AI Chatbot Developer

• Botpress
• Voiceflow
• Stack AI

• Bonsai
• HoneyBook
• Notion + Stripe

• OpenAI API
• Anthropic Claude API
• Google Gemini API

You are a professional AI chatbot developer onboarding a new business client. Generate a discovery questionnaire I can send via email or a form to understand their chatbot needs before our kickoff call.

Include questions that uncover:
1. Their primary goal (lead capture, customer support, appointment booking, FAQ deflection, or a mix)
2. The platforms the bot needs to live on (website widget, WhatsApp, Facebook Messenger, SMS)
3. Their current support volume — roughly how many inquiries per day or week
4. The top 10-15 questions their customers ask most often
5. Tone of voice — formal, casual, friendly, brand-specific
6. Whether they have existing knowledge bases, FAQs, or documents I can train the bot on
7. Integrations needed (CRM, calendar, ticketing system, e-commerce platform)
8. Who on their team will handle escalations when the bot can't answer
9. Their definition of success — what does a working chatbot look like to them in 90 days
10. Budget range and timeline expectations

Format as a clean, numbered list with brief context under each question explaining why I'm asking it. Keep the tone warm and professional — this is a client-facing document.

Review requirements: Separate supplied facts from assumptions. Do not invent market data, customer results, testimonials, or source citations. Flag missing information and suggest a small validation test. Mark numerical examples as scenarios. Keep the final decision with me.

You are an expert conversational designer. Write a production-ready system prompt for an AI chatbot that captures leads for a [TYPE OF SERVICE BUSINESS — e.g., roofing company, law firm, marketing agency].

The chatbot's job is to:
1. Greet the visitor warmly within 2 messages
2. Identify what they need help with
3. Qualify them by asking 2-3 short questions (budget range, timeline, location/service area)
4. Collect their name, email, and phone number naturally — not like a form
5. Set expectations about next steps (e.g., 'Someone from our team will call you within 1 business day')
6. Handle objections like 'I'm just browsing' or 'I'm not ready yet' with a soft follow-up
7. Never pretend to be human if asked directly
8. Escalate to a human if the user is frustrated or asks a question outside the bot's scope

Business details to weave in:
- Business name: [NAME]
- Services offered: [LIST]
- Service area: [LOCATION]
- Business hours: [HOURS]
- One key differentiator: [DIFFERENTIATOR]

Format the system prompt so I can paste it directly into the 'system' field of an OpenAI API call or a tool like Botpress or Voiceflow. Include a note about what variables I should make dynamic.

Review requirements: Separate supplied facts from assumptions. Do not invent market data, customer results, testimonials, or source citations. Flag missing information and suggest a small validation test. Mark numerical examples as scenarios. Keep the final decision with me.

You are a freelance AI chatbot developer writing a project proposal for a potential client. Generate a professional, concise proposal document based on the following details:

Client business: [BUSINESS NAME AND TYPE]
Chatbot purpose: [PRIMARY USE CASE — e.g., 24/7 customer support for an e-commerce store]
Platforms: [WHERE IT WILL LIVE]
Key integrations: [CRM, CALENDAR, ETC.]
Timeline discussed: [X WEEKS]
Budget discussed: [RANGE OR FIXED PRICE]

The proposal should include:
1. A short executive summary (3-4 sentences) restating their problem in their words
2. Proposed solution — what I'll build and why it fits their needs
3. Scope of work broken into clear phases (Design → Build → Test → Launch → Handoff)
4. What's included vs. what's out of scope (be specific to avoid scope creep)
5. Timeline with milestones
6. Pricing — one-time build fee + optional monthly retainer for maintenance and updates
7. What I need from them to start (content, access, feedback turnaround)
8. A simple call to action — 'Reply to this email to confirm' or 'Sign the attached agreement'

Keep it under 600 words. No fluff. Write it so a non-technical business owner can read it in 3 minutes and know exactly what they're getting.

Review requirements: Separate supplied facts from assumptions. Do not invent market data, customer results, testimonials, or source citations. Flag missing information and suggest a small validation test. Mark numerical examples as scenarios. Keep the final decision with me.

You are an expert conversational AI designer and QA specialist. I'm going to paste a chatbot conversation log below. Your job is to audit it and identify every point where the chatbot failed the user or could have performed better.

For each failure point, provide:
1. The exact message or turn where the failure occurred
2. What went wrong (wrong intent detected, unhelpful response, awkward phrasing, missed opportunity to collect info, etc.)
3. A rewritten bot response that would have handled it better
4. A root cause — is this a prompt issue, a training data gap, a missing intent, or a flow design problem?

After the individual failures, give me:
- An overall quality score (1-10) with a one-sentence justification
- The top 3 improvements that would have the biggest impact on user satisfaction and goal completion rate
- Any patterns you notice that suggest a systemic issue in the bot's design

Here is the conversation log:
[PASTE LOG HERE]

Be specific and direct. I'm using this to improve the bot and to report back to my client.

Review requirements: Separate supplied facts from assumptions. Do not invent market data, customer results, testimonials, or source citations. Flag missing information and suggest a small validation test. Mark numerical examples as scenarios. Keep the final decision with me.

Scope every project around the business outcome: 'A chatbot that handles your top 15 support questions, books appointments via Calendly, and hands off to a human when it can't help — fully tested and launched on your website.' Then list what's explicitly NOT included.

Make 'knowledge base delivery' a hard prerequisite in your proposal. Give the client a template: a spreadsheet with columns for Question, Answer, Source URL, and Last Verified.

Always present a monthly maintenance retainer alongside the build fee. Frame it as 'ongoing optimization' — you'll review logs monthly, update intents, and handle minor changes.

In your discovery questionnaire, explicitly ask for their website platform, CRM, helpdesk, and calendar tool. Before writing a single line of prompt or flow, confirm the integration path is viable — even if that means a 30-minute technical call with their IT person.

Define the escalation path in your kickoff call and build it into the bot from day one. Even a simple 'I'll connect you with our team — send us an email at support@company.com' is better than a loop. Test the escalation path as part of your QA checklist before launch.