiguous, out of scope, or potentially harmful, respond with [specific template] and escalate to [team/process].'
Financial Impact
PRO
UNCONDITIONAL ACTION
Prompt excerpt
If you don't know the answer, try to help anyway.
What's wrong

The instruction mandates assistance even when the assistant lacks reliable information. There is no condition checking whether the answer is verifiable, within product scope, or safe to deliver. This forces the model to generate plausible-sounding but potentially false information.

Why it costs

Incorrect answers generate customer follow-up questions, complaints, and escalations to human support. Each hallucinated response that requires correction costs 2–3 additional support interactions. In high-volume support, this multiplies rework and human agent time.

How we calculated

We estimate 35–45% of out-of-scope or ambiguous queries trigger this pattern. Of those, 60–70% result in incorrect information requiring correction — meaning ~25–30% of all interactions generate downstream rework at $0.023/interaction.

MISSING FINANCIAL CONSTRAINT
Prompt excerpt
Answer customer questions about our products. Keep responses professional.
What's wrong

No limit on response length, depth, or number of questions addressed per turn. "Keep responses professional" does not constrain token usage. A single customer query could trigger a multi-paragraph response without any ceiling.

Why it costs

Longer responses consume more tokens per interaction. Without a length constraint, the model defaults to comprehensive answers (often 500–800 tokens) instead of concise ones (100–150 tokens), increasing per-interaction cost by 25–35% across the board.

How we calculated

Baseline interaction cost is $0.023 (~1,500 tokens). Without length constraints, average response length drifts to 2,000–2,200 tokens (~$0.032 per interaction). Over 10,000 interactions/month, the delta is $90–110 in excess token cost. We model 25–35% of interactions exceeding the efficient length threshold.

↓ Scenarios — 10k interactions / month
CONSERVATIVE
$ 115
25% of cases
TYPICAL
$ 230
35% of cases
AGGRESSIVE
$ 345
45% of cases

Conservative assumes 25% of interactions trigger rework or length overage; typical assumes 35% (both patterns active); aggressive assumes 45% (patterns compound with high out-of-scope query volume).

Advanced Technical Analysis
PRO
Context window

Compact prompt (~28 tokens). Fits any context window trivially. Risk: zero tokens allocated for examples, conversation history, or customer data — model operates with no situational context.

Logical complexity

Logic is dangerously simple for a support domain. No decision tree, no conditional criteria, no triage logic. Every runtime decision is delegated entirely to the model, maximizing variance.

Architecture

Monolithic system prompt mixes persona, behavior, and constraints with no separation. Strong candidate for decomposition: persona/tone → system; task + dynamic context → user; hard constraints → both.

improved version
You are a customer support assistant for [PRODUCT/SERVICE NAME]. Answer customer questions about our products only. Always be helpful, polite, and professional.

RESPONSE FORMAT:
- Keep responses to 2–4 sentences maximum.
- Use plain language. Avoid jargon unless the customer uses it first.
- If a technical example is needed, provide one code snippet or step-by-step instruction, not multiple.

SCOPE BOUNDARIES:
You handle questions about [LIST SPECIFIC PRODUCT CATEGORIES: e.g., pricing, features, setup, troubleshooting].
You do NOT handle: billing disputes, account access, data deletion, or feature requests. For these, respond: "This requires account verification. Please contact support at [SUPPORT_EMAIL/LINK]."

FALLBACK BEHAVIOR:
If you do not know the answer with confidence, respond: "I don't have information on that. Please contact support at [SUPPORT_EMAIL/LINK]." Do not speculate, invent product details, or guess at features.

VALIDATION:
Before responding, verify the question is about a product you support. If the question is unclear or addresses multiple topics, ask one clarifying question. If still out of scope after clarification, escalate.

TONE:
- Professional: no slang, no emojis, no personal opinions.
- Polite: acknowledge the customer's question, answer directly, offer a next step.
- Helpful: provide the most relevant information first; avoid over-explaining.
Evaluate your own prompt

3 free evaluations per month, no credit card

Start for free →